Adsorption and its Applications in Industry and Environmental Protection : Applications in Environmental Protection (Studies in Surface Science and Catalysis, 120 a-B.)

Adsorption and Its Applications in Industry and Enviromntal Protection: Vol. 2: Applications in Environmental Protection, by A. Dabrowski

• ISBN: 0444501665 • Publisher: Elsevier Science & Technology Books • Pub. Date: January 1999

PREFACE Every aspect of h u m a n activity is closely connected with the natural environment. Whether or not we are aware, or care, every day each of us interacts with and affects our environment. The rapid development of technology, especially at the end of 20th century, has increased enormously man's ability to produce goods which, in turn, have enhanced his standard of living. On the other hand, this development has also generated a secondary phenomenon, the environment pollution. Such effect led to deterioration of life quality. Thus, improvement of the life quality owing to innovative technologies caused negative effects for the environment. In order to keep the balance between technology development and main components of the man's environment the appropriate technologies should be used which appear to be a powerful force for the improvement of the environment. The relevant activities for upgrading the quality of ground water, drinking water, soil and air have to be developed. The environmental changes affect also the h u m a n health. Only few chemical compounds present in the h u m a n close surrounding may be considered as beneficial for health. The majority of them act harmfully on humans, even in minimal doses. They occur in our environmental media - air, water and soil and that is why we observe the increasing efforts devoted to the h u m a n environmental protection. One of the most important factors in this field are the possibilities and results of modern chemical analyses of pollutants in biological fluids to maintain h u m a n health. Water is one of the most important components of our environment. Nowadays, the drinking water is becoming more and more scarce, but our demand for water is becoming greater and greater. A very important problem is concerned with the rising levels of nutrients such as nitrates and phosphates in the surface water. Their presence has caused a serious deterioration in the water quality of many rivers, lakes and reservoirs. Therefore the attention has to be given to the removal of nutrients originating from sewages and fertilizers by adsorption methods, ion-exchange and relevant biotechnological techniques. Phosphorous and its compounds dissolved in the ground waters are responsible for the eutrophication in the closed water system, especially in lakes and highly enclosed bays where water is stagnant. Slag media, wasted by - products from steel industries, are effective adsorbents for phosphorous and its compounds. The earth atmosphere along with water, is the main component of our environment. One essential cause of pollution of the air is the tendency to decrease the cost of manufacturing goods by the use of contaminated raw materials without purifying or enriching them before their application. A preliminary desulfurization of coal is still rare. When air is used as a source of

vi oxygen, nitrogen in the air is a diluent which, after the oxygen consumption, is discharged into the atmosphere together with other impurities. Dusts and smogs are another group of air contaminants. The modern technologies should restrict emissions of carbon dioxide to prevent from increasing the amount of heat being dispersed into the atmosphere. This increase, leading to a change of climate, is the greenhouse effect. The other fundamental problem is connected with the removal of volatile organic chloride (VOC) compounds from ground water and recovery of chlorofluorocarbons (CFCs), which are still used in refrigeration and cooling systems. Emission control of ozone depletion by CFCs is very urgent. The pressure on industry to decrease the emission of various pollutants into the environment is increasing. A broad range of methods is available and developed to control and remove both natural and anthropogenic, municipal, agricultural and other pollutants. In relation to the price/performance, adsorption technologies are the most important techniques to overcome the degradation of environmental quality. They play a significant role both in environmental and h u m a n health control and in prevention from global warming and ozone layer depletion. The neccessity to reduce the ozone depletion gases like CFCs and the demand for primary energy diversification in the air conditioning sector, are the main reasons for the increasing interest in adsorption devices considered as alternative to the traditional compressor heat pumps in the cooling systems. Adsorption processes are the ,,heart" of several safety energy technologies which can find suitable applications in the domestic sectors as reversible heat pumps, and in the industrial sectors as refrigerating systems and heat trasnformers using industrial waste heat as the primary energy source. They can also be used for technologies to be applied in the transportation sectors, for automobile air conditioning or for food preservation in trucks. The adsorption dessicant dehumidification technology is also emerging as an alternative to vapour compression systems for cooling and conditioning air for a space. Dessicant base systems can improve indoor air quality and remove air pollutants due to their coadsorption by the dessicant materials. Moreover, a number of microorganisms are removed or killed by the dessicant. Other problems are production of drinking water, removal of anthropogenic pollutants from air, soil and water as well as removal of microorganisms from the indoor air and other important tasks to solve in terms of adsorption technologies. Adsorption can also be expected to play a significant role in the environmental control and life supporting systems or planetary bases, where sorbents may be used to process the habitat air or to recover useful substances from the local environments. Another environmental dilemma deals with the removal of thermal SOx and NOx from hot combustion gases. The above mentioned problems may be solved by advanced adsorption techniques. Among them, the rapid pressure swing adsorption (PSA) methods are very efficient for solving both global and local environmental issues. By the term of global environmental problem is meant emission of ozone depletion gases like CFCs, VOC and emission of green-house gases (CO2, CH4, N20, etc.), but the term local environmental problem deals with flue gas recovery (SOx and NOx),

vii solvent vapour fractionation and solvent vapour recovery, wastewater treatment and drinking water production. Other environmental issues concern the industrial solid aerosols, which are the incomplete combustion products. They are harmful as precursors to the synthesis of strong toxins, carcinogenes and mutagenes. Automobiles contribute substantially to man-made hydrocarbon emissions. A new type of activated carbon filtres for the application in Evaporative Loss Central Devices (ELCD) were developed by NORIT. Automobiles had to pass the so-called SHED emission test, which was legislated in Europe in 1992. Adsorption of metals into living or dead cells has been termed biosorption. Biosorption dealing with the metal - microbe interactions include both terrestrial and marine environments. Biosorption by the sea bacteria plays a significant role in detoxification of heavy metals in the aqueous systems. The literature on the influence of biosorption in metal crystal formation is rather scant. The subject of microbe participation in nucleation and halite crystal growth is important with regard to the influence of cell surface layer (S-layer) components on the crystal habit. As follows from the above considerations, the subject of utility of modern adsorption technologies has enormous environmental, economic and legal importance and constitutes a serious challenge with the prospects for further intense development. Likwise to volume I which contains the most important industrial applications of adsorption, this volume includes the chapters written by authoritative specialists on the broad spectrum of environmental topics to find a way for intense anthropogenic activities to coexist with the natural environment. Some of the topics presented in this volume were mentioned above. However, I would like to highlight a wide spectrum of themes referring to the environmental analysis and environmental control, molecular modelling of both sorbents and adsorption environmentally friendly processes, new trends in applications of colloidal science for protecting soil systems, purification and production of drinking water, water and ground water treatment, new environmental adsorbents for removal of pollutants from waste waters and sewages, selective sorbents for hot combustion gases, some corrosion aspects and ecological adsorption of heating and cooling pumps. This book is divided into two volumes, consisting of chapters arranged in a consistent order, though some chapters could be connected with the industrial (volume I) or environmental (volume II) fields. In order to highlight for readers all topics and considerations each volume of the monograph comprises the complete contents and the complete list of authors, but ncludes its own subject index only. It should be emphasized that all contributions were subjected to a rigorous review process, with almost all papers receiving two reviews from a panel of approximately fifty reviewers. The presented chapters give not only brief current knowledge about the studied problems, but are also a source of topical literature on it. Thus each chapter constitutes an excellent literature guide for a given topic and encourages

viii the potential reader to get to know a problem in detail and for further specialistic studies. At the end of the volume the comprehensive bibliography on adsorptive separations, environmental applications, PSA, parametric pumping, ion-exchange and chromatography is presented which includes the period 1967-1997.All the articles give both the scientific background of the phenomena discussed and indicate practical aspects to a great extent. Consequently, this monograph is addressed to a large group of research workers both in academic institutions and industrial laboratories, whose professional activities are related to widely understood surface environmental problems, including environmental analysis, environmental catalysis and biocatalysis,modern adsorption ecologicallyfriendlly technologies, etc. This book is meant also for students of graduate and postgraduate courses. I am aware, that the panorama of the researches presented is incomplete.On the other hand, I believe that this monograph is a substantial step presenting the current trends and the state of the art. I would like to express my warmest thanks to all the contributors for their efforts to develop the topical environmental fields of great importance. Finally, I wish acknowledge the great help I had my wife, Mrs. Iwona D@rowska, during all stages of the growth of the monograph.Her patience, encouragment and support made it possible to appear this book in present form.

Lublin, September, 1998.

A.Dqbrowski (ed.)

Complete List of Authors

1.

A l e x a n d r a t o s S.D. Department of Chemistry, University of Tennessee at Knoxville, Knoxville, TN 37996-1600, USA 2. A n d r u s h k o v a O.V. Department of Total and Bioorganic Chemistry, Novosibirsk Medical Institute, Krasny Prospekt 52, Novosibirsk 630091, Russia 3. Baldini F. Instituto di Ricerca sulle Onde Elettromagnetiche ,,Nello Carrara", CNR, Via Panciatichi 64, 50127 Firenze, Italy 4. B a n d o s z T.J. Department of Chemistry, City College of New York, New York, NY 10031, USA 5. Blom J. Tauw Milieu P.O.Box 133, 7400 AC Deventer, The Netherlands 6. Bl~dek J. Institute of Chemistry, Military University of Technology, Kaliskiego 2, 01-489 Warsaw, Poland 7. Boere J.A. NORIT N.V., Research & Development, Nijverheidsweg - Noord 72, P.O.Box 105, 3800 AC Amersfoort, The Netherlands 8. Bogillo V.I. Institute of Surface Chemistry, National Academy of Sciences, Prospekt Nauki 31, 252022 Kiev, Ukraine 9. B r a c c i S . Centro di Studio sulle Cause di Deperimento e Metodi di Conservazione Opere d'Arte, CNR, Via G.Capponi 9, 50121 Firenze, Italy 10. Billow M. The BOC Group Gases Technical Center, 100 Mountain Ave., Murray Hill, NJ 07974, USA 11. B u c z e k B. Faculty of Fuels and Energy, University of Mining and Metallurgy, 30-059 Cracow, Poland

12. B u r k e M. University of Arizona, Old Chemistry Bldg., Tucson, AZ 85721, USA 13. Cacciola G. National Council of Research, Institute for Research on Chemical Methods and Processes for Energy Storage and Transformation, S.Lucia sopra Contesse, 98126 Messina, Italy 14. Carey T.R. Radian International, LLC, 8501 N.Mopac Blvd., Austin, TX 78759, USA 15. Cerofolini G.F. SGS-THOMSON Microelectronics, 20041 Agrate MI, Italy 16. C h a n g R. Electric Power Research Institute, 3412 Hillview Ave., Palo Alto, CA 94403, USA 17. Chen J. Georgia Institute of Technology, School of Civil and Environmental Engineering, Atlanta, GA 30332-0512, USA 18. Chen S. Illinois State Geological Survey, 615 E. Peabody Dr. Champaign, IL 61820, USA 19. D a b o u X. Chemical Process Engineering Laboratory, Department of Chemical Engineering, Aristotle University of Thessaloniki and Chemical Process Engineering Research Institute, PO Box 1520, Thessaloniki 54006, Greece 20. D a l l B a u m a n L.A. NASA Johnson Space Center, Houston, TX 77058, USA 21. D ~ b r o w s k i A. Faculty of Chemistry, M.Curie-Sktodowska University, 20031 Lublin, Poland 22. Deka R.C. India Catalysis Division, National Chemical Laboratory, Pune - 411008, India 23. Deng S.G. USA Department of Chemical Engineering, University of Cincinnati, Cincinnati, Ohio 45221, USA 24. D o b r o w o l s k i R. Faculty of Chemistry, M.Curie-Sklodowska University, 20031 Lublin, Poland 25. D o m i n g o - G a r c i a M. Grupo de InvestigaciSn en Carbones, Dpto. de Quimica Inorganica, Fac. de Ciencias, Universidad de Granada, 18071 Granada, Spain

xi 26. D y b k o A. Department of Chemistry, Warsaw University of Technology, Noakowskiego 3, 00-664 Warsaw, Poland 27. F a d o n i M. Department of Physical Chemistry and Electrochemistry, University of Milan, Via Golgi 19, 20133 Milan, Italy 28. F e r n a n d e z - M o r a l e s I. Grupo de InvestigaciSn en Carbones, Dpto. de Quimica Inorganica, Fac. de Ciencias, Universidad de Granada, 18071 Granada, Spain 29. F i n n J.E. NASA Ames Research Center, Moffett Field CA, USA 30. F l e m i n g H. Cochrane Inc., 800 3nd Avenue, King of Prussia, 19406 PA, USA 31. G h o s h T.K. Particulate Systems Research Center, Nuclear Engineering Program, E 2434 Engineering Building East, University of Missouri-Columbia, Columbia, MO 65211, USA 32. G h z a o u i A.E1. UM II LAMMI ESA 5079, Case 015, Place Eugene Bataillon, 34095 Montpellier Cedex 5, France 33. Golden T.C. Air Products and Chemicals, Inc., 7201 Hamilton Boulevard, Allentown, PA 18195-1501, USA 34. G r o s z e k A.J. MICROSCAL LTD, 79 Southern Row, London W 10 5 AL, UK 35. H a u k k a S. Microchemistry Ltd., P.O.Box 132, FIN-02631 Espoo, Finland 36. H e i j m a n S.G.J. KIWA Research and Consultancy, P.O.Box 1072, 3430 BB Nieuwegein, The Netherlands 37. Hines A.L. Honda of America Mfg.Inc., 24 000 Honda Parkway, Marysville, OH 43040, USA 38. H o p m a n R. KIWA Research and Consultancy, P.O.Box 1072, 3430 BB Nieuwegein, The Netherlands 39. H o r v a t h G. University of Veszprem, H-8201 Veszprem, P.O.Box 158, Egyetem u.10, Hungary

xii 40. H s i H-C.

41.

42.

43. 44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

University of Illinois, Environmental Enegineering Program, 205 N.Mathews Ave., Urbana, IL 61801, USA H u b i c k i Z. Faculty of Chemistry, M.Curie-Sktodowska University, 20031 Lublin, Poland I s u p o v V.P. Institute of Solid State Chemistry and Raw Mineral Processing Kutateladze-18, 630128, Novosibirsk, Russia I v e r s o n I. Department of Chemistry, University of Nevada, Reno, NV 89557, USA Izmailova V.N. Moscow State University, Department Colloid Chemistry, Vorob'evy Gory, 119899 Moscow, Russia J a k o w i c z A. Faculty of Chemistry, M.Curie-Sklodowska University, 20031 Lublin, Poland J a n u s z W. Faculty of Chemistry, M.Curie-Sktodowska University, 20031 Lublin, Poland Kalvoda R. J.Heyrovsky Inst.Phys.Chem., Czech Acad. Scis, Dolejskova 3, 18223 Prague 8, Czech Republic K a n e k o K. Chiba University, Department of Chemistry, Faculty of Science, 1-33 Yayoi, Inage, Chiba 263, Japan Kanellopoulos N. Institute of Physical Chemistry NCSR ,,DEMOKRITOS", Aghia Paraskevi Attikis, GR- 153 10, Athenes, Greece K i k k i n i d e s E.S. Institute of Physical Chemistry NCSR ,,DEMOKRITOS", Aghia Paraskevi Attikis, GR- 153 10, Athenes, Greece K i r i c h i e n k o O.A. Institute of Solid State Chemistry, SB RAS, Kutateladze 18, Novosibirsk 630128, Russia Kleut D.v.d. NORIT N.V., Research & Development, Nijverheidsweg - Noord 72, P.O.Box 105, 3800 AC Amersfoort, The Netherlands Kobal I. Department of Physical and Environmental Chemistry, J.Stefan Institute, 61000 Ljubljana, Slovenia

xiii 54. Kotsupalo N.P. Ekostar - Nautech Company, B.Chmielnitsky 2, 630075 Novosibirsk, Russia 55. Krebs K.-F. Merck KGaA, LAB CHROM Synthese, D-64271 Darmstadt, Germany 56. Kubo M. Department of Molecular Chemistry and Engineering, Faculty of Engineering, Tohoku University, Sendai 980-77, Japan 57. L a k o m a a E.-L. Neste Oy, Technology Center, P.O.Box 310, FIN-06101 Porvoo, Finland 58. Lemcoff N.O. The BOC Group, 100 Mountain Avenue, Murray Hill, NJ 07974, USA 59. Lin Y.S. USA Department of Chemical Engineering, University of Cincinnati, Cincinnati, Ohio 45221, USA 60. Liu Y. Department of Chemical Engineering, Swearingen Engineering Center, University of South Carolina, Columbia, SC 29208, USA 61. Long R. Department of Chemical Engineering, The University of Michigan, Ann Arbor, Michigan 48109-2136, USA 62. Lopez-Cortes A. Center for Biological Research, P.O. Box 128, La Paz 23000, BCS, Mexico 63. Lopez-Garzon F.J. Grupo de Investigaci6n en Carbones, Dpto. de Quimica Inorganica, Fac. de Ciencias, Universidad de Granada, 18071 Granada, Spain 64. Lucarelli L. ThermoQuest Italy S.p.A., Strada Rivoltana, 20090 Rodano (Milan), Italy 65. L u o R.G.

66. 67. 68.

69.

Department of Chemical Engineering, Chemistry and Environmental Science, New Jersey Institute of Technology, University Heights, Newark, NJ 07102-1982, USA Lutz W. Holzmarktstrasse 73, D-10179 Berlin, Germany Lodyga A. Fertilizers Research Institute, 24110 Putawy, Poland L u k a s z e w s k i Z. Poznafl University of Technology, Institute of Chemistry and Technical Electrochemistry, Piotrowo 3, 60-965 Poznafl, Poland MacDowall J.D. NORIT United Kingdom Ltd., Clydesmill Place, Cambuslang Industrial Estate, Glasgow G32 8RF, Scotland

xiv 70. Matyska M. Department of Chemistry, San Jose State University, San Jose, CA 95192 USA 71. Matijevic E. Center for Advanced Materials Processing, Clarkson University, P.O.Box 5814, Potsdam, New York 13699-5814, USA 72. Meda L. EniChem - Istituto Guido Donegani, 28100 Novara NO, Italy 73. Menzeres L.T. Ekostar - Nautech Company, B.Chmielnitsky 2, 630075 Novosibirsk, Russia 74. Meyer K. Bundesanstalt ffir Materialforschung und -prfifung (BAM), Zweiggelande Adlershof, Rudower Chaussee 5, D-12489 Berlin, Germany 75. Mitropoulos A.Ch. Institute of Physical Chemistry NCSR ,,DEMOKRITOS", Aghia Paraskevi Attikis, GR- 153 10, Athenes, Greece 76. Miyamoto A. Department of Molecular Chemistry and Engineering, Faculty of Engineering, Tohoku University, Sendai 980-77, Japan 77. M i z u k a m i K. Department of Molecular Chemistry and Engineering, Faculty of Engineering, Tohoku University, Sendai 980-77, Japan 78. Moon H. Department of Chemical Technology, Chonnam National University, Kwangju 500-757, Korea 79. Moreno-Castilla C. Grupo de Investigaci6n en Carbones, Dpto. de Quimica Inorganica, Fac. de Ciencias, Universidad de Granada, 18071 Granada, Spain 80. Neffe S. Institute of Chemistry, Military University of Technology, Kaliskiego 2, 01-489 Warsaw, Poland 81. N e m u d r y A.P. Institute of Solid State Chemistry and Raw Mineral Processing, Kutateladze-18, 630128, Novosibirsk, Russia 82. Nijdam D. Tauw Milieu, P.O.Box 133, 7400 AC Deventer, The Netherlands 83. Ochoa J.L. Center for Biological Research, P.O.Box 128, La Paz 23000, BCS, Mexico 84. P a n G. Department of Earth Sciences, University of Leeds, Leeds LS2 9JT, UK

XV

85. P a r t y k a S. UM II LAMMI ESA 5079, Case 015, Place Eugene Bataillon, 34095 Montpellier Cedex 5, France 86. P a t e l D.C. Department of Chemical Engineering, Chemistry and Environmental Science, New Jersey Institute of Technology, University Heights, Newark, NJ 07102-1982, USA 87. P e s e k J. Department of Chemistry, San Jose State University, San Jose, CA 95192, USA 88. P o k r o v s k i y V.A. Institute of Surface Chemistry, National Academy of Sciences, Prospekt Nauki 31, 252022 Kiev, Ukraine 89. Raisglid M. University of Arizona, Old Chemistry Bldg., Tucson, AZ 85721, USA 90. R a m a r a o B.V. Syracuse University, Faculty of Paper Science and Engineering and Engineering, SUNY, College of Environmental Science and Forestry, Syracuse, NY 13210, USA 91. R a o M.B.

92.

93.

94.

95. 96.

97.

Air Products and Chemicals, Inc., 7201 Hamilton Boulevard, Allentown, PA 18195-1501, USA Ray M.S. Department of Chemical Engineering, Curtin University of Technology, GPO Box U1987, Perth 6845, Western Australia R e i m e r i n k W.M.T.M. NORIT N.V., Research & Development, Nijverheidsweg - Noord 72, P.O.Box 105, 3800 Ac Amersfoort, The Netherlands R e s t u c c i a G. National Council of Research, Institute for Research on Chemical Methods and Processes for Energy Storage and Transformation, S.Lucia sopra Contesse, 98126 Messina, Italy R i c h a r d s o n C.F. Radian International, LLC, 8501 N.Mopac Blvd., Austin, TX 78759, USA R i p p e r g e r K.P. Department of Chemistry, University of Tennessee at Knoxville, Knoxville, TN 37996-1600, USA R i t t e r J.A. University of South Carolina, Department of Chemical Engineering, Swearingen Engineering Center, Columbia, South Carolina 29208, USA

xvi 98. Robens E. Institut ffir Anorganische Chemie und Analytische Chemie der J.Gutenberg-Universitat D-55099 Mainz, Germany 99. R o d r i g u e s A.E. Laboratory of Separation and Reaction Engineering, University of Porto, 4099 Porto Codex, Portugal 100. Rood M. University of Illinois, Environmental Engineering Program, 205 N.Mathews Ave., Urbana, IL 61801, USA 101. R o s e n h o o v e r W. CONSOL, 4000 Brownsville Rd., Library, PA 15129, USA 102. Rostam-Abadi M. Illinois State Geological Survey, 615 E. Peabody Dr. Champaign, IL 61820, USA 103. R u l e J.

College of Sciences, Old Dominion University, Norfolk, VA 23529-0163, USA 104. Saba J. Faculty of Chemistry, M.Curie-Sklodowska University, 20031 Lublin, Poland 105. Sakellaropoulos G.P. Chemical Process Engineering Laboratory, Department of Chemical Engineering, Aristotle University of Thessaloniki and Chemical Process Engineering Research Institute, PO Box 1520, Thessaloniki 54006, Greece 106. S a m a r a s P. Chemical Process Engineering Laboratory, Department of Chemical Engineering, Aristotle University of Thessaloniki and Chemical Process Engineering Research Institute, P.O. Box 1520, Thessaloniki 54006, Greece 107. S h i n t a n i H.

108.

109.

110.

111.

National Institute of Hygienic Sciences, 18-1 Kamiyoga 1-Chome, Setagaya-ku, Tokyo 158, Japan Silva da F.A. Laboratory of Separation and Reaction Engineering, University of Porto, 4099 Porto Codex, Portugal Silva J.A.C. Laboratory of Separation and Reaction Engineering, University of Porto, 4099 Porto Codex, Portugal Sircar S. Air Products and Chemicals, Inc., 7201 Hamilton Boulevard, Allentown, PA 18195-1501, USA S i v a s a n k e r S. Catalysis Division, National Chemical Laboratory, Pune - 411008, India

xvii 112. Stubos A.K. Institute of Nuclear Technology and Radiation Protection, NCSR ,,DEMOKRITOS", 15310 Aghia Paraskevi Attikis, GR-15310, Athenes, Greece 113. S u b r a m a n i a n D. University of South Carolina, Department of Chemical Engineering, Swearingen Engineering Center, Columbia, South Carolina 29208, USA 114. Suckow M. Fachhochschule Lausitz, Grossenhainer Strasse, D-01968 Senftenberg, Germany 115. S u n t o l a T. Microchemistry Ltd., P.O.Box 132, FIN-02631 Espoo, Finland 116. Suzuki M. Institute of Industrial Science, University of Tokyo, 7-221 Roppongi, Minato-ku, Tokyo 106, Japan 117. Szczypa J. Faculty of Chemistry, M.Curie-Sktodowska University, 20031 Lublin, Poland 118. S y k u t K. Faculty of Chemistry, M.Curie-Sklodowska University, 20031 Lublin, Poland 119. Swi~ttkowski A. Institute of Chemistry, Military Technical Academy, Kaliskiego 2, 01-489 Warsaw, Poland 120. T a k a b a H. Department of Molecular Chemistry and Engineering, Faculty of Engineering, Tohoku University, Sendai 980-77, Japan 121. T a m - C h a n g S.-W. Department of Chemistry, University of Nevada, Reno, NV 89557, USA 122. T a r a s e v i c h Yu.I. Institute of Colloid Chemistry and Chemistry of Water, 42 Vernadsky avenue, Kiev 252680, Ukraine 123. TSth J. Hungarian Academy of Sciences, Research Laboratory for Mining Chemistry, 3515 Miskolc-Egyetemvaros, P.O. Box 2, Hungary 124. Tzevelekos K.P. Institute of Physical Chemistry NCSR ,,DEMOKRITOS", Aghia Paraskevi Attikis, GR-153 10, Athenes, Greece 125. Unger K.K. Institut ffir Anorganische Chemic und Analytische Chemic der J.Gutenberg-Universitat, D-55099 Mainz, Germany

xviii 126. U s h a k o v V.A. Institute of Solid State Chemistry, SB RAS, Kutateladze 18, Novosibirsk 630128, Russia 127. V a n s a n t E.F. Laboratory of Inorganic Chemistry, University of Antwerpen (U.I.A.), Universiteitsplein 1, 2610 Wilrijk, Belgium 128. Vetrivel R. Catalysis Division, National Chemical Laboratory, Pune - 411008, India 129. V i g n e s w a r a n S. University of Technology, Sydney, Faculty of Engineering, Building 2, Level 5 P.O.Box 123 Broadway, NSW 2007, Australia 130. W a g h m o d e S.B. Catalysis Division, National Chemical Laboratory, Pune - 411008, India 131. W r 6 b l e w s k i W. Department of Chemistry, Warsaw University of Technology, Noakowskiego 3, 00-664 Warsaw, Poland 132. Y a m p o l s k a y a G.P. Moscow State University, Department Colloid Chemistry, Vorob'evy Gory, 119899 Moscow, Russia 133. Y a n g R.T. Department of Chemical Engineering, The University of Michigan, Ann Arbor, Michigan 48109-2136, USA 134. Y i a c o u m i S. Georgia Institute of Technology, School of Civil and Environmental Engineering, Atlanta, GA 30332-0512, USA

xix

Contents of V o l u m e I Preface Complete List of A u t h o r s

v IX

F u n d a m e n t a l s of A d s o r p t i o n 1. Adsorption - its development and applications for practical purposes (A.D@rowski) 2. Industrial carbon adsorbents (A.Swi~tkowski) 3. Standarization of sorption measurements and reference materials for dispersed and porous solids (E.Robens, K.-F.Krebs, K.Meyer, K.K.Unger) 4. Spectroscopic characterization of chemically modified oxide surfaces (J.Pesek, M.Matyska) 5. Advances in characterisation of adsorbents by flow adsorption microcalorimetry (A.J.Groszek) 6. Temperature programmed desorption, reduction, oxidation and flow chemisorption for the characterisation of heterogeneous catalysts. Theoretical aspects, instrumentation and applications (M.Fadoni, L.Lucarelli) 7. Adsorption with soft adsorbents and adsorbates. Theory and practice (G.F.Cerofolini, L.Meda, T.J.Bandosz)

3 69 95 117 143

177 227

A p p l i c a t i o n in I n d u s t r y 1. Advanced technical tools for the solution of high capacity adsorption separation (G.Horvath, M.Suzuki) 2. The mutual transformation of hydrogen sulphide and carbonyl sulphide and its role for gas desulphurization processes with zeolitic molecular sieve sorbents (M.B(ilow, W.Lutz, M.Suckow) 3. Nitrogen separation from air by pressure swing adsorption (N.O.Lemcoff) 4. Methodology of gas adsorption process design. Separation of propane/propylene and rgiso- paraffins mixtures (Jose A.C.Silva, F.Avelino da Silva, Alirio E.Rodrigues) 5. Fractionation of air by zeolites (S.Sircar, M.B.Rao, T.C.Golden) 6. Production, characterization and applications of carbon molecular sieves from a high ash Greek lignite (P.Samaras, X.Dabou, G.P.Sakellaropoulos) 7. Development of carbon-based adsorbents for removal of mercury emissions from coal combustion flue gas (M.Rostam-Abadi, H-C.Hsi, S.Chen, M.Rood, R.Chang, T.R.Carey, C.F.Richardson, W.Rosenhoover) 8. Sorption properties of gas/coal systems, degasification of coal seams (J.TSth) 9. The influence of properties within particles of active carbons on selected adsorption processes (B.Buczek)

275

301 347

371 395 425

459 485 507

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

Environmental pollutants and application of the adsorption phenomena for their analyses J. Btadek and S. Neffe Institute of Chemistry, Military University of Technology, 00-908 Warsaw, Kaliskiego St., 2, Poland 1. I N T R O D U C T I O N H u m a n activity is now altering the global environment on an unprecedented scale and thus contributes to the environmental change affecting h u m a n health. Only few of chemical compounds present in direct h u m a n surrounding may be considered as beneficial for health; the majority of them act harmfully on humans, even in minimal doses. They occur in all environmental media (air, water and soil) and that is why we observe the increasing attention to the environmental protection. One of the most important factors in this field are the results of chemical analysis of pollutants. It is obvious that only reliable analytical data obtained during monitoring can be a base for environmental protection activities. The term monitoring means systematic and planned collection of analytical activities realised in any space to define the quality of air, water and soil. Volatile organic compounds, pesticides, polycyclic aromatic hydrocarbons, polycyclic aromatic heterocycles, phenols, polychlorinated biphenyls, organotins, chemical warfare agents and inorganic pollutants belong to the most important environmental pollutants. The need of monitoring leads to the development of independent branch of instrumental analysis - environmental analysis. It is a discrete, and sophisticated branch of instrumental analysis which concerns the treatment of environmental samples from their sampling to receiving the final result of analysis. The fundamental requirement of environmental analysis is for a fast, modern and reliable methodology, especially as the data produced are increasingly drawn upon as the decisive basis for regulatory measures. Consequently, specific conditions need to be fulfilled for the detection of pollutants in trace and ultra-trace quantities, within a short time and with a high degree of precision. To define the analytical process, Skoog and co-workers [1] mention the following steps: selecting method of sampling, obtaining representative samples, preparing laboratory samples, defining replicate samples, dissolving samples,

eliminating interference and measuring features of analyses. The aims of these activities are: 9 making the sample suitable physical parameters, removing interference and transferring the analytes to matrix being compatible with analytical technique; liquid, gas, solid phase and supercritical fluid extraction is usually applied for transferring analytes directly from samples into media being subjected to final instrumental analysis, as well as to liberate analytes trapped on sorbents during preconcentration steps; 9 cleaning-up the analytical samples and analytes enrichment; liquid-liquid partitioning, solid phase extraction, preparative column and thin layer chromatography are usually applied as clean-up and preconcentration techniques, 9 separation of sample components to obtain the chemical individuals; in environmental analyses the partition of analysed mixtures is most often realised by chromatographic methods, 9 detection, identification and quantitation; detectors which are parts of chromatographic apparatus or can co-operate with them in on-line mode are predominantly used. There are many various methods of sampling, sample preparation and analyses, which w a r r a n t correctness of obtained analytical results. Extraction, chemisorption, absorption, adsorption, distillation or freezing are used in them inter alia. Features and applications of these methods are presented in numerous compilations and monographs. In this elaboration we present only these techniques in which phenomena of adsorption are used. They are applied mainly to the sampling of pollutants in fluid, sample preparation and such analytical techniques, which w a r r a n t separation of components of analysed mixture (mainly chromatographic techniques of analyses). In these processes compounds of interest are selectively removed from the bulk sample matrix, preconcentrated, cleaned-up~ separated into individual substances and analysed.

2. SHORT CHARACTERISTIC OF MONITORED S U B S T A N C E S The term environmental pollution means any physical, chemical, or biological change disturbing ecological equilibrium in the environment. It may be a result of random, accidental events, emission of certain pollutants due to activity of nature itself, or h u m a n activities. As a result of the activity of nature, natural pollutants are emitted into atmosphere; h u m a n activity leads to the emission of pollutants called anthropogenic pollutants. Natural and anthropogenic pollutants emitted from a given source are called primary pollutants. A number of primary pollutants can undergo some changes due to reactions with other pollutants, as well as with some components of the environment. In this way, new compounds, often of higher toxicity, can be formed. They are called secondary pollutants. Primary as well as secondary pollutants occur in all of the environment media:

atmosphere, hydrosphere, and soil. The following groups of substances are considered as the most important environmental pollutants: 9 Volatile organic c o m p o u n d s . Volatile organic compounds (VOCs), originating from anthropogenic sources, are the monocyclic aromatic hydrocarbons and the volatile chlorinated hydrocarbons. Both groups of compounds are considered as priority pollutants; they are present in all parts of the environment. Monocyclic aromatic hydrocarbons are mainly emitted by industrial processes and combustion of fossil fuels, while chlorinated hydrocarbons are widely applied as solvents for dry cleaning, as degreasing agents in metal industries or as fumigants [2]. Due to their lipophilic properties, they can be taken up in lipophilic matrices. Uptake of xenobiotic VOCs in plants used for h u m a n nutrition (vegetables and fruits), results in an exposure of man through the food chain, next to a direct exposure to air pollutants through inhalation. VOCs are also the most frequently encountered contaminants at hazardous waste sites. 9 Pesticides comprise a group of compounds that are given great attention in environmental studies. They are introduced into environment due to wilful h u m a n activity; economic production in the cultivation of vegetables and fruits, as well as in agriculture, can not be achieved without pesticides. Pesticides belong to different chemical groups of compounds; the most important of them are: organophosphorous, organochloride, carbamate, triazine compounds and chlorophenoxy acids. With respect to the biological activity they are classified as insecticides, herbicides and fungicides. Well known compounds such as DDT, lindane or aldrin belong to the organochloride group which, in the past, was widely used all over the world. Although their manufacture and application are now largely prohibited, they can still, due to their persistence, be found in the soil, in animals, plants and food products. Pesticides are poisons; some of them or their degradation products also demonstrate carcinogenic potential and teratogenic activity. They are present in all parts of the environment. 9 Polycyclic a r o m a t i c h y d r o c a r b o n s (PAHs) are compounds whose molecules can contain 2-13 aromatic rings arranged in linear, cluster, or angular shapes. They may contain some number of alkyl substituents. PAHs arouse much interest mainly due to their carcinogenic and mutagenic properties. They are widespread environmental contaminants emitted from a variety of sources, including industrial combustion and discharge of fossil fuels, residential heating, or motor vehicle exhaust. In processes of monitoring, PAHs have been measured in a variety of environmental matrices including air, water, soil, sediments and tissue samples. 9 Polycyclic a r o m a t i c heterocycles. In the environment, carbon atoms in PAHs rings can be substituted with oxygen, sulphur, or nitrogen atoms. In this way polycyclic aromatic heterocycles are formed, and they usually occur together with PAHs. The most dangerous of these, polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans, are by-products formed during the

9

9

9

9

manufacture of chlorophenols and related products; other sources include the pulp and paper industry and accidental fires that release polychlorinated biphenyls. Dibenzotiophene and some of its methyl-substituted compounds are persistent residues in sea environment after oil accidents. In the natural environment, polychlorinated thianthrenes and polychlorinated dibenzothiophenes also exist. As with their oxygen analogues, they are hazardous substances. Azaarenes, mainly benz(c)acridine and many of its related compounds, have been shown to exhibit carcinogenic activity. Nitrorelated compounds are mutagenic and carcinogenic. Polycyclic aromatic heterocycles are continually found in many natural and environmental samples. P h e n o l s form a group of aromatic compounds with one or more hydroxyl groups. Phenols and substituted phenols are products of manufacturing processes used in plastics, dyes, drugs, antioxidants, and pesticides industries. They pose the serious danger to the environment, especially when they enter the food chain as water pollutants. Even at very low concentration phenols affect the taste and odour of fishes and drinking water. Because of this, many phenol derivatives (mainly nitrophenols and chlorophenols, which are also poisons) are considered as priority pollutants of the environment. P o l y c h l o r i n a t e d b i p h e n y l s (PCBs) are a group of compounds derived from biphenyl by substituting one to ten hydrogen atoms with chlorine. There are 209 possible PCB configurations. They have extensive application because of high chemical and thermal stability, low or no flammability, low vapour pressure at ambient temperature and high permeability. PCBs are utilised alone or in mixtures as heat-transfer fluids, dielectrics for capacitors and transformers, hydraulic fluids, lubricants, additives in plastics and dyes, etc. PCBs are different in their physical and chemical properties as well as toxic potencies; some of them are inducers of drug-metabolising enzymes also being able to affect various physiological processes such as reproduction, carcinogenesis or embryonic development. O r g a n o t i n s . These compounds have been widely used as biocides incorporated in antifouling paints, and are accumulated by the biota, especially by filtrating organisms. The organotins are much more toxic than inorganic tin. Contamination of the marine environment by organotins has been well documented. Tributyltin is the most often used organotin compound, followed by triphenyltin. In water these substances can be step-wise decomposed to less substituted and down to inorganic tin, absorbed by lipid fraction of organisms or adsorbed onto particulate matter. C h e m i c a l w a r f a r e a g e n t s . The need of the monitoring on the presence of these substances in the environment results not only from the need of the verification of the Chemical Weapon Convention [3] but also because certain chemical warfare agents can be spread in the environment as the old or abandoned chemical warfare agents. Out of this group of compounds organophosphorous (O-ethyl S-2-diisopropylaminoethyl methyl phosphono-

thiolate, O-pinacolyl methylphosphono-fluoridate, etc.) and bis(2-chloroethyl) sulfide ( m u s t a r d gas), tris(2-chloroethyl) amine (nitrogen gas), 10-chloro-5,10dihydrophenarsazine (adamsite) have importance due to their toxicity or persistence in the environment. 9 E x p l o s i v e s . 2,4,6-trinitrotoluene (TNT) is known first of all as an explosive, but it appears t h a t this compound and its degradation products have been found as c o n t a m i n a n t s in w a t e r and soil. TNT and its degradation products have been identified in the blood and urine of the explosives m a n u f a c t u r i n g plants personnel. Because of the m u t a g e n i t y of these compounds, environmental t r e a t m e n t of TNT and its degradation products (2- and 4-monoaminodinitrotoluenes as well as 2,4- and 2,6-diaminonitrotoluenes) is an i m p o r t a n t issue. 9 I n o r g a n i c p o l l u t a n t s . Among inorganic environmental pollutants aerosols, heavy metals, radionuclides and some anions are monitored. Aerosol or particulate m a t t e r refer to any substance, except pure water, t h a t exists as a liquid or solid in the atmosphere under normal conditions and is in microscopic or submicroscopic size. Even non-toxic aerosols are harmful; they can cause eye or t h r o a t irritation, bronchitis or lung damage. Heavy metals (mainly As, Cd, Cr, Cu, Se, Ni, Mo, Hg and Pb) can pose serious t h r e a t s to the h u m a n health even at very low concentrations in air and water. For instance, lead causes damage of brain, mercury affects several areas of the brain, as well as the kidneys and bowels, arsenic causes cancer etc. After pollution of soil they can be incorporated into the food cycle via vegetables or, alternatively, be washed towards surface or underground water. Farming, industrial and u r b a n activities are most often mentioned as pollution sources of heavy metals. The radioactivity in environment originates from both n a t u r a l sources and h u m a n activities. The l a t t e r include operations concerned with the nuclear fuel cycles, from mining to reprocessing, medical uses etc. Radionuclides cause cancer. The common anions, such as cyanides (CN-), halides (Br-, CI-, F-) or the oxy-ions (SO3-, 304-, NO2-or NO3-) are monitored mainly in w a t e r and wastewaters. When listing the most i m p o r t a n t environmental pollutants it is impossible to forget industrial gases such as SO2, NOx, CO2, etc., which are emitted in huge quantities to the atmosphere. First two of t h e m cause respiratory illness and lung damage. They also cause the acid rains which are responsible for corrosion of metals, acidification of soil and surface waters, as well as degradation of forests. NO2 and CO2 are, like as CH4, tropospheric 03 and chlorofluorocarbons, greenhouse gases. These gases absorb in the spectral range where t h e r m a l energy r a d i a t e d from the e a r t h is at a m a x i m u m . All of them, analogically as above mentioned organic and non-organic pollutants m u s t be systematically monitored.

3. A D S O R P T I O N IN S A M P L I N G A N D S A M P L E P R E P A R A T I O N Basic feature which distinguishes environmental analysis is the need of sampling and sample preparation of substances existing in matrix on trace levels. Monitoring of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans can be a good example of such needs. Because of high toxicity the level of quantitation of these substances equals 10 -~2 g/kg; it is also important that these substances usually exist in natural environment in neighbourhood of other organic chlorine compounds whose concentration can be twice or three times higher. So to cope with the demands of environmental analyses such as techniques of sampling, sample preparation and analyses, which have proper ability to separation, high sensitivity, good selectivity, ability to generate reliable identification data should be applied. Adsorption phenomena play an important, if not decisive, role in many of these processes. 3.1. S a m p l i n g The term sampling is used for the description of the process by which a representative fraction of matrix is acquired. In environmental analyses various sampling techniques (and equipment submitting them) are used; adsorption phenomena are usually applied for the sampling of air, surface water and wastewater; in these processes sampling is realised together with the enrichment of analytes. Owing to the adsorption processes compounds of interest are selectively removed from the bulk sample matrix and preconcentrated (an enrichment factor of 103-107 can be usually obtained). There are two main groups of sampling and preconcentration methods of air samples: passive and aspirative (denudatic or dynamic) [4]. The idea of passive method is diffusion or permeation of analytes to the trapped medium surface. Analytes which are present in the nearest surrounding of the enriching device (dosimeter) are transferred due to the molecular diffusion forces towards the semipermeable membrane and are penetrating through it. Phenomena of absorption, chemisorption and adsorption are used in aspirative methods. Passive samplers are suitable for large scale measurements. As they do not require pumping of air during sampling they can be employed at virtually every location. Passive samplers can be sent by mail and stored before and after sampling for periods of several months. On the other hand, passive samplers require at least 24-hour exposure and therefore cannot be used for short-term sampling. Aspirative denudatic method of preconcentration consist in a junction of a forced gas stream flow and diffusive transfer of analytes in the direction of denuder wall which acts as an analyte trap. The advantage of denudating techniques is the possibility of differentiation of so called physical speciation of analytes, it means differentiation between gaseous and aerosol form of preconcentrated substances. Aspirative dynamic enrichment is the oldest method of air sampling. It allows to determine the time weighted average concentration or short term exposure level. Absorption in liquid solutions, freezing out in

cryogenic traps and adsorption belongs to these methods. Adsorption aspirative dynamic methods are used to separate the volatile and non-volatile organic pollutants. The applied techniques differ from each other in volume of sample, shape of sorbent container, and first of all in dissimilarity of used sorbents (usually they are carbonaceous, inorganic or polymeric sorbents). The scheme of the set for sampling and preconcentration of atmospheric air pollutants on adsorbent is presented in Figure la. In Figure lb the crossection of adsorption tube is shown.

1

5

6

7

8 r

Figure la. The set for collecting samples. 1-probe, 2- adsorption tube, 3- filter, 4-capillary tubes 5-vacuum-gauge, 6-flow controller, 7-pressure reducing valve, 8-vacuum pump. Reprinted from [4].

1 2 .~/3 #y5. ~_.#6#'L.~ 8

1

Figure lb. Adsorption tube. 1-plastic caps, 2-fused ends of tube (they are broken before using), 3-glass sorption tube, 4- spring, 5-glass wool, 6-adsorbent layer, 7-polyurethane plug, 8-adsorbent protective layer. Reprinted from [4].

Among carbon sorbents active carbons and carbon molecular sieves with specific surface area between 600 and 1200 me/g, and relatively high adsorption

10 capacity for most organic compounds are used. For specific non-polar analytes graphitized carbon blacks with a small specific surface area are used. Disadvantage of carbon adsorbents is an irreversible adsorption of many analytes and substantial variability of adsorption properties between different batches of the same product. Detailed description of application of carbon adsorbents in analyses of organic environmental pollutants is presented in work of Matiskowa and Skrabakov~ [5]. Among the inorganic sorbents, silica is the most widely used. Chromatographic silica is amorphous, porous solid which can be prepared in a wide range of surface areas and average pore diameters. Variation of solution pH during the acid gelation of sodium silicate yields silica with surface areas varying from about 200 m2/g (pH ~ 10) to 800 m2/g (pH < 4). Silica may be treated as a typical polar adsorbent. The raw material for the production of chromatographic alumina (aluminium oxides) are different aluminium hydroxides, e.g. hydrargillite. Like silica, alumina can be regarded as a typical polar adsorbent, and sample separation order on alumina and silica is generally similar. The presence of carbon-carbon double bonds in a pollutant molecule generally increases adsorption energy on alumina more than on silica. Aromatic hydrocarbons which contain different numbers of aromatic carbon atoms are much better separated on alumina t h a n on silica. Adsorption sites are used for the selective adsorption of u n s a t u r a t e d or polar molecules onto a hydroxylated silica surface. Three distinct site types can be recognised on the alumina surface: acidic or positive field sites, basic or proton acceptor sites and electron acceptor (charge transfer) sites [6]. Each of these is important in the adsorption of certain samples on alumina. Florisil is co-precipitate of silica and magnesia and this is why the retention and separation on its surface is generally intermediate between alumina and silica. Inorganic adsorbents have a high adsorption capability, even to polar and volatile organic compounds. This property is limited in the case of moisture samples (adsorption of water vapours cause the deactivation of adsorption centres and lowers the retention of analytes). Porous polymers and co-polymers are the most universal group of adsorbents used for sampling of air; they are synthesised in the processes of the bead polymerisation. A suitable selection of cross-linking polymers and other polymerisation parameters allows to control polymerisation processes. Therefore it is possible to obtain adsorbents with desirable specific surface area, porosity and polarity (for example Tenax | Porapak | Chromosorb | or XAD| Tenax is a porous polymer based on 2,6-diphenyl-p-phenylene oxide. The high thermal stability and its compatibility with alcohols, amines, amides, acids and bases together with good recovery characteristics make Tenax very suitable as sorbent medium in air and water analysis [7]. Porapak is a series of cross-linked porous polymers, for example divinylbenzene/ethylene glycol dimethacrylate (Porapak N). That sorbent is used for preconcentration of many substances [8]. Porapaks have the following polarity: N>S>P>Q, T>R. Chromosorbs or XAD are produced by copolymerising monofunctional monomers with bifunctional monomers. For

11 instance Chromosorb 102 is a styrene/divinylbenzene copolymer with specific surface area in the range of 300-400 m2/g; the surface is non-polar. Chromosorb 108 is moderately polar acrylic ester resin with the specific surface area between 100 and 200 m2/g. They are also commonly used for air sampling and preconcentration of analytes [9, 10]. Disadvantage of polymeric adsorbents is their sensibility to oxidative action of ozone or chlorine. Among the adsorption methods applied for isolating analytes from liquid matrixes (mainly from water) and for their preconcentration, practical importance has the solid phase extraction (SPE) technique. The idea of this technique consists in retention of analytes from a large sample volume on a small bed of adsorbent (placed in cartridge or shaped in the disk form), and following elution of analytes, with a small volume of solvent. The selection of appropriate parameters of adsorbents and solvents is the basic condition for successful employment of this method. Details on the SPE are presented in chapter 23, vol. 2 of this book. An alternative to the SPE, solvent-free sampling technique is a solid phase microextraction [SPME]. Typically, a fused-silica fibre, which is coated with a thin layer of polymeric stationary phase, is used to extract analytes from fluid (for analysis the retained analytes are thermally desorbed). The application of the SPME for sampling of polycyclic aromatic hydrocarbons [PAHs] from aqueous samples is presented in the work of Yu Liu et al. [11]. The porous layer coatings were prepared by the use of silica particles (5 ~m diameter) bonded with phenyl, Cs, and monomeric or polymeric Cls stationary phases. It was proved that several factors affected the selectivity for extraction of PAHs, including functional group in the bonded phase, and phase type (monomeric or polymeric). The distribution coefficients of PAHs in the porous layer increased with an increasing number of carbon atoms. A greater selectivity towards solute molecular size and shape were obtained using a polymeric Cls porous layer. The effect of solution ionic strength on recovery was also investigated. There are many papers describing the testing of usefulness of various adsorbents for fluid sampling [12, 13]. Adsorption capacity for the defined groups of the analytes, breakthrough capacity and influence of adsorbent bed length, as well as enrichment conditions on these parameters were investigated. The recovery of analytes by their thermal or liquid desorption is an essential element of such investigations.

3.2. Sample preparation Only few analytical methods provide the possibility for examination of samples in their original state, without preliminary preparation. In case of the environmental analysis such examinations are in practice almost impossible. Complexity of environmental samples is the reason why analytical processes are very difficult and usually multistages. Analytes need to be determined at extremely low concentrations over a wide polarity range, and frequently there is little or no information about the analysed sample. This is why the sample

12 preparation is the most important and often the most difficult step of analysis in environmental studies. The experiment described by Falcon et al. [14] can be a very good example of complications of sample preparation process. They developed the procedure for trace enrichment of benzo(a)pyrene in extracts of smoked food products. All steps of this analysis are presented in Figure 2. As it was mentioned above, the

[

Lyophilization

[ 25/15/10 ml hexane (lh)

[

Ultrasound

I

50 ml hexane extract

]

Centrifugation

[

Supernatant

I

Concentration to 5 ml

]

Silica purification

I

I I ! I

Elution 10 ml hexane [

Extraction DMSO

[ 15/10/5 ml DMSO (5 min)

30 ml DMSO extract

I 75 ml water

[

C- 18 purification

[ Elution 5 ml hexane

[

Hexane evaporation

I 1 ml acetonitryle

]

Filtration

[

HPLC

I

I [

Figure 2. Flow-chart summarising treatment sample prior to the HPLC analysis of benzo(a)pyrene. Reprinted from [ 13].

13 transfer of analytes to matrix being compatible with analytical technique, usually by means of liquid, gas or supercritical fluids extraction, is one of the steps of sample preparation process. Unfortunately, in this process very undesirable substances (interferences) penetrate to the matrix. This is why a cleaning-up of analytical samples, connected usually with preconcetration of analytes, is a very essential step of environmental analyses. Among the adsorption methods, preparative column chromatography and thin layer chromatography are commonly used. Aluminium oxide, cellulose powder or microcrystalline cellulose, silica, diatomaceous earth (Kieselguhr), polyamide and Florisil | are employed in column or layer preparation. Nowadays, a large variety of chemicaly bonded stationary phases are applied. Such phases are prepared by anchoring specific organic moieties to inorganic oxides (mainly silica), under defined reaction conditions. Organic moieties can be attached to the silica by mono-, di-, or trifunctional silane reaction. After derivatisation of the silica substrate to yield a bonded phase, a network of socalled structure elements can be distinguished at the silica surface. This includes organic moieties bound to the surface, like cyano-, NH2-, phenyl-, octyl-, or octadecyl groups. The residual silanols, approximately 50% of the originally present silanols, have different properties as they consist of lone, vicinal and geminal groups. Consequently, besides the attached organic ligands, also the residual silanols play an important role in the final properties of the chemically bonded stationary phases. Carlsson and Ostman [15] presented a method for the isolation of polycyclic aromatic nitrogen heterocyclic (PANHs) compounds from complex sample matrix. They are known to be mutagenic and /or carcinogenic. PANHs with a single endocyclic nitrogen heteroatoms can be divided into two classes: acridines (containing a pyridine ring) and carbazoles (containing a pyrrole ring). They were isolated and separated as carbazole and acridine type PANHs with an absolute recovery in the range between 79-98%. The open column chromatography was used as an initial step for isolating a PANH fraction. By applying a normal-phase liquid chromatography using a dimethylaminopropyl silica stationary phase and utilising back-flush technique it was possible to separate the PANHs fraction into two fractions containing acridine type and carbazole type PANHs respectively. The method applied on a sample of solvent refined coal heavy distillate; acridines and carbazoles were identified by gas chromatography (GC). Rimmer's and co-workers work [16] is a good example of application of highresolution gel permeation chromatographic clean-up technique (prior to GC). The method for the determination of phenoxy acid herbicides in vegetable samples was presented. Macerated samples were extracted with acetone, filtered and acidified; the herbicides were then partitioned into dichloromethane, cleaned-up using high-resolution gel permeation chromatography before undergoing rapid and efficient methylation using trimethyl-silyldiazomethane. The resultant methyl esters were than selectively and sensitively analysed by GC/MS

14 technique. The procedure has been applied for grass samples spiked with four phenoxy acid herbicides: 2,4-D, dichlorprop, MCPA and mecroprop. Environmental monitoring is often realised by using the non-direct methods; in such investigations the results of contamination, e.g. presence of pollutants or products of their transformation in food are determined. For example milk; being at one of the highest levels of the tropic chain and due to its lipophilic nature, milk has been usually studied as an indicator of the bioconcentration process of environmentally persistent organic micropollutants. Di Muccio and co-workers [17] developed a rapid procedure that allows a single step selective extraction and clean-up of organophosphate pesticide residues from milk, dispersed on solid matrix diatomaceous material into disposable cartridges by means of light petroleum saturated with acetonitrile and ethanol. Recovery experiments were carried out on homogenised commercial milk spiked with solutions of 24 pesticides. Bernal and co-workers [18] presented a method for determination of vinclozolin (agrochemical fungicide) in honey and bee larvae. LL or SPE extraction techniques were used and two clean-up procedures (chromatography on Florisil or Cls column) were assayed after the solvent extraction. A clean-up method for organochloride compounds in fatty samples based on normal-phase liquid chromatography is described in work of van der Hoff et al. [19]. The use of liquid chromatography column packed with silica enables complete fat/organochloride pesticide separation in total fraction volume of 12 ml and results in a fully automated clean-up procedure. Adsorption phenomena in the soil sampling and sample preparation is rarely applied; it is used mainly to the clean-up of extracts. 4.

TIIE C H R O M A T O G R A P H I C M E T H O D S

The detection and determination of pollutants in complex environmental systems by conventional and biochemical methods is difficult and timeconsuming, and the results are often doubtful. These methods are now being systematically replaced by instrumental analytical methods, among which adsorption procedures play an imporatan role; crucial meaning have the chromatographic methods. The idea of all chromatographic methods is the partition of components of analysed mixture between two phases. One of these phases is stationary; the second is the mobile phase which moves along the stationary phase. Gas, liquid or supercritical fluid can be the mobile phase; the separation techniques which use these phases are called respectively: gas chromatography (GC), liquid chromatography (LC) and supercritical fluid chromatography (SFC). A solid or liquid can be the stationary phase; in the first case it is adsorption chromatography (GSC), in the second one - partitioning chromatography (GLC). If the stationary phase is in a column we call it column chromatography (GC or High Performance Liquid Chromatography - HPLC). In the case when adsorbent

15 is spread on a solid carrier plate in the form of thin layer and attached to it we call it thin-layer chromatography (TLC). In every case the separation is achieved by repeating distribution of analytes between two phases of given chromatographic system. In the column chromatography the compounds are eluted with the mobile phase to a detector (universal or selective), which produces a signal proportional to the amount of a particular substance in this phase. The proper choice of column, injection technique and temperature program will ensure the separation of interesting substances from the background ones. Good separation efficiency is one of the most critical parameters for reliable identification of pollutants by a detector. Pollutants can be identified by means of the absolute or relative retention times; a very useful parameter of identification is also retention index. Quantitation can be realised by internal or external standards. In cases of environmental analyses very frequently compounds cannot be separated from each other. These problems can often be solved by chromatographic technique utilising two or more columns. In multicolumn chromatography the columns may have widely varying measurements and separation characteristics. The columns may be connected either off-line or, nowadays much more often, on-line technique. Volatile or semi-volatile environmental pollutants which are the subject of monitoring are usually analysed by GC. In this technique sensitive and selective detectors such as the electron capture detector (ECD) or the mass spectrometer (MS) are used. They enable identification and quantitation of trace components in complex mixtures. HPLC has been recommended for the analyses of thermally labile, non-volatile and highly polar compounds. Application of high performance adsorbents in TLC and sophisticated equipment (apparatus for automatically spotting and developing chromatograms, scanning densitometry) caused, that present instrumental TLC can compete with the HPLC in terms of analytical efficiency, sensitivity, and precision. Other chromatographic methods such as SCFC and capillary electromigration have been currently developed but for the time being their application in environmental analysis is limited. The studies on applications of chromatographic methods for environmental investigation can be classified on the criteria of goals of experiments. According to this criterion they can be divided into three groups. These ones which refer to the monitoring are represented the most frequently. The reports which can be entitled "behaviour" are relatively numerous too. They refer to behaviour (in term of resolution possibilities) of pollutants in various chromatographic systems. The third group consists of the works in which physical and chemical properties of pollutants, i.e. their mobility, bioaccumulation, biotransformation etc. are examined.

4.1. High Performance Liquid Chromatography High Performance Liquid Chromatography (HPLC) is a form of column liquid chromatography. In this technique the mobile phase is pumped through the

15

packed column at high pressure and therefore HPLC is also called High Pressure Liquid Chromatography. Columns are made of stainless steel tubes 10-, 20 cm long and internal diameters (I.D.) of a few millimetres. Depending on the type of interaction between stationary phase, mobile phase and a sample, following separation mechanisms can take place: adsorption, partition, ion exchange, ionpair and size exclusion. In adsorption liquid chromatography mainly silica and (rarely) aluminium oxide, cellulose and polyamide are used as stationary phases. The separated molecules are reversibly bonded to the solid surface by dipole-dipole interactions. Because the strength of interaction is different for different molecules, residence time at the stationary phase varies for different compounds; thus, separation can be achieved. This technique is used mainly for resolution of polar, non-ionic substances; in environmental analyses it is used occasionally. In the case of liquid- liquid partition chromatography stationary phases (liquids) can coat a support or can be chemically bonded to that support. Distribution mechanism is called partitioning because separation is based on the use of relative solubility differences of the sample in the two phases (in fact the separation is also achieved through the adsorption by non-protected silanol groups). In the normal phase (NP) liquid-liquid partition chromatography, the stationary phase is more polar than the mobile phase, in the reversed phase (RP) liquid-liquid partition chromatography, the mobile phase is more polar than stationary phase. The NP liquid-liquid partition chromatography is used for separation of very polar organic substances, while the RP chromatography (nowadays more popular technique) is used for the non-polar or weakly polar compounds. An example of using the liquid-liquid partition chromatography for the environmental analyses can be the above mentioned work of FalcSn et al. [14]. They used a HPLC-fluorescent detection method for the determination of benzo[a]pyrene in the enriched extract of the smoked food products. It should be stressed that the determination of polycyclic aromatic hydrocarbons (PAHs) by HPLC requires separation columns of high selectivity and efficiency. Reupert and co-workers [20] proposed a method for the separation of PAHs by the application of PAH 16-Plus column under optimal operating conditions. A very good separation of 16 PAHs was obtained (Figure 3). Liquid-liquid partition chromatography is often employed in the analysis of pesticides. The analysis of pesticide residues in the environment is of great current interest due to the possible risks that may arise from the exposure of humans and animals to such agents. From among the latest papers concerning that problem the special issue of Journal of Chromatography "Chromatography and Electrophoresis in Environmental Analysis: Pesticide Residues" is worthy to notice [21]. A good example of taking advantage of liquid-liquid partition HPLC can be the paper by Somsen and co-workers [22]. Precolumn packed with Cls (Polygosil) material for the enrichment of herbicides was combined on-line with the column liquid chromatography and Fourier-transform infrared spectrometry.

17

100

-

80-

9 6040

11

11213 20 O-

I

0

10

I

20

I

I

30

40

Time, min Figure 3. HPLC chromatogram of 10 ~tl PAHs standard (EPA) in CH3CN; concentration of individual substances 90 pg/~tl. Emission signals. Column- Bakerbond PAH 16-Plus; mobile phase H20 - CH3CN (gradient elution). 1-naphthalene, 2-acenaphthene, 3-fluorene, 4-phenanthrene, 5-anthracene, 6-fluoranthene, 7-pyrene, 8-benzo[a] anthracene, 9-chrysene, 10-benzo[e]pyrene, 11-benzo[b]fluoranthene, 12-benzo[k]fluoranthene, 13-benzo[a]pyrene, 14-dibenzo[a,h]anthracene, 15-benzo[g,h,i] perylene, 16-indeno[ 1,2,3oc,d]pyrene. Reprinted from [20].

The isocratic separation was carried out on a 200x2.1 mm I.D. C18 column (Rosil) using acetonitrile-phosphate buffer (40:60) as eluent. The method was based on post-column on-line liquid-liquid extraction and solvent elimination, followed by Fourier-transform infrared spectroscopy. The feasibility of the complete system was demonstrated by analysing river water spiked with triazines and phenylureas at the ~g/1 level. Identifiable spectra were obtained for all analytes. The authors showed that on-line trace enrichment in combination with column liquid chromatography and Fourier-transform infrared detector offers a selective method for the characterisation of moderately polar analytes such as phenylureas and triazines in water samples. In the analysis of pesticides the degradation products also have to be determined because these products will often possess such activities as the parent pesticides. One ought to emphasise that the analysis of pesticide degradation in environmental samples is often difficult to perform due to the different polarities and lower concentrations of the degradation products relative to the parent compounds. Taking into account these difficulties Rollang, Beck-Westermeyer and Hage [23] applied the RP liquid-liquid partition chromatography and the high performance immunoaffinity chromatography for determining the degradation products of the herbicide atrazine in water. A high performance

18 immunoaffinity chromatography column containing anti-triazine antibodies was first used to extract the degradation products of interest from samples, followed by the on-line separation of the retained components on C18 analytical column. The limits of detection for hydroxyatriazine, deethylatriazine and deisopropylatriazine were 20-30 ng/1. Usefulness of this method was demonstrated in the analysis of both river water and groundwater samples. Rapid methods for the isolation and determination of alkylphenols from crude oils with the use of partitioning chromatography were described by Bennett et al. [24]. Determinations were performed by RP liquid-liquid partition HPLC. The authors have proved that the method affords rapid and accurate quantitation of phenol, cresols, dimethylphenols and is suitable for screening large number of samples. They illustrated the methods with two petroleum geochemical examples: determination of the partition coefficients of alkylphenols in oil/brine systems under high pressure and temperature conditions. Leira, Botana and Cela [25] applied an effect of differences in the retention capacity and selectivity of C18 and graphitized carbon column to resolve complex mixtures of non-flavonoid polyphenols (Table 1). Separation of mixture components was accomplished in a single switching operation by using mobile phase of the same composition but a different eluting strength in both separation steps. The elution conditions used in both columns were simplified by means of simulation software in order to obtain multiple fractions. The potential of this technique was realised by resolving a mixture of 38 very similar species (Figure 4).

2.5

@ 2.0

--

1.5

--

1.0

--

AU

V 7 II

Ii.

I TM ,

]

jl !1 fl

iI A it

I

i

II

II

t I

Pl Iii II [I

if II II II

iI II iI li

I I I I

I I-- -- -li

0.5

1 I I i

I

Q

,A ,ll ,,!i, |

i I

!AN_

0

vii I I I I

I I I I

'

I

I

~.~_ ~.

0.0

VI

I

I

I

I

5

10

15

20

Minutes Figure 4. Chromatogram of 38 non-flavonoid polyphenols. Reprinted from [25].

'

19 Table 1 Listing of the non-flavonoid species studied; key numbers match the spectrum labels in the figures, and heart-cut groups the labels in Figure 4 Key Compound Heart-cut number Group 3-Hydroxybenzoic acid III 4-Hydroxybenzoic acid II 2,4 Dihydroxybenzoic acid (~-resorcylic acid) II 2,5 Dihydroxibenzoic acid (gentisic acid) II 2,6 Dihydroxybenzoic acid (7-resorcylic acid) I 6 3,4 Dihydroxybenzenzoic acid (protocatechuic acid) I 7 3,3 Dihydroxybenzoic acid (a-resorcylic acid) I 3,4,5-Trihydroxybenzoic acid (gallic acid) 8 4-Hydroxy-3-methoxybenzoic acid (vanillic acid) III 9 10 3-Hydroxy-4-methoxybenzoic acid (isovanillic acid) III 11 4-Hydroxy-3,5-dimethoxybenzoic acid (syringic acid IV 2,4 Dimethoxybenzoic acid 12 2,6 Dimethoxybenzoic acid IV 13 14 3,4-Dimethoxybenzoic acid V 15 3,5- Dimethoxybenzoic acid VII 2-Hydroxycinnamic acid (o-coumaric acid) VI 16 3-Hydroxycinnamic acid (o-coumaric acid) V 17 18 4-Hydroxycinnamic acid (p-coumaric acid) 19 3,4-Dihydroxycinnamic acid (caffeic acid) III 20 4-Hydroxy-3-methoxycinnamic acid (ferulic acid) V 21 3,5-Dimethoxy-4-hydroxycinnamic acid (sinapic acid) V 22 3,4,5-Trimethoxycinnamic acid VII 23 2-Hydroxybenzaldehyde (salicyl aldehyde) V 24 3- Hydroxybenzaldehyde III 25 4-Hydroxybenzaldehyde III 26 2,5-Dihydroxybenzaldehyde III 27 3,4-Dihydroxybenzaldehyde(protocatechialdehyde) III 28 3,5-Dimethoxy-4-hydroxybenzaldehyde 29 2-Hydroxy-3-methoxybenzaldehyde (o-vanillin) V 30 4-Hydroxy-3-methoxybenzaldehyde (vanillin) IV 31 3-Hydroxy-4-methoxybenzaldehyde (isovanillin) IV 32 2,4-Dimethoxybenzaldehyde 33 3,4-Dimethoxybenzaldehyde (veratraldehyde) V 34 3,5-Dimethoxybenzaldehyde VII 35 3-Methoxybenzaldehyde (m.-anisaldehyde) VI 4- Methoxybenzaldehyde (p-anisaldehyde) VI 36 37 3,4,5-Trimethoxybenzaldehyde VI Chlorogenic acid II 38 Reprinted from [25].

20

Ion-exchange chromatography is a separation procedure in which ions of similar charges are separated by elution from a column packed with a finely divided resin. The stationary phase consists of acidic or basic functional groups bonded to the surface of the polymer matrix. Charged species present in the mobile phase are attracted to appropriate functional groups in the ion exchanger and separated. Mixtures of bases and acids can be separated by this technique. The stationary phases used in ion-pair chromatography are the same as in RP chromatography. Ionic organic compounds (e.g. C7H15803- - heptane sulfonic ion for bases or Bu2N § - tetrabutyl ammonium ion for acids), which form the ion-pair with the analysed sample component of opposite charge, are added to the mobile phase. This ion-pair is a salt, which behaves chromatographically like a non-ionic organic molecule that can be separated by RP chromatography. These methods found only limited application in environmental analysis. D. Krochmal and A. Kalina [26] proved that coupling the ion-exchange chromatography with active or passive sampling of air pollutants gives the possibility of simultaneous determination of sulphur dioxide and nitrogen dioxide. Both gases can be quantitatively absorbed in aqueous solution of triethanolamine and subsequently determined with ion chromatography as sulphates and nitrates. Absorbing solutions were analysed with a single column ion chromatograph equipped with a packed column. Size-exclusion chromatography is a powerful technique applicable for separation of high-molecular-weight pollutants. Packing material for sizeexclusion chromatography consists of a small silica or polymer particles containing network of uniform pores into which solute and solvent molecules can diffuse. In the chromatographic process molecules are effectively trapped in pores and removed by the flowing mobile phase. The compounds with higher molecular weight cannot penetrate into the pores and are retained to a less extend t h a n smaller ones. Some of size-exclusion packing materials are hydrophilic and are used with aqueous mobile phase (gel filtration); others are hydrophobic and are used with non-polar, organic solvents (gel permeation). In environmental analyses size-exclusion chromatography is used for sample clean-up and fractionation. For example, gel permeation chromatography is a standard technique for the isolation of herbicides and fungicides from samples that contain high-molecular-weight interferences, such as solid waste extract, oil or fats [16]. In the case of environmental analyses information about pollutants may be obtained not only from environmental matrix. Kabzifiski [27] proposed a new analytical method for the quantitative determination of metallothioneins protein in h u m a n body fluids and tissues, in order to determine the level of environmental and industrial exposition to heavy metals. For metallothioneins isolation covalent affinity chromatography with thiol-disulfide interchange was applied, which is a modern technique of separation of high affinity, good repeatibly and reproducibility, allowing specific isolation of the thiolproteins and metallothiolproteins. Fundamentals of indirect determination of the contents of metallothioneins protein were worked out throughout estimation of the

21 quantities of metals bound with metallothionein protein and adsorbed on covalent affinity chromatography gel as on the solid-phase extraction support during a separating process.

4.2. Gas c h r o m a t o g r a p h y The term gas chromatography (GC) is used to denote the chromatographic techniques in which the mobile phase is a gas (the carrier gas, mostly N2, H2, Ar or He). The stationary phase is placed in the column; it may be a porous solid (GSC-gas solid chromatography, adsorption chromatography) or viscous liquid (GLC-gas liquid chromatography, partition chromatography). In both cases the transport of components of analysed mixtures (adsorbates, analytes) is realised exclusively in the gas phase, separation - exclusively in the stationary phase. The time of passing of particular analytes through the stationary phase and the frequency of interactions of analytes with this phase are the decisive factors in the separation process. In case of GSC separation occurs because of differences in the adsorption equlibria between the gaseous components of the sample and the solid surface of the stationary phase. In case of GLC, in contrast to HPLC, there is no interaction between the mobile phase and the analyte. Glass, metal (copper, aluminium, stainless steel) or Teflon tubes long 2-3 m and I.D. 2-4 mm are used for making the packed columns to GC. Open tubular columns (capillary columns) are of two basic types: wall coated open tubular (WCOT) and support coated open tubular (SCOT). WCOT is the traditional capillary column made of glass or stainless steel. The liquid phase is applied as a continuous thin film on the inside walls of the tube. The newest WCOT are fused silica open tubular columns (FSOT). This is a very small outer diameter thin wall column that is inherently a straight tube but is flexible enough to be coiled to diameters c.a. 20 cm. FSOT are drawn from specially purified silica that contains minimal amounts of metal oxides. Compared to packed column these capillaries show inert surfaces and higher reproducibility with at last equal separation efficiencies. PLOT (porous layer open tubular) column is similar to a SCOT except for the fact, t h a t the support material is responsible for the separation through the adsorption process. In a PLOT columns there is no coating liquid phases. There are two basic types of packing materials employed in GC. The first type is porous materials used in GSC. The second type are the support materials which are covered with a layer of liquid phase used in GLC. The adsorption properties and selectivity of adsorbents applied in GSC depend first of all on the chemical composition and geometrical structure of their surface. There are several kinds of attractive adsorbate-adsorbent interactions occurring during the separation of mixturecomponents. The most important interactions are: dispersion or London forces, electrostatic forces, induction forces and specific interaction (mainly charge-transfer, which occur between one component with nbonding electrons and showing small ionisation energy and the second component showing high electron affinity). Among dozens of different solids which have been

22 used in adsorption chromatography only few adsorbent types have wide application today. Non-organic adsorbents such as silica, aluminium oxide or Florisil and polymeric adsorbents type of Tenax, Chromosorb or Porapak belong to the porous packings (which do not need to be coated with stationary phases). They can directly be used for adsorption chromatography. The carbonaceous adsorbents are today used in gas adsorption chromatography rather occasionally. In case of GLC the stationary phase is a liquid (often rubber-like), it is immobilised on the surface of a solid support by adsorption or by chemical bonding. Liquid stationary phases are applied both in packed and capillary columns. Packed columns are completely filled with a packing, liquid stationary phases coating an inert support such as diatomite (Kieselguhr), rarely Teflon or glass spheres. Capillary columns do not require a support because their inert walls are coated with the stationary phases. The most important feature of liquid stationary phase is its polarity. The very popular non-polar phases are Squalane (hexamethyltetracosane) and Apolane-87 (24,14-diethyl-19,29-dioctadecylhaptatetracontane). Squalane is used as reference for determination of polarity of other liquid phases in packed column. Apolane-87 is high temperature standard phase used in capillary chromatography. In environmental analyses semipolar phases are used most often. That group of phases is mainly represented by Silicones. Depending on the kind of substituent in oxosilanes chain (dimethyl-, phenyl-, trifluoropropyl-, cyano- etc.), the weak-, medium- and strong polar phases can be prepared. Polygethylenelycol is an example of strong-polar phase. Among specific liquid phases a family of polysiloxane stationary phases (Chirasil), developed for the separation of optical enantiomers, has a great practical importance. Chemicaly bonded phases used in GLC are identical as twere used in HPLC. Barrefors et al. [28] showed that furan and alkylfurans might be selectively analysed on PLOT (aluminium oxide) columns, since other oxygen-containing compounds are normally not eluted. Furan, 2-methylfuran, 3-methylfuran, 2,5-dimethylfuran and the five isomeric C6 alkylfurans, two C7 and three C6-C7 alkenylfurans were determined by adsorbent sampling and GC/MS technique. Separation on PLOT column is presented on Figure 5. Furan elutes after isoprene and cyclopentadiene in the same region as minor pentadienes and branched hexanes. Several minor C6 and C7 furans appear.in the chromatographic range before and after methylbenzene. The purpose of this study was to characterise volatile furans in birdwood smoke which may be of interest with respect to human exposure and as indoor and outdoor wood-smoke tracers in studies of air pollutants. An analytical method to determine highly volatile saturated aldehydes, degradation products of lipid peroxidation, was developed for the capillary GC [29]. The carbonyl compounds were derivatized quantitatively with 2-hydrazinobenzothiazole at room temperature to form their corresponding water-insoluble hydrazones. The derivatives were extracted and detected with high selectivity (Figure 6) by high-resolution GC with nitrogen-phosphorous

23

m

9

C

9

k m

0

I

5

I~I

4~ min

-1

10

20

3O

L.| v,

200~ isothermal

Figure 5. Gas chromatographic separation on aluminium oxide column of prominent furans, alkadienes. Reprinted from [28]. 1 - Methanal 2 - Ethanal 3 - Propanal 4 - Butanal 5 - Isopentanal 6 - Pentanal 7 - Hexanal

34 ISTD,

21

5

6

I

I

I

I

0

5

l0

15

Time (min)

Figure 6. Typical gas chromatogram of the 2-hydrazinobenzothiazole-derivative aldehydes. ISTDinternal standard: 2,4 pentanedione-2-hydrazinobenzothiazole-derivative. Reprinted from [29].

24 detection due to their high nitrogen content. Analyte concentration, pH and type of extraction technique (LLE and SPE) were studied to determine optimal recovery conditions. The method was applied to the analysis of the volatile aldehydes generated during the t h e r m a l oxidation of olive oil at 220~ Begerov and co-workers [30] applied the screen method for the simultaneous d e t e r m i n a t i o n of 28 volatile organic compounds in the indoor and outdoor air at environmental concentrations. Using passive (sorption-diffusive) samplers, the volatile organic compounds were adsorbed onto charcoal during a four-week sampling period and subsequently desorbed with carbon disulphide. The eluate was split via an Y-connector and led onto two capillary columns of different polarity switched in parallel. This dual column configuration provides additional information about the volatile organic compound components and can be obtained for verification purpose. Detection was in both cases performed by connecting each column with a non-destructive electron-capture detector and a flame ionisation detector switched in series. The procedure has been successfully applied in the context of a large field study to measure outdoor air concentration in three areas with different traffic density (Figure 7). It is applicable to indoor air m e a s u r e m e n t s in a similar manner.

a)

b) 7

0.345

1

0.275]--~

4

11

0.305

0.265 13

0.305

2

5

10

3

0.285

17 0.255 1

8

14

1516 /

6

0.245,,

,"

"

r

0.265 0.245

,,, 5

,, ,,,,,,, 10 15

, , , 20

, ,,,,, 25

--"""

.-

18 . . - ' " ""

,, ,,,,,,,,, 30 35

, , 40

Time (min) Figure 7. Typical gas chromatograms of an indoor air sample obtained by flame ionisation detection. (a) more polar column, (b) less polar column. Reprinted from [30]. Lobiafiski et al. [31] studied the potential of the microwave-induced p l a s m a atomic emission detector for capillary GC (GC/AED) as a tool for the specification of organotin compounds in environmental samples. The operational variables are optimised for chromatographic resolution and detection limits. A comprehensive

25 method for the determination of mono-, di-, tri-, and some tetraalkylated organotin compounds in water and sediments by GC/AED was developed. Ionic organotin compounds were extracted as diethyl dithiocarbamates into pentane and, after its evaporation, dissolved in a small volume of octane and derivatized by pentylmagnesium bromide to give the solution suitable for gas chromatography. The phenoxy acids were first introduced as herbicides in the late 1940s. They have found widespread usage in the post-emergence control of annual and perennial broad leafed weeds cereals and grasses. Functioning as synthetic plant growth regulators these herbicides accumulate in the roots and stems of the plants. A method for the determination of phenoxy acid herbicides in vegetation samples is described among other things in the work of Rimmer and co-workers [32]. Macerated samples were extracted with acetone. After filtration and acidification they were introduced into dichloromethane. The herbicides were than cleaned-up using high-resolution gel permeation chromatography. Analysis of PCB normally includes extensive sample clean-up and preconcentration followed by high resolution capillary GC either with electron capture or mass-selective detection. Although both techniques provide the high sensitivity required for PCB investigations, quantitative analysis is complicated by structural variations of detectors-response factors. The quantitative aspect of GC with atomic emission detection (GC/AED) used for the analysis of PCB is presented in work of Bjergaard et al. [33]. Since Cl-responses were almost independent on the PCB structure, individual PCBs were quantitated with an accuracy not better than 10% by utilising a Cl-calibration plot based on a single randomly selected congener (universal calibration). In addition, within 5-10% accuracy, GC/AED enabled estimation of total PCB residue levels and calculation of the percentage by weight of chlorine in mixtures containig PCB. Thus PCB detection limits were higher with GC/AED than GC/ECD. The GC/AED technique was very attractive for PCB and enabled significant simplification of PCB quantitation. A fraction of polycyclic aromatic nitrogen heterocycles (PANHs) from the environmental samples consists of a complex mixture of compounds, due to the large number of isomers. This cause problems with co-eluting peaks when using chromatographic separation technique. Thus, chemical analysis of PANHs requires a group separation of acridine- and carbazole-type compounds in order to facilitate identification, as well as quantitation. The method of solving this problem is presented above [15]. Separated acridine and carbazole groups were analysed by means of GC technique (Figure 8) with using nitrogen-phosphorous detector (NPD). There was no overlap between the PAH and PANH fractions. PAHs were detected with coupled LC/GC-flame ionisation detection. Isomeric selectivity (for the separation of anthracene/phenanthrene) of new monomeric and polymeric liquid crystalline stationary phases, as well as of common non-polar and polar stationary phases in capilary gas chromatography were compared in the work of Kraus and co-workers [34]. The high isomeric

26

"~N

I

13

I

I

15

I

I

17

I

I

19

Figure 8. GC/NPD chromatogram of acridine fraction from solvent refined coal heavy distillate. Reprinted from [15].

selectivity of monomeric liquid crystalline phase [4,4'-biphenylene-bis-(4-n-butyloxybenxoate] was used for the separation of critical pairs of polycyclic aromatic hydrocarbons. Liquid crystal stationary phases are characterized by their chemical structure, as well as by the ordered structure of the mesophase type at given temperature. This is why using liquid crystal stationary phase the solutes retention does not only depend on their vapour pressure and their interaction with the chemical structure of the stationary phase but also on their molecular shape. The isomers having higher length to breadth ratio of the molecule and/or planar molecular shape have an increased retention in the liquid crystalline mesophase. The advantage of using liquid crystals as stationary phases in GC for separation of isomeric PAHs was also demonstrated by Apfel at al. [35] and Hahne [36].

4.3. Thin Layer Chromatography TLC is one of the best known and thoroughly tested method of the analysis of environmental pollutants [37]. The one-dimensional ascending or horizontal techniques have usually been applied for the development of chromatograms in a closed chamber. Multiple and two-dimensional development techniques have seldom been used. Recent, instrumental techniques such as forced flow planar layer chromatography (FFPC), automated multiple development (AMD) and gradient development techniques are becoming more frequently used. Multimodal TLC (TLC/GC, TLC/MS, etc.) have occasionally been applied. Details concerning the new TLC techniques were described in the review of Jork [38].

27 The separation of environmental mixture samples is usually performed on commercial chromatographic plates. The adsorbent is spread as a thin layer onto a suitable solid support (e.g. glass plate, polyester or aluminium sheet). In TLC the same adsorbents are used as in the HPLC technique. Most often separations are performed with silica 60 (pore diameter = 6 nm). Other commercial adsorbents are Kieselguhr, aluminium oxide, cellulose, polyamide and ion exchangers. At present the modified silica (amino -NH2 or cyano -CN and reversed phases such as octyl RP-Cs, octadecyl RP-Cls,) are used. Impregnated silica is mainly applied for analysis of PAHs and heavy metals. H. Engelhardt and P. Engeld [39] showed the possibility of application of TLC technique to the quantitative determination of hydrocarbons in waste water after extraction with n-heptane by means of a micro separator. Chromatographic development was carried out with n-heptane. Quantitation was done by IR spectroscopy or after dyeing with acid violet reagent. Application of acid violet reagent results in no difficulties relating to the proper selection of the hydrocarbon standards in confrontation with IR quantitation technique, and the standard TLC scanner could be used for quantitation. As it was mentioned above, aminoarenes are a class of compounds usually accompanying PAHs in environmental samples. Janoszka an co-workers [40] applied semipreparative TLC to separate these substances from other polar compounds present in sludge extract isolated by SPE technique. Chromatograms were developed on A1203 plates with ternary mixture of organic solvents to distance of 9 cm in DS chamber. The data obtained by using TLC were confirmed by GC/MS analysis. TLC has found frequent extreme application for environmental analysis of pesticides. Rathore and Begun [41] refer to ca. 300 papers in their recapitulation of TLC methods for pesticide residue analysis. Advances in the application of TLC for separation, detection, and qualitative and quantitative determination of pesticides, other agrochemicals, and related compounds are reviewed in Sherma's article [42]. The author showed the possibility of application of TLC for pesticide analysis in different matrices such as food and environmental samples, and for analyses of residue pesticides of various types, including insecticides, herbicides, and fungicides, belonging to different chemical classes. Bt~dek and co-workers [43] proved the possibility of application of thin layer chromatography and SPE to the analyses of pesticide residues in strongly contaminated samples of soil. Modern TLC equipment was used in these investigations. Chromatograms were developed in a normal-phase system by automated multiple development gradient elution. Limitations of detectability by TLC were compensated for by the application of relatively large volumes (by spray-on technique) of analysed solutions on start lines. Quantitative assessment was achieved by UV absorption measurement scanning of the chromatograms by a ,,zig-zag" technique (Figure 9) Recovery and error of the method was estimated - the recovery level was 80% and the R.S.D. was less than 9%.

28

0.4 ' 0.3

1

6

-

5

0.20.]

-

0.0-0.1

'

2

!

25

5'0

75

X (mm)

Figure 9. A chromatogram of pesticide residues: tree scanning at wavelengths 220, 265 and 300 nm. S-absorbance, x-distance of bands. Peaks: 1-oxamyl, 2-pirimicarb, 3-carbaryl, 4-phosalone, 5-malathion, 6-fenitrtion, 7-tetradifon, 8-methoxychlor. Reprinted from [43].

Petrovi6 and co-workers [44] described a study on the retention behaviour of series of thiosemicarbazide derivatives and their 1,2,4-triazoline-3-thiones by normal- and reversed- phase TLC using silica gel, alumina, and C ls- modified silica gel layers, as well as non-aqueous and aqueous eluents. Two types of compounds were studied: thiosemicarbazide derivatives and 1,2,4-triazoline-3thione derivatives. Retention data were discussed in relation to the molecular structure of the solute, and the nature of the stationary phase and eluent. The study can be an example of application of TLC for cognitive purposes. Substituted phenylureas (such as diuron, linuron, metoxuron etc.) are widely used as selective herbicides. Lautie and Stankovic [45] applied an instrumental TLC for determination of six phenylurea herbicides in food. The pesticides were extracted with acetone and purified by SPE. Analysed and standard solutions were spotted to the plates by means of spray-on technique. Chromatograms were developed in 25 steps with the use of gradient elution. Quantitative analysis were based on the measurements of UV (k = 245 nm) absorbance by using a scanner densitometer. The high selectivity, high detectability and reliability of analysis under fairly simple conditions contribute to the effective use of TLC for the detection of chemical warfare agents. It is proper to add that the problem connected with the determination of substances classified as potential warfare agents lie also in the non-military sphere of interest. This concerns, for instance, the uncontrolled spread of toxic substances as a result of industrial break-down or agrotechnical operations, and the generation of poisons, i.e. fluoroacetic acid in plants or phosgene in the troposphere. Application of TLC for military purposes, including analytical procedures for chemical warfare agents, has been recommended by many workers. The review on application of chromatographic methods for

29 chemical warfare agents analysis has been done by Witkiewicz at al. [46]. The same authors, testing a new instrument to the overpressured thin layer chromatography, proved [47] that the instrument can be used (among many other applications) to the analyses of organophosphorous chemical warfare agents (Figure 10). Currently, due to the Chemical Weapon Convention requirements, interest in the TLC technique is increasing; it can be used as "screening method"[48].

A

OPLC

2.000

.

.

GB

~(3iA

VXl

.

.

OD

~

DFP

I. n

0.o00

oooo

Qo

OQ

oOo

oO

Figure. 10. Densitogram (A) and chromatogram (B) obtained from separation of organophosphorous warfare agents by overpressured TLC: adsorbent- silica gel; mobile phase - diisopropyl ether:benzene:tetrahydrofuran:n-hexane (10:7:5:11 v/v). VX1 and VX2 isomers of VX; GA-tabun, GB-sarin, GD-soman and DFP. * indicate impurities. -

Listing the application of adsorption TLC for military purposes, it is impossible to forget about explosives. The most widely found explosives are trinitrotoluene, hexogen, oktagen, tetryl and pentrit. Application of TLC for determination of high explosive residues in water and soil samples is presented in work of Bt~dek and co-workers [49]. In this group of compounds the most dangerous (because of

30 toxicity and wide application) is trinitrotoluene (TNT); in environment it can be reduced by bacteria to aminodinitrotoluenes, dinitrotoluenes and both di- and trinitrobenzenes. The most dangerous are aminodinitrotoluenes; they show severe toxicity and mutagenity [50]. TLC has been also used to solve a variety of analytical problems relating to the identification, separation and determination of inorganic compounds in environmental samples (mainly heavy metals). The review by Mohammad [51] summarises the application of TLC in the analysis of environmental samples for harmful and toxic metals (inorganic and organic substances containing Ag, A1, As, Be, Bi, Cd, Co, Cr, Cu, Fe, Hg, Mn, Ni, Sb, Sn, and Zn). In most cases, extraction of analytes from the matrix, clean-up of the extracts and concentration of the analytes precede TLC analysis. The author showed that spectrophotometry, titrimetry, atomic absorption spectroscopy, densitometry, fluorimetry and solvent extraction techniques have been combined with TLC for sensitive identification, quantification, and selective separation of toxic heavy metals present in rivers and sea water, waste water, sludge, aquatic plants, cosmic dust, air and airborne dust.

4.4. Supercritical Fluid Chromatography Supercritical fluid chromatography (SFC) combines some characteristics of both gas and liquid chromatography. In recent years the interest in the use of SFC as a separation technique has been increasing rapidly because of the unique properties of supercritical fluid; its higher diffusivity and lower viscosity enable analysis to be 3-10 times faster than HPLC. On the other hand, it has relatively similar density to liquid and viscosity comparable to gas, so SFC can be used to analyse a wide range of compounds, particularly those that are thermally labile, non-volatile and of high molecular mass, that cannot be satisfactorily analysed by GC. The most widely used supercritical fluid is CO2, however, for analyses of polar and high-molecular-mass solutes, polar modifiers such as methanol must be incorporated to increase the solvent polarity. Supercritical CO2 is an ideal solvent for preparing samples for GC, LC, SFC and other analyses. CO2 has easily accessible critical point and its solvating power can be controlled to match a wide range of hazardous organic solvents. Yet even as it assumes liquid-like solvent properties, supercritical CO2 retains gas-like viscosity and penetrates solid samples quickly. E. Pocurull and co-workers [52] studied the possibility of application of SFC with diode array detection system to determine phenol and nitrophenol in water. Several columns and the influence of chromatographic conditions (temperature, pressure, flow-rate and adding methanol in the mobile phase) were studied in order to separate the compounds. To decrease the detection limits of the method, SPE on-line coupled to SFC was tested. Tetrabutylammonium bromide was used as ion-pair reagent in the extraction process to increase the breakthrough volumes. The separation of five phenolic compounds, in the time period less than

31 6 min., with good resolution for all compounds was achieved. The performance of the method was checked for tap and river water samples. 5. MISCELLANEOUS METHODS In recent years, capillary electrophoresis (CE) has developed into an versatile and powerful technique. It was applied for separation of a wide variety of compounds, ranging in size from small ions to large biomolecules. The main separation modes of CE are: capillary zone electrophoresis (CZE), capillary gel electrophoresis (CGE), micellar electrokinetic capillary chromatography (MECC), capillary electrochromatography (CEC), capillary isoelectric focusing (CIEF) and capillary isotachophoresis (CITP). Various modes of separation, high resolving power and small sample requirement (CE is know as a nanoscale technique) have made possible a wide range of applications. CE is an analytical tool, possessing some typical features of chromatographic techniques and some features of traditional slab gel electrophoresis [53]. For example, micellar electrokinetic capillary chromatography couples both the electrophoresis and chromatographic partitioning element for simultaneous separation of charged and neutral compounds. In Figure 11 the versatility and

6 7

0.015

8

0.010

23

0.005

-

0.000

-

9

10

J 0

'

I

4

'

I

8

'

I

12

Time (min) Figure 11. MECC separation of eleven herbicides in distilled water at the mg/l level. 1-tribenuron, 2-chlorsulfuron, 3-metsulfuron, 4-paraquat, 5-simazine, 6-atrazine, 7-1inuron, 8-terbuthylazine, 9-alachlor, 10-metolachlor, 11-trifluralin. Reprinted from [53].

32 the efficiency of CE in the separation of herbicides is shown (conditions of experiment: herbicide concentration range 1-2 mg/1, capillary 500 mm x 7,5 mm I.D.; operating voltage 25 kV at 30~ UV detection at 214 nm; separation buffer: 30 mM sodium borate and 30 mM sodium dodecyl sulphate, pH 8.0). During application of CE for monitoring analyses and environmental studies of pollutants, particular attention has to be devoted to the optimisation of the separation process in order to obtain the best selectivity in a complex matrix, where many potential compounds may interfere. In order to optimise CE separation, several parameters have to be taken into account, such as electrolyte buffer composition, capillary dimensions, capillary temperature, applied voltage and mode of injection and detection. J un Liu et al.. [54] employed in CE a palladium particle-modified carbon microdisk electrode for the simultaneous detection of hydrazine, methylhydrazine and isoniazid. Analytes were separated by CZE and MECC techniques with the Pd-modified microdisk electrode; it had high catalytic activity for hydrazine and exhibited good reproducibility and stability. Quantitative determination of total molecular concentration of bioaccumulatable organic micropollutants in water, using Cls empore disk is presented in van Loon and co-workers paper [55]. Chemical group parameters such as dissolved organic carbon, absorbable organohalogen and chemical oxygen demand determination are routinely used for water quality monitoring. These parameters gave information about the degree of anthropogenic contamination and potential aquatic toxicity of water systems. A new, highly sensitive and quantitative group parameter to determine total molar concentrations of organic micropollutants that can bioaccumulate in the lipid phase of aquatic organisms from effluents, surface water and drinking water has been developed in this work. Cls empore disk was used as a surrogate lipid phase. The partition between water and Cls empore disk was employed to simulate the bioaccumulation process. After partition extraction of the water sample, the empore disk was extracted with cyclohexane and total molar concentration of organic micropollutants was determined. Vapour pressure osmometry and GC/MS were used in these investigations. Measurements of pesticides mobility in soil by application of TLC are presented in work of Camazano et al. [56]. The effect of soil improvement by using urban compost, agricultural organic amendments and surfactants on the mobility of two sparingly-soluble pesticides (diazinon and linuron) was studied. The modifications in Rf values due to the addition of the amendments were similar for both pesticides. No significant correlation was found between the Rf values and the content of total organic carbon in the amended soils. Authors demonstrated that not only the organic carbon content of amended soils but also the amendment nature, especially their contents in a soluble fraction play a very important role in the pesticide mobility. The surfactants gave rise to important alterations in pesticide mobility. The mobility of pesticides changed from being immobile in the soil sample modified with tetradecyltrimethylammonium

33 bromide to being slightly mobile in natural soil and to being mobile in the soil sample amended with sodium dodecyl sulphate. In recent years repid development of small, simple and portable devices for detection and quantitative determination of pollutants in air, water, soil and biological materials is observed. These devices are composed of detecting system and electronic part with displaying possibilities. Adsorption of analytes at the detector surface results in changing of its physical, chemical or biological properties which can be easily converted to the changes of electric signals. In o p t i c a l s e n s o r s a change of the factor of light refraction or effect of fluorescent quenching because of adsorption of analytes are usually applied. E l e c t r o c h e m i c a l s e n s o r s are based on the changes of cell electromotance due to adsorption of analytes at the surface of ion selective electrodes. In the construction of piezoelectric sensors quartz resonators are used. They are called quartz microbalances or adsorption detectors because changes of crystal frequencies are caused due the adsorption of analysed pollutants. To increase the sensitivity and selectivity of such sensors, special liquid coating materials are used. Piezoelectric sensors for SO2, NH3, H2S, Hg, nitrocompounds, chlorinated hydrocarbons and other environmental pollutants are commonly used. The principle of s u r f a c e a c o u s t i c w a v e s e n s o r s action is similar. CONCLUSIONS Environmental analysis is a broad branch of analytical chemistry in which different analytical techniques are applied. Besides chromatographic methods described above, the spectrophotometric, spectrometric and electrochemical methods are of great importance. In this work we focused our attention only on processes and methods in which the adsorption phenomena play an important role. Quoted examples of application of adsorption phenomena in sampling, preconcentration, clean-up and analyses processes are a small part of the huge collection of works devoted to the discussed problems. These examples confirm that adsorption phenomena present one of the most important and promising tools for characterisation, identification and determination of trace pollutants in various environmental samples.

REFERENCES .

.

D. A. Skoog, D. M. West and F. J. Holler, Fundamentals of Analytical Chemistry, A. Harcourt Brace Colleg. Publisher, Six Edition, Orlando 1992, 3. R. Keymeulen and H. Van Langenhove, Application of gas chromatography/mass spectrometry for the determination of volatile organic pollutants in plants. II OgSlnopolskie Sympozjum Chromatograficzne, B. Buszewski (ed.), Toruh 1995, 125.

34 3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

27.

Convention on the Prohibition of the Development, Production, Stockpiling and Use of Chemical Weapons and Their Destruction, United Nations, 1993. J. Namie~nik, J. Lukasiak and Z. JamrSgiewicz, Pobieranie prSbek ~rodowiskowych do analizy, PWN, Warszawa, 1995, 158. E. Matiskowa and S. Skarbakov~, J. Chromatogr., 707 (1995) 145. L.R. Snyder, Principles of Adsorption Chromatography, The Separation of Nonionic Organic Compounds, E. Arnold (Publishers) LTD., London, 1968, 166. P. Ciccioli, G. Bertoni, E. Brancalegeli and F. Bruner, J. Chromatogr., 126 (1976) 757. J.W. Russel and L. Shadoff, J. Chromatogr., 134 (1977) 375. R.L. Harris, R. J. Huggett and H. D. Slone, Anal. Chem., 52 (1980) 52. P. Van Rossum and R. S. Webb, J. Chromatogr., 150 (1978) 381. Yu Liu, M. L. Lee, K. J. Hageman, Yu Yang and S. B. Hawthorne, Anal. Chem., 69 (1997)5001. K. Figge, W. Rabel and A. Wieck, Fresenius Z. Anal. Chem., 327 (1987) 261. J. Namie~nik, L. Torres and J. Mathieu, Int. J. Environ. Anal. Chem., 22 (1985) 169. M. S. G. Falc6n, S. G. Amigo, M. A. L. Yusty, M. J. L6pez de Alda Villaiz~n and J. S. Lozano, J. Chromatogr., 753 (1996) 207. H. Carlsson and C. Ostman, J. Chromatogr., 790 (1997) 73. D. A. Rimmer, P. D. Johnson and R. H. Brown, J. Chromatogr., 755 (1996) 245. A. Di Muccio, P. Pelosi, I. Camoni, D. A. Barbini. R. Dommarco, T. Generali and A. Ausil, J. Chromatogr., 754 (1996) 497. J. L. Bernal, M. J. del Nozal, J. M. Rivera, J. J. Jim~nez and J. Atienza, J. Chromatogr., 754 (1996) 507. G. R. van der Hoff, A. C. van Beuzekom, U. A. Th. Brinkman, R. A. Baumann and P. Van Zoonen, J. Chromatogr., 754 (1996) 487. R. Reupert, G. Brausen and S. Witkowski, Chem. Anal. (Warsaw), 38 (1993) 463. Chromatography and Electrophoresis in Environmental Analysis: Pesticide Residues, J. Chromatogr., 754 (1996). G. W. Somsen, I. Jagat, C. Gooijer, N. H. Velthorst, U. A. Brinkman and T. Visser, J. Chromatogr., 756 (1996) 145. J. G. Rollang, M. Beck-Westermeyer and D. S. Hage, Anal. Chem., 68 (1996) 3631. B. Bennett, B. F. J. Bowler and S. R. Larter, Anal. Chem., 68 (1996) 3697. E. Leira, A. Botana and R. Cela, J. Chromatogr., 724 (1996) 67. D. Krochmal and A. Kalina, Application of ion chromatography for determination of air pollutions. II OgSlnopolskie Sympozjum Chromatograficzne, B. Buszewski (ed.), Toru~ 1995, 57. A. K. M. Kabzi~ski, Polish J. Environ. Studies, 6 (1997) 61.

35 28. G. Barrefors, S. Bj0rkqvist, O. Ramn~is and G. Petersson, J. Chromatogr., 753 (1996) 151. 29. E. E. Stashenko, J. W. Wong, J. R. Martinez, A. Mateus and T. Shibamoto, J. Chromatogr., 752 (1996) 209. 30. J. Begerow, E. Jermann, T. Keles, T. Koch and L. Dunemann, J. Chromatogr., 749 (1996) 181. 31. R. Lobiahski, W. M. R. Dirkx, M. Ceulemans and F. C. Adams, Anal. Chem., 64 (1992) 159. 32. A. Rimmer, P. D. Johnson and R. H. Brown, J. Chromatogr., 755 (1996) 245. 33. S. Pedersen-Bjergaard, S. I. Semb, E. M. Brevik and T. Greibrokk, J. Chromatogr., 723 (1996) 337. 34. A. Kraus, G. Kraus, R. Kubinec, I. Ostrovsk~ and L. Soj~k, Chem. Anal. (Warsaw), 42 (1997) 497. 35. A. Apfel, H. Finckelmann, G. M. Janini, R. J. Laub, B. L(ihmann, A. Price, W. L. Roberts, T. J. Shaw and C. A. Smith, Anal Chem., 57 (1985) 651. 36. F. Hahne, Disertation, Martin-Luther-Universit~it, Halle 1990. 37. J. Bt~dek, Thin Layer Chromatography in Environmental Analysis, in: Practical Thin Layer Chromatography, B. Fried and J. Sherma (eds.), CRC Press, Boca Raton, 1996, 153. 38. H. Jork, J. Planar Chromatogr., 5 (1992) 4. 39. H. Engelhardt and P. Engeld, J. Planar Chromatogr., 5 (1997) 336. 40. B. Janoszka, K. Trypieh and D. Bodzek, J. Planar Chromatogr., 6 (1996) 450. 41. H. S. Rathore and T. Begum, J. Chromatogr., 643 (1993) 271. 42. J. Sherma, J. Planar Chromatogr., 2 (1997) 80. 43. J. Bt~dek, A. Rostkowski and M. Miszczak, J. Chromatogr., 754 (1996), 273. 44. S. M. PetroviS, E. LonSar, N. U. PrisiS-Janjid and M. LazareviS, J. Planar Chromatogr., 1 (1997) 26. 45. J. P. Lautie and V. Stankovic, J. Planar Chromatogr., 2 (1996) 113. 46. Z. Witkiewicz, M. Mazurek and J. Szulc, J. Chromatogr. 503 (1990) 293. 47. Z. Witkiewicz, M. Mazurek and J. Bt~dek, J. Planar Chromatogr., 6 (1993) 407. 48. J. Bt~dek, S. Neffe and A. Rostkowski, Estimation of the Possibilities of Application of Thin-Layer Chromatography for Chemical Weapon Convention, NATO SICA Meeting in Baltimore, May 1997. 49. J. Bt~dek, A. Paplihski, S. Neffe and A. Rostkowski, Chem. Anal. (Warsaw), in press. 50. N. G. McCornick, F. E. Feeherry and H. S. Levenson, Appl. Environ. Microbiol., 31 (1976) 949. 51. A. Mohammad, J. Planar Chromatogr., 1 (1997) 48. 52. E. Pocurull, R. M. Marc~, F. Borrull, J. L. Bernal, L. Toribio and M. L. Serna, J.Chromatogr., 755 (1996) 67. 53. G. Dinelli, A. Vicari and P. Catizone, J. Chromatogr., 733 (1996) 337. 54. Jun Liu, W. Zhou, T. You, F. Li, E. Wang and S. Dong, Anal. Chem., 68 (1996) 3353.

35 55. W. M. G. M. van Loon, F. G. Wijnker, M. E. Verwoerd and J. L. M. Hermes, Anal. Chem., 68 (1996) 2916. 56. M. Sanches-Camazano, M. J. Sanches-Martin, E. Poweda and E. Iglesias-Jimdnez, J. Chromatogr., 754 (1996) 279.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

37

F u n d a m e n t a l s of s o l i d p h a s e e x t r a c t i o n a n d i t s a p p l i c a t i o n to environmental analyses M. E. Raisglid and M. F. Burke Department of Chemistry, University of Arizona, Tucson, AZ, USA 1. S A M P L E P R E P A R A T I O N O V E R V I E W

It has been the convention to extract an analyte as a means of sample preparation using a liquid-liquid extraction technique. In liquid-liquid extraction, the analytes are partitioned between to immiscible solvents. The partitioning is based on a difference in solubility of the analyte in each of the phases, and is a very non-selective process. Efficient separation is dependent on obtaining intimate contact between the two immiscible phases. One limitation of the liquidliquid extraction method is that for intimate contact to be achieved, the phases must be shaken vigorously, which often results in an emulsion that can be difficult to break. An additional limitation is that large sample and extraction volumes are required. After the extraction step, further concentration of the extraction solvent is often necessary in order to obtain detectable levels of the analyte. LLE can be a very time consuming process. The type of manipulations that are required to perform a liquid-liquid extractions lend this technique to being a serial and largely manual process. A different approach to sample preparation is to exploit the interactions at the liquid - solid interface. This has come to be known as solid phase extraction, or SPE. This technique has found its roots in liquid-solid extraction (LSE) which was classically carried out by adding the sorbent to a vessel containing the analytes in a liquid phase, and shaking for a controlled length of time. After the distribution of the analytes reached equilibrium, the phases were separated by either filtering or decantation. The analytes could then be desorbed using a suitable solvent [1]. 2. SOLID P H A S E E X T R A C T I O N O V E R V I E W

There are a wide variety of applications for solid phase extraction as a sample preparation technique, including many environmental applications such as the determination of pesticides and other contaminants in drinking water, waste water, soils and sludge. SPE can be defined as the separation or removal of an analyte or analytes from a mixture of compounds by selectively partitioning the compounds between a

38

stationary solid phase (sorbent) and a mobile liquid phase (solvent). This process is dependent first on our ability to extract the analytes from their matrix and retain them on a solid surface. Once the analyte is extracted from the matrix, it can be eluted from the sorbent using a selective solvent. In solid phase extraction, a "liquid" phase is immobilized onto a solid surface. The analyte is partitioned between the liquid phase (sample matrix) and the modified solid surface. There are a variety of species that can be immobilized on the solid surface, ranging from hydrophobic, to polar, to ionizable species, and a particular solid phase can be chosen to retain analytes having a specific functional group. Recoveries are often improved using solid phase extraction versus LLE. In a review article, Chladek and Marano compared a solid phase extraction procedure with the standard EPA method 625 liquid -liquid extraction and obtained an average recovery with C8 that was 20% higher t h a n that obtained with the standard method [2]. The emulsion problem which often plagues liquid-liquid extractions is eliminated, since the extraction no longer requires the mixing of two liquid phases. Intimate contact is obtained between the two phases by utilization of a solid phase that is made up of small particles packed into a bed. Typically the solid phase consists of particles that are nominally 50 microns in diameter. The particles are very porous, which provides the large surface area. A sample containing the analytes of interest is then passed through the packed bed of particles. The small packed bed provides an inherent concentration step since the sorbent has a relatively small void volume (-1.2 mL/g). The solvent volume required to elute the analytes from the solid phase is correspondingly small (typically two bed volumes), and so the concentration step which is often required for liquid-liquid extractions is eliminated or greatly reduced for solid phase extractions. Solid phase extraction, unlike liquid-liquid extraction is also very amenable to batch processing and to automation. One of the important functions that solid phase extraction can serve is the isolation and purification of an analyte. Analytes are often present in a matrix that contains a significant number of interfering species. If a sample is analyzed directly (without an extraction step), the interfering species may generate a background that makes it difficult to quantify the analyte of interest (Figure 1). Since SPE is a selective process, the interfering species can be either passed through the column during the loading step while the analyte of interest is retained, removed during a rinse step, or retained on the column during the elution of the analyte. Another important function that SPE can serve is the trace enrichment of an analyte. Often the limiting factor in quantifying a compound is that the species is at or below the detection limit. Using solid phase extraction, the analyte can be extracted from a large volume onto a small sorbent mass, and eluted using a small volume of solvent. Concentration factors as great as 500 to 1000 fold can be achieved. In solid phase extraction, both the extraction and elution steps are impacted by interactions between the analytes and the sorbent, the analytes and the

39

, > _

q

I

I

'-.._ [

Figure 1. Removal of background interferences using SPE.

matrix, and finally the matrix and the sorbent. Therefore, the three components to a solid phase extraction that need to be considered are the sample matrix, the analyte and the sorbent. In order to retain the analytes efficiently, it is necessary to optimize the interactions between the analytes and the sorbent while minimizing the interaction between the analytes and the matrix, as well as the sorbent and the matrix. During elution, it is important to optimize interactions between the matrix and sorbent, and matrix and analyte, while minimizing interactions between the sorbent and analyte. It is instructive to compare solid phase extraction to elution chromatography. In liquid chromatography, the volume of solvent required for the elution of a species (Vr), is equal to the volume of the mobile phase plus the product of a constant and the volume of the stationery phase (Vr = V m + KVs). Under the conditions of elution chromatography, it is desirable for the constant, K, to be greater than 0.2 so that separation between species can be obtained, but less than 200, so that the species are eluted within a reasonable amount of time. In extraction chromatography (SPE), ideally the species would behave digitally, stopping at the top of the column upon being loaded (K> 1000), and traveling with the solvent front during the elution (K<0.001). Typically the analyte is eluted within 10 bed volumes (void volume of the sorbent).

2.1. P h y s i c a l c h a r a c t e r i s t i c s of the s o r b e n t One of the challenges in investigating interactions at the liquid-solid interface is in characterizing the surface of the sorbent. Figure 2a shows a sample of silica material that was analyzed using an electron microscope. The particle sizes can be seen to be very uniform at a magnification of x60. At the greater magnification (x600) shown in Figure 2b, the particles can be seen to be very irregularly shaped, which helps with the chromatography.

40

Figure 2a. 50 p silica magnified x60.

Figure 2b. 50 ~t silica magnified x600.

On a n a n o m e t e r scale, the silica is found to have a very porous structure, which increases the effective surface area of the material (Figure 3). If the surface area of the silica particles is calculated strictly based on the diameter of an average particle, for 50 micron particles, the result would be 0.1 meters 2 per gram. Due to the porosity of the silica, the effective surface area is 400 to 550 meters 2 per gram. The porous n a t u r e of silica used in solid phase extraction plays an i m p o r t a n t role in the extraction as well as in the elution of analytes. The porosity of the silica provides the analyte with sufficient surface area to interact with the sorbent during analyte retention. The pores must be of adequate size as to be accessible to the bonding agent, solute, and analyte. It has been shown t h a t as the pore size increases, analyte retention increases, until the pore size gets sufficiently large as to diminish the available surface area [3].

j i J

:~

[

Figure 3. Porous nature of silica sorbent.

\

41 2.2. C h e m i c a l c h a r a c t e r i s t i c s o f t h e s o r b e n t On an unmodified silica surface, the sorbent is very heterogeneous in nature. Some of the variations on the surface include the presence of free silanols, geminal silanols, siloxanes, bound and reactive silanols, and water absorbed onto the surface (Figure 4). Under various conditions of pH, silanols may be largely protonated, protonated and ionized to approximately an equal extent, or be largely ionized.

I

I

9---S i'--O Free SUanol

/H

~

/OH

"----]i--OH .... O\H

/Si\oH

Adsorbed Water

Geminal Silanol

I

--~Si\

%,o -'~Ii

I

--"~P'-O k

o

I Dehydrated Oxide or Siloxa ne

,,H

--SilO\

.

Bound and Reactive Silanols

Figure 4. Heterogeneous nature of silica surface.

Silica may be chemically modified to create a covalently bonded surface capable of offering interactions to a wide variety of analytes (Figure 5). Depending on the bonding chemistry used, the modification of the surface may be monomeric or polymeric. In monomeric type bonding (Figure 6), a compound with the general formula R~SiC1 (such as octadecyldimethylchlorosilane) is reacted with the silica surface, and HC1 is eliminated. This modified surface now has hydrophobic character, and can retain species that are hydrophobic in

\

~/

Figure 5. Chemically modified silica surface.

42

R R~I IR

OH R 3SiCI

OH

I

+

I

,,Si x ,,Si x O

O

o I /Six

---O

O

OH I /Si~ O

O

Figure 6. Monomeric bonding chemistry on silica surface.

nature. Since geometry will not allow every silanol to be modified, a significant number of the silanol surface variations mentioned earlier will still be present, and so the surface is still heterogeneous in nature. In the polymeric modification of the silica surface, a compound having the formula RSiC13, such as octadecyltrichlorosilane, is reacted in the presence of a trace amount of water. The chlorine group is hydrolyzed, allowing the product to polymerize. The polymer is then brought into contact with the silica surface, forming covalent bonds (Figure 7). This results in a surface having a greater coverage of the modifying species than in the monomeric phase, although once again, all of the variations discussed earlier are still present [3,4]. To further deactivate silanol groups on the silica, the surface may be "endcapped". Endcapping is done as a final step using trimethylchlorosilane, which results in "capping off' some of the remaining silanol groups with a trimethylsilyl group. This process reduces, but does not eliminate, all of the silanols on the silica surface. There is a wide variety of different sorbents that are manufactured using these chemistries, including hydrophobic phases ranging from C2 to C18, carboxylic acid (CBA), propylamino (NH2), propylsulfonic acid (SCX), quaternary amine (SAX), propylcyano (CN) and diol (2OH) [9]. The most hydrophobic of the sorbents that are available, and one that is not silica based, is polystyrene divinylbenzene (Figure 8). This material is made by cross-linking oligomers of styrene with divinylbenzene. The sorbent has a very high capacity, with an effective surface area of approximately 1100 m 2 per gram. It is subsequently hydroxylated to provide a water wettable surface. RSiCl3

R

I

H20 ~

R

I

RSi(OH)3

-~

R

R

JOH

~H

I I --O---~i--O---Si--O---

R

I

HO--Si--O--Si--O--Si---OH / H ~/ I .O u OH .O I / I Si $1 $1

?. ?. ?. Si

= o/

\0 /

Si

Si "0 /

o/ " o / " o / " o

Figure 7. Polymeric bonding chemistry on silica surface.

"0

43

OH

{~ ~~~~--OH

pH

~O

H~~

"Ct'I, Ct.I2

C~..c~(C=/'/,e OH--~~ O// C,~..c~.CH.CH2.

OH

C-CH2 CH'CH2"CH-GH2" C C"H2. CH-CH2"

HO Figure 8. Chemical structure of polystyrene divinylbenzene.

2.3. Sorbentselection The sorbent chosen for the extraction is dependent on the types of interactions that can be exploited to remove an analyte from a particular matrix. If an analyte has sufficient hydrophobic character, a sorbent that has been modified to have a hydrophobic surface can be utilized. Examples of such sorbents are C18, C8, C6, C4, C2, CH (cyclohexane) and PH (phenyl). For analytes with limited hydrophobic character, the polystyrene divinylbenzene is most likely the best choice. In addition to hydrophobic interactions, it is possible to capitalize on ionic interactions between an analyte and the sorbent in an aqueous matrix. If the analyte is ionized, a sorbent that is modified with an ion exchanger can be utilized. These ion exchangers may be permanently charged (commonly referred to as a strong ion exchanger or ionizable (weak ion exchanger). Examples of some of these are sorbents are. Cation exchangers: Propyl sulfonic (PRS)

S03 with propyl linkage

Benzyl sulfonic (SCX)

SO3 with benzyl linkage

Carboxylic acid (CBA)

Carboxylic acid group with propyl linkage

P e r m a n e n t negative charge P e r m a n e n t negative charge pKa approximately 4.8

44 Anion exchangers: Q u a t e r n a r y amine (SAX) Amino propyl (NH2)

NH(CH3)3 group with propyl linkage NH2 group with propyl linkage

P e r m a n e n t positive charge pKa approximately 7.8

Although the primary interaction between an analyte and the surface may be either polar, ionic or hydrophobic, it has been demonstrated that most analytes are retained through multiple interactions, and that retention mechanisms may be largely dependent on pH. As an example, aniline at a pH of six may be retained on a C8 column through ionic interactions with the silica surface, as well as secondary hydrophobic interactions with immobilized C8. Although polar interactions are exploited in normal phase chromatography (non-aqueous matrix), they may also contribute as secondary interactions in reverse phase (aqueous matrix). Even among the hydrophobic phases, there are significant differences in polarity and retention characteristics. For example, the C2 surface is quite polar. It has been shown that there is a significant amount of water associated with a C2 surface. Since C2 chains are short, water has easy access to the surface. A phenyl phase is slightly more hydrophobic than C2, but the surface can still be easily accessed by water molecules. Although there is still some water associated with a C8 surface, it is much less polar than a C2 surface. The least polar of the modified silica hydrophobic phases is C18, where the long C18 chains effectively serve to water-proof the surface. In an aqueous solution containing organic solvent, much more of the organic solvent the will associate with the C18 chains. 2.4. S P E f o r m a t The most common format in which the solid phase extraction sorbent is utilized is a polypropylene cartridge with a lure tip, containing sorbent particles packed into a bed. The sorbent bed is retained on the top and bottom by porous polyethylene frits (Figure 9).

Reservoir

Sorbent Bed Frits Luer Tip

Figure 9. Cartridge format for solid phase extraction.

45 Typical column sizes range from 1 to 6 mL in volume, although barrel sizes up to 70 mL are available. Commonly used sorbent bed masses range from 100 to 1000 mg, however there in an increasing trend toward minimizing bed size, and 1 mL columns containing 10, 25 and 50 mg of sorbent are now being used. A second format for SPE is known as the disk, where Teflon or glass fiber is i m p r e g n a t e d with small silica particles (- 8 microns). The disks range in depth from 0.8-2 mm, and in diameter from 0.5 to 90 mm. The smaller disks are retained in a syringe barrel, while the larger disks are use in conjunction with a vacuum filtering setup. The disk format generally has a lower capacity, and requires larger solvent volumes for the elution.

2.5. Basic steps of an SPE p r o c e d u r e A typical SPE procedure involves six steps, and there are a variety of factors t h a t can have an impact on analyte recovery in each one of these steps. These steps are as follows, and are described in detail in the following sections. 1. 2. 3. 4. 5. 6.

Sample P r e - t r e a t m e n t Column Solvation Column Equilibration Sample Loading Interference Removal Analyte Elution

2.5.1. S a m p l e p r e - t r e a t m e n t The first step in an SPE procedure is p r e - t r e a t m e n t of the sample. Samples are p r e t r e a t e d in order to maximize the interactions between the analyte and the sorbent, while minimizing the interactions between the analyte and the matrix. One example of sample p r e - t r e a t m e n t is the addition of an organic solvent to an aqueous sample. In order for analytes to be extracted effectively from a sample, they m u s t be free and available in solution, and able to diffuse into the pores of the sorbent. Since the pores are nominally 60 a n g s t r o m s in diameter, any analytes t h a t are agglomerated or associated with a particles larger t h a n 60 angstroms will be unable to diffuse into the pores, and the effective surface area is decreased from 500 to 0.1 meters 2. Such is also the case when analytes are bound to proteins or colloidal particles. The addition of an organic solvent to the sample helps to disrupt the interactions between the analyte and other particles, rendering t h e m available to interact with the sorbent. In this manner, the addition of organic solvent serves to minimize inter-analyte interactions, minimize interactions between the analyte and vessel walls, and minimize interactions between particulate m a t t e r and other matrix components. It should be noted t h a t a concentration of solvent t h a t is too high will enhance the matrix ~ analyte interactions, and retention of the analyte on the surface will be reduced.

46 Additional sample pre-treatment can include adding chemicals to remove interfering species, such as the addition of sodium sulfite or sodium thiosulfate to reduce residual chlorine from drinking water. The pH of a sample is often adjusted during sample pre-treatment to insure that the analytes are either ionized or not, depending on the requirements for the extraction. If the extraction is dependent on hydrophobic interactions for retention of the analyte, it will most likely be desirable to adjust the pH conditions so that the analyte, if ionizable, is largely in a neutral form. In this manner, the interactions between the analyte and matrix are minimized, and those between the analyte and bonded phase are optimized. To ensure quantitative retention and elution, an appropriate pH needs to be selected. We know from the acid-base approximation that: [A-] pH = pK a + log [HA] If the pH is equal to the pKa, there will be equal concentrations of the acidic and basic forms, which leads to inefficient chromatography. Therefore, a pH of about two units away from the pKa should be selected to achieve quantitative retention and elution of the analyte. As an example, if the analyte has an acidic function, it would be necessary to raise the pH to a value roughly two pH units above the pKa, where approximately 99% of the species present will be ionized. Alternatively, if the analyte is basic, it would be necessary to lower the pH to roughly two pH units below the pKa for the base to be ionized, or two units above the pKa if it necessary for the base to be neutralized. The pH of the sample must also be considered with respect to its impact on the sorbent, since this too may impact analyte retention. The silanol surface has a pKa of approximately four to six (Figure 10). At a pH of seven, the surface is

d

O-

QH

'Si~

~

%

'.O.j o, Protonated Figure 10. Impact of pH on the charge of silica surface.

47 largely ionized, and negatively charged. At a pH of three, the surface is largely protonated and neutral. This will have a dramatic effect on those analytes that are ionizable species, where the mechanism for retention is ion exchange. Even when a hydrophobic phase is being used, the analyte may be retained through polar or ionic secondary interactions with the surface, and so the pH needs to be considered [5]. In section 2.3, ion exchangers were described as being either strong or weak, depending on whether the charge was permanent or could be controlled by pH. For weak ion-exchangers, the pH of the sample must be adjusted to ensure that a charge on the ion exchanger is maintained. For example, since an amino propyl (NH2) column has a pKa of 9.8, the sample pH must adjust to no greater than 7.8. Since the weak cation exchanger, CBA, has a pKa of 4.8, the sample pH should be no lower t h a n 6.8. Ions in the sample can compete for ion exchange sites on the sorbent, resulting in low analyte recoveries. When the pH of the sample is adjusted, a buffer is selected such that the ionic strength of the adjusted sample does not exceed 50 mM. In addition, the type of buffer used should be such that the anions or cations are not more strongly retained t h a n the analyte itself. The following series lists cations on the left that will displace those on the right: Ba+>Ag+>Ca+2>Zn+2>K+>NH4+>H+>Li § The following series lists anions on the right that will displace those on the left: O H > acetate> formate> HPO4>HCO~>CI-> HS03> Citrate 2.5.2.

Column solvation After sample pre-treatment, the next step in an SPE procedure is to solvate (condition or wet) the sorbent. It has been demonstrated that the modified surface must be conditioned in order for it to be active and available to the analyte [6]. If the surface has been modified with a hydrophobic species such as octadecylchlorosilane (C18), and then is placed in an aqueous environment, the long hydrophobic chains will collapse upon themselves. The surface can be conditioned with an organic solvent, such as methanol. This will result in solvation of the hydrophobic chains, allowing the chains to become extended, and hence available to interact with the analyte (Figure 11). When sample volume

Unconditioned

Conditioned

Figure 11. Impact of conditioning modified silica surface.

48 exceed 100 mL, it is often pretreated with methanol, which is added as a "wetting agent" to keep the chains extend during the extraction process. Methanol is a good choice for the solvent since the hydrophobic end of the molecule can interact with the hydrocarbon chains, while the polar end can interact with the surface silanols. If no wetting agent is present in the sample, there will be a tendency for large volume aqueous samples to drag the conditioning solvent away from the surface of the sorbent, allowing the chains to collapse. With an adequate concentration of wetting agent present, the organic solvent can remain in a steady state concentration around the immobilized chains. The concentration of the wetting agent needs to be high enough to maintain a conditioned surface, and is typically added at a concentration of 0.5 to 1%.

2.5.3.

Column equilibration

The third step in a solid phase extraction procedure is typically column equilibration. Here, excess organic solvent is removed from the sorbent so that that analyte retention is not hindered. If conditioning solvent is present during sample loading, analytes may remain in a highly organic mobile phase, and pass through the column. In addition to removing excess conditioning solvent, the equilibration step serves to set the conditions of the column, such as pH and ionic strength. These parameters are normally adjusted to the conditions of the pre-treated sample, since the same considerations need to be made with respect to the charge on the analyte and sorbent surface. Equilibrating the surface is necessary to ensure that analytes reaching the sorbent during the early part of sample loading are extracted under the same conditions as those loaded at the end of the extraction.

2.5.4.

Sample loading

In step four of the solid phase extraction procedure, the sample is loaded onto the column. The loading rate necessary may vary significantly depending on the nature of the analytes, and on which type of interactions are being relied to extract them. For compounds having a strong affinity for the sorbent, sorption can take place in a small segment of the sorbent bed, allowing large volumes to be handled at high sampling speeds [7]. The important aspect of sample loading in solid phase extraction is the amount of contact time that an analyte is allowed with the serbent. A sufficient amount of time is required for an interaction to occur. At high flow rates, non-equilibrium conditions may exist, resulting in lower partition coefficients [8]. This will result in analyte breakthrough and consequently reduced recoveries (Figure 12). For analytes that are retained primarily through hydrophobic interactions, significantly higher loading rates are possible as compared to analytes retained through an ion exchange mechanism. Sample loading rates for the retention of hydrophobic species as high as 120 mL per minute have yielded quantitative recoveries on a 6 mL extraction cartridge [9]. It is recommended that a loading rate of five to ten mL per minute not be exceeded for a 1 mL, 100 mg sorbent bed [6]. Five mL per minute represents a linear velocity of 0.42 cm/sec, and a

49

100 % Recovery 50

Increasing Flow Rate Figure 12. Impact of loading rate. residence time of 1.9 seconds. For a 100 mg column, having a bed height of eight millimeters, there is an estimated residence time of 0.19 seconds. For ionic species, the analytes can be surrounded by a solvent sphere, hindering the interaction of analytes with the sorbent surface. Therefore, for analytes that are retained by ion exchange, the linear velocity for sample loading should be to the order of five cm per minute (one mL per minute on a one mL extraction cartridge) [6].

2.5.5. I n t e r f e r e n c e r e m o v a l An interference removal step generally follows sample loading. The extraction cartridge is treated to remove species that could interfere with the analytical determination. This usually involves rinsing the cartridge with a suitable solvent (one that will remove interferences without loss of the analyte). Appropriate retention conditions, such as pH and ionic strength, should be maintained to avoid loss of analyte. The equilibration solvent is commonly used in the interference removal step. If the analytes are well retained, a fraction of organic solvent can be added to the equilibration solvent to remove additional interferences. Drying the sorbent under vacuum or with a gas such as nitrogen or carbon dioxide may be required to remove water. Water may interfere with the elution of analytes from the sorbent if water immiscible solvents are used. Alternatively, interferences may be removed by selecting a solvent that is able to elute analytes from the sorbent, while interferences are retained on the column. Removal of water is critical if the analytical determination is gravimetric. Water that remains on the column can be eluted with the analyte and contribute to the final weight. 2.5.6. Analyte e l u t i o n The final step of the SPE procedure is elution of the analytes from the sorbent. A suitable solvent will preferably have a low viscosity, be readily available in a

50 pure form and have low flammability and toxicity. Important properties are strength and selectivity [9]. The solvent should be one that can to some degree solubilize the analyte. An appropriate solvent must be chosen to enhance the matrix~-~sorbent, and matrix~-~analyte interactions, while minimizing s o r b e n t ~ a n a l y t e interactions. The strength of the elution solvent is related to the mechanism by which the analyte is being retained. Table 1 shows the relationship between solvent strength and type of mechanism by which an analyte is being retained.

Table 1 Mechanism versus solvent strength

Mechanism

Increasing Solvent Strength

Non-polar

Water =v Methanol r Hexane

Polar

Hexane ~ Methanol =~ Water Methanol or water r Methanol and water

Multiple interaction

Since an analyte may be retained through multiple interactions on a heterogeneous surface, a mixed elution solvent is often most effective. An important consideration when choosing an elution solvent is the final analysis. If the analyte concentration is to be determined by HPLC, the mobile phase can often be used to elute the analyte. If the elution solvent must be evaporated to dryness and reconstituted, then a volatile solvent should be selected. Appropriate elution conditions, such as pH and ionic strength, should be considered. It may be necessary to adjust the pH of the elution solvent to neutralize the analyte or surface of the sorbent. Ionic species are typically eluted by adjusting the ionic strength of the elution solvent to 0.1 molar for monovalent analytes and 0.2 molar for divalent analytes. It may be useful to elute the analyte with a buffer containing counter ions that are better retained on the surface than the analyte. Refer to section 2.5.1 for ion selectivities. In addition to selecting an appropriate solvent, there should be sufficient contact time between the sorbent and solvent to ensure quantitative removal of the analyte from the surface. In section 2.1, the porous nature of the sorbent was described. During the loading step, analytes can diffuse deeply into the pores. When possible, the elution should be preformed using two aliquots of solvent, and allowing a 1-2 minute soak step between the two elutions. This allows sufficient time analytes to diffuse back out of the pores. In addition, the inclusion of a soak step allows the analyte to be eluted using minimal solvent volumes, which maximizes trace enrichment.

51

3. IMPACT OF VARIOUS FACTORS ON SOLID P H A S E EXTRACTION In the previous section, each of the steps involved in an SPE procedure was described. There are a variety of factors that can have an impact on the selectivity and efficiency of the extraction. Some of these factors include the sample loading and elution rates, choice of sorbent materials, as well as selection and volume of conditioning, equilibration, rinse and elution solvents. The influence of sample pH and the choice of solvents added to samples prior to loading on to the SPE column can have a significant impact on analyte recovery. Failure to consider these various aspects in solid phase extraction procedures can result in non-robust methods, lengthy development times and excessive costs.

3.1. I m p a c t of v a r i o u s b o n d e d p h a s e c h a i n l e n g t h s on a n a l y t e s e l e c t i v i t y It is known that C18 is the most hydrophobic phase of the bonded silicas, while C2 is the most polar of the hydrophobic phases [6]. It is instructive to look at the impact of a variety of bonded phases on analyte selectivity. By understanding the extent of these interactions, analyte recoveries can be optimized. The impact of bonded phase chain length can be examined by utilizing the EPA Method 525.1 analytes as probes [9]. Samples loaded on ISOLUTE | bonded silica modified with hydrophobic phases C2, phenyl, C8 and C18 were used to generate the data in Table 2. The behavior of these analytes varied with respect to the differences in their structures (size and degree of hydrophobicity), and the type of sorbent selected. For the purposes of analyzing the data, this discussion will characterize the analytes by grouping together those compounds that tended to behave similarly. Those compounds that were eluted early in the gas chromatographic run are grouped together. These compounds were generally small and/or fairly polar. The compounds eluted in the middle of the run were also grouped together, and then again, the late eluting compounds. The late eluting compounds tended to be fairly large and hydrophobic in character.

Table 2 Impact of sorbent type on recovery of EPA 525.1 analytes COMPOUND CI8(EC) CI8(EC) 89 lg 1 Hexachlorocyclopentadiene 62 66 2 Dimethylphthalate 92 100 3 Acenaphthalene 94 92 4 Acenaphthene-dl0 i n t e r n a 1 st 5 2-chlorobiphenyl 87 85 6 Diethylphthalate 100 100 7 Fluorene 89 88 8 2,3-dichlorobiphenyl 79 77 9 Hexachlorobenzene 66 64 10 Simazine 78 69

PH lg 41 8 27 andard 80 58 61 78 64 10

C2 lg 27 1 8 26 10 22 54 53 0

52 Table 2 (continued) I m p a c t of sorbent type on recovery of EPA 525.1 analytes COMPOUND C18(EC) CI8(EC) 89 lg 11 Atrazine 60 33 12 Pentachlorophenol 13 Lindane 99 96 internal 14 P h e n a n t h r e n e - d l 0 15 P h e n a n t h r e n e 83 82 16 Anthracene 76 75 17 2,4,5-trichlorobiphenyl 70 66 100 97 18 Alachlor 76 72 19 Heptachlor 95 93 20 di-n-butylphthalate 21 2,2', 4,4'-tetrachlorobiphenyl 74 71 22 Aldrin 70 63 23 Heptachlor epoxide 88 85 24 2,2', 3',4,6-pentachlorobiphenyl 72 70 25 G a m m a - c h l o r d a n e 78 74 76 74 26 Pyrene 79 74 27 alpha-chlordane 76 72 28 t r a n s nonachlor 29 2,2', 4,4', 5,6'-hexachlorobiphenyl 67 69 62 67 30 Endrin 31 B u t y l b e n z y l p h t h a l a t e 83 78 83 92 32 di(2 -ethylhexyl)adipate 70 71 33 benz[a] a n t h r a c e n e internal 34 Chrysene d-12 76 78 35 Chrysene 77 78 36 2,2'3,3',4,4',6-heptachlorobiphenyl 83 83 37 Methoxychlor 76 79 38 2,2',3,3',4,5',6,6'-octachlorobenzene 78 79 39 di(2-ethylhexyl)phthalate 72 71 40 benzo[b]fluoranthene 68 70 41 benzo [k]fluoranthene 58 60 42 benzo[a]pyrene 83 85 43 perylene-d12 57 58 44 indeno[1,2,3,c,d]pyrene 58 58 45 dibenz [a,h] a n t h r a c e n e 61 61 46 Benzo[g,h,i]perylene Avg: 77 76 early eluting: mid eluting: late eluting:

87 78 65

89 75 66

PH lg 1 24 st andard 80 75 66 94 72 97 73 67 87 74 78 80 79 76 78 56 90 82 75 st andard 82 78 89 76 100 78 79 72 99 71 75 79 69 46 72 79

C2 lg 0 3 38 40 61 16 66 95 70 66 52 73 78 68 78 78 74 32 100 86 74 77 70 100 68 100 80 78 75 100 81 87 82 58 16 61 83

53 For those compounds that were eluted early in the chromatographic run, there was a significant reduction in recoveries as the chain length of the bonded phase decreased. Recoveries for these earlier eluting compounds averaged 89% when extracted onto one gram of C18 material, dropping to 46% when extracted using a phenyl phase, and down to 16% when extracted onto C2. This suggests that the earlier eluting compounds require a sorbent having long hydrophobic chains in order to be extracted from a very polar matrix, since the extraction is based primarily on non-polar interactions. As the hydrocarbon chains on the sorbent get shorter, the silanols on the surface become more accessible causing the surface to become more polar in nature, and the recoveries of the earlier eluting compounds dropped off. For those compounds that were eluted in the middle group of the gas chromatographic run, there was not a significant difference in the recoveries from the phenyl, C8 and C18 phases, ranging from 72 to 78%. There was a drop in recoveries on the C2 phase, averaging 61%. The decline in the recoveries for these somewhat larger and less polar compounds was not as dramatic as that for the earlier eluting compounds, suggesting that the choice of bonded phase for analytes of intermediate polarity is less critical. It can be seen from the data that for the later eluting compounds, the best recoveries were obtained when the extraction was performed using bonded phases with modified with shorter hydrocarbon chains. In contrast to the earlier eluting compounds, as the hydroca,'bon chain length of the bonded phase increased, the recovery of these analytes dropped. For these later eluting compounds, it is energetically less favorable to remain in an aqueous matrix as compared to the earlier eluting compounds. Since this group of compounds tends to be large and quite hydrophobic, they are retained very strongly to a C18 surface during sample loading. This strong interaction between the analytes and sorbent also makes it difficult to remove these compounds from the surface during the elution step. In the case of the EPA Method 525.1 analytes, where a broad range of compounds present, it is possible to select a "compromise phase" such as C8. In this case, the chain length of the bonded phases is sufficiently long to provide a reasonable amount of hydrophobic interaction with the earlier eluting compounds, while having a surface that is sufficiently polar to allow for the elution of the later eluting compounds.

3.2. I m p a c t o f t e m p e r a t u r e on a n a l y t e r e c o v e r y The EPA Method 525.1 compounds were extracted at room temperature (20~ as well as at 4~ An improvement in recoveries can be observed for analytes extracted at a lewer temperature. For analytes extracted using an ISOLUTE | C18 phase, average overall recoveries improved from 75% when extracted at 20~ to 85% when extracted at 4~ The impact of temperature was less dramatic for analytes extracted on ISOLUTE | C8, where the extraction at 20~ gave an average recovery of 85% versus 90% when extracted at 4~ The impact of analyte solubility was considered when examining the phenomenon of improved recoveries at lower loading temperatures. It is likely

54 t h a t the analytes are less soluble at lower t e m p e r a t u r e s , and could therefore be more easily extracted from an aqueous matrix. However, when samples were extracted on cartridges stacked in series at the elevated t e m p e r a t u r e , there was no evidence of analyte breakthrough. Since lower recoveries can not be a t t r i b u t e d to analyte breakthrough, this implies t h a t lower t e m p e r a t u r e s do not improve analyte retention, and the i m p r o v e m e n t in recoveries m u s t therefore be a t t r i b u t e d to an i m p r o v e m e n t in analyte elution. An alternative explanation would be to consider the impact of t e m p e r a t u r e on the diffusion r a t e of analytes. It is known t h a t at elevated t e m p e r a t u r e s , the rate of diffusion of most molecules increases. T h a t the diffusion rate of a molecule is a function of t e m p e r a t u r e and is described by Fick's Law, where it is proportional to the square root of t e m p e r a t u r e . Calculating the difference in diffusion rate at 20~ versus 4~ gives the following: D293=k x (294) 2 D277=k x (277) 2 D293 / D277 = 1.13 It can be therefore estimated t h a t there is an increase in diffusion rate by 13%. For those samples extracted at a higher t e m p e r a t u r e , as the rate of diffusion of the molecules increases, analytes can diffuse more deeply into the pores of the silica. The more deeply the analytes are retained in the pores, the more time t h a t is required to allow the molecules to diffuse back out of the pores (Figure 13). This is consistent with experiments described in the section 3.4, addressing the inclusion of a soak step during analyte elution.

Figure 13. Diffusion into pores of silica.

55

3.3. I m p a c t o f e l u t i o n s o l v e n t The broad range of analyte characteristics of the EPA Method 525.1 analytes m a k e t h e m suitable to illustrate the impact of the elution solvent on analyte recovery. A variety of solvents were tested, both as pure and as mixed solvents as shown in Table 3. Two elutions per column were performed. It can be seen from the data t h a t the average recovery for elution solvents one through four in Table 3 is 69%. In each case, the elution was performed using a pure solvent for both elution steps, although the solvent wasn't necessarily the same for each step. Elutions 5-13 were done with mixed solvents, with the average recovery being 79% and a s t a n d a r d deviation of 5%. From these results it can be seen t h a t the highest recoveries were obtained when mixed solvents were used, as compared to pure solvents. This is consistent with our picture of the mechanisms by which analytes are r e t a i n e d on the surface of the silica. Since compounds may be retained through multiple interactions, analyte elution can be optimized by i n t e r r u p t i n g those interactions with a solvent system t h a t can solvate the analyte using multiple interactions. If a very polar elution solvent is used, analytes can be retained through hydrophobic interactions with the bonded phase. If a nonpolar solvent is used, analytes can be retained through interactions with silanol groups on the silica surface. Solvents five, seven and eight in Table 3 were mixtures of ethyl acetate and acetone in ratios of three to one, one to one, and one to three, respectively. The average recovery for each was 86, 86 and 79 percent, with a relative s t a n d a r d deviation of 5%. Therefore it can be noted t h a t relative concentrations of the mixed solvents is not critical for this solvent combination. For achieving lower detection limits, however, the fraction of w a t e r miscible solvent should be considered with respect to the limit to which the solvent can be concentrated. If it is necessary to concentrate the extract after elution, there m u s t be a sufficient volume of w a t e r miscible solvent present to prevent phase separation if traces of water are eluted along with the analytes. 3.4. I m p a c t o f e l u t i o n s o a k step on a n a l y t e r e c o v e r y Silica gels are an agglomeration of particles resulting in a very porous in structure. The structure of the silica varies with respect to pore size and surface area. The presence and dimensions of pores has a significant impact on the surface area t h a t is available with which an analyte can interact, dramatically improving the efficiency of the extraction. Commercially available silica used in solid phase extraction typically have a nominal pore size of fifty to sixty angstroms. This section addresses the impact t h a t the pores have on analyte recoveries, due to the diffusion of molecules into and out of the pores. A series of experiments were devised to study the impact t h a t including a soak step between elution volumes would have on analyte recoveries [9]. It can be seen from the d a t a in Table 4 t h a t for equal volumes of elution solvent, there are an improvements in recoverieswhen the elution is performed using two separate elution volumes with a two minute soak step between elutions, versus eluting the

56

oo

o 0 ,--.I

9 .,-i

~

oo

c)

@

9 OO 9

.,-i

o o o (D

N

o,-~

l~l

~

, ~

~

~

~

~

~

~

~,

o - ~ Oe~z'~O0oOl~-~. I ~ l ' ~ r ~ ~ 1 7 6

~'~ r

o

~

~

~o c x l ~ D o o ~ ' ~

~

o o

o

,_o

~

9

o

o

o

.;,.~

9 o

o

o

o

o

OO

r

A

9

~

0

"'~

~

~ ~ ~ ~ ~..

=

O

~ ~ 0 ~ o 0 0 ~

~.o

0

~- o ~

~ o o ~

0

N

O

_

o ~ o ~

~ o ~ ~ ~

dH

o 0~

o

n~

o

o

5"7

o

o

58

Table 4 % Rec. of EPA 525.1 analytes from ISOLUTE | C18 with and without 2 min soak step Two aliquots One Aliquot 3 + 2 mL 2 + 2 mL 1 . 5 + 1 . 5 m L 3 mL 4 mL 5 mL 82 77 80 74 72 74

columns with a single solvent volume. The average recovery for analytes eluted in a single step is shown to be 72%. When the elution was performed in two steps, which included a two minute soak step, the average recovery increased to 78%. The data suggests t h a t for analytes t h a t have diffused more deeply into the pores of the silica, the addition of a soak step allows more time for the analytes to diffuse back out of the pores, and recoveries are improved. This indicates a dependence on time required for the analytes to diffuse out of the pores of the silica. Simply increasing the elution volume did not have a significant impact on recoveries, as seen for the three mL versus five mL elution, where the recoveries were 73 and 72 percent respectively. An additional experiment was performed to determine the impact of elution rate on analyte recovery. Analytes were eluted using flow rates from five to forty mL per minute. Recoveries at five mL per minute average 79% while those at forty mL per minute averaged 77%, S t a n d a r d deviations are four to five percent, and so the difference in recoveries between five and forty mL per minute are statistically insignificant. The most i m p o r t a n t aspect of the elution is the contact time t h a t the elution solvent has with the surface of the silica. For samples eluted with a total volume of five mL, the difference between an elution rate of five versus forty mL per m i n u t e is only a difference in contact time of 60 seconds versus eight seconds. Both of these contact times are less t h a n half of the total contact time allowed when a two minute soak step is included between elution volumes, and so the benefit of reducing the elution rate is insignificant when a soak step is included. It has been shown t h a t the porous n a t u r e of silica used in solid phase extraction plays and i m p o r t a n t role in the extraction as well as in the elution of analytes. It is known t h a t the porosity of the silica provides the analyte with sufficient surface area to interact with the sorbent during analyte retention. During the extraction step, molecules can diffuse deeply into the pores of the silica. As a result, t h a t sufficient contact time m u s t be allowed between the solvent and sorbent during the elution step to provide analyte molecules to diffuse back out of the pores. It has been shown t h a t when samples are loaded at a reduced t e m p e r a t u r e , decreasing the diffusion rate of the analytes, the depth to which analytes p e n e t r a t e into the pores is reduced, and recoveries are improved. It has also been shown t h a t including a soak step between elution volumes can improve recoveries by allowing analytes sufficient time to diffuse back out of the pores.

59 3.5. I m p a c t o f c h a i n l e n g t h o n t h e r e t e n t i o n o f w a t e r In section 3.1, the impact of the chain length on the bonded phase on the retention of analytes was discussed. This section will address the impact that the hydrocarbon chain length has on the water retention characteristics of the sorbent. Earlier in this chapter, it was described that the silanol groups on a C2 surface are well exposed, and there is a significant amount of water associated with this surface. On a C8 surface, the hydrocarbon chains are longer. Access to the surface is somewhat restricted, but the chain lengths are still short relative to the distance between the chains. The silanol groups are still exposed, and the surface is still somewhat polar. On a C18 surface, the long chains, which can reach each other across silanol groups, serve to "water proof' the surface. Although it is still possible for water to reach the surface of the silica, the hydrocarbon chains are mostly associated with organic solvent molecules.

Nitrogen Drying of Bonded Phases 20 PSIG at AutoTrace (4.2 L/min) W 1.6 C 12 (WP)

A

1.4

T 1.2 E R, 1

C 2 (UC)

c 8 (uc)

0.8 G 0.6

C 12 (UC)

R A 0.4

c 18 (uc)

M 0.2 S

o

5

10

15

20

25

30

C 18 (EC)

Time, min Figure 14. Comparison of drying times for various ISOLUTE| phases.

Various phases were tested with respect the time required to dry the sorbent to constant weight (Figure 14). It has been demonstrated by Fung Kee Fung using differential scanning calorimetry that the silica surface can organize water in the near surface region, imparting an "ice-like" structure [11]. Since silanol groups are more easily accessed by water on sorbents modified with shorter hydrocarbon chains, it is expected that the surface would have a greater impact on structuring the water in the near surface region than those surfaces modified with longer hydrocarbons. It can be seen from the graph in Figure 14 that the bonded phase having the shortest chains required longer drying times, while those with longer hydrophobic chains (C8, C18) required increasingly shorter drying times.

60 Surface Coverage, ~tM2/m Hydrocarbon Volume, mL/t

Void Volume, mL/g: 0.67

2.3

1.6

0.08

0.20

Void Volu~ mL/g: 0.55

Figure 15. Comparison of void volumes for C2 versus C 18.

Figure 15 illustrates the difference in pore volumes for a C2 versus a C18 bonded silica. The pore void volume is decreased for a C18 phase, corresponding to the increase in volume of hydrocarbon present. There is only a 20% difference in volume between the two phases, which is insufficient to account for the two to three fold differences in drying times. For each of the curves in Figure 14, there are two distinct regions. The region having the greatest slope for time versus grams of water removed represent water that is being removed form between sorbent particles, and from the pores of the silica. The less steep region of each of the curves represents water removed form the near surface region. The slope of each line decreases with decreasing hydrocarbon chain length, indicating that it becomes more difficult to remove more highly structured water from the near surface region. The y-intercept gives an indication of the amount of water that is associated with the surface after the bulk (water between particles and in the pores) water is removed. Using these volumes, the depth of the water in the near surface region may be calculated from the known surface areas of the material (approximately 550 m2). The endcapped material dried slightly more quickly than the uncapped material, which would be expected, since there are fewer silanol groups on the endcapped material, reducing the ability to structure water in the near surface region. 3.6. I m p a c t o f p h y s i c a l p a r a m e t e r s on d r y i n g t i m e s In the previous section, the difference in water retention characteristics on various phases was examined. The differences observed were due to differences in surface chemistry. In this section, an experiment is described that examines the difference in the retention of water due to physical differences in the particles, specifically, the impact of the distribution of particle sizes. Sorbent beds

61

versus

versus

Figure 16. Flow characteristics through columns cotaining fines versus fines-free. should ideally contain material that has a narrow particle size distribution (nominally fifty microns), with very few fines present (particle size less than 20 microns). When fines are present in the bulk silica, there is a tendency for the smaller particles to agglomerate. When this material is packed into a cartridge, agglomerations of large and small particles can result in poor flow characteristics (channeling of flow through the bed) as shown in Figure 16. When the sorbent is being dried, channeling results in areas of the bed where there is the least resistance to flow. This results in those areas having contact with the greatest volume of gas, and so drying occurs quickly. In regions of the bed where there is a greater pressure drop (where smaller particles have agglomerated), there is exposure to a smaller volume of gas, and drying times are greatly extended. The drying data for silica from different manufactures is consistent with the data obtained for particle size distribution. The presence of fines could have the same impact on liquid flow as well as on gas flow, resulting in increased volume or soak requirements for conditioning and elution steps, and channeling during loading steps. 3.7. I m p a c t o f w a t e r o n a n a l y t e r e c o v e r y

The importance of removing water from the sorbent bed varies, depending on the type of analysis being performed. If water is present in the elution solvent, additional steps may be required to remove water from the elution solvent, such as passing the extract through sodium sulfate. This can contribute to loss of analytes, as well as being a source of contamination. For samples that require evaporating to dryness, the presence of water can greatly extend the drying time. In addition, the presence of water may greatly limit the ability to use derivatization chemistry. Adequate drying time for the extraction cartridge is essential after an aqueous sample has been loaded, if the sorbent is to subsequently be eluted with a water immiscible solvent. Insufficient drying of the cartridge can result in inadequate contact between the elution solvent and the sorbent (Figure 17), as well as partitioning of the analytes between phases.

62

Methylene chloride

=,-,===~jWm~

H /%

/%

..

o /

/ON

H

,2

A

/0

?H RH

/S{'o

\

N A

L~I

T

E

H

H\ H/O

S

Si~"

,s,.,

\ 0 S{"O"Si

\

Six

l"

Figure 17. Water as an interference to analyte elution.

The sorbent is often dried with a gas such as nitrogen or carbon dioxide. An a l t e r n a t i v e to e x h a u s t i v e l y drying the sorbent with a gas is to use a watermiscible solvent, such as acetone, as a component of an elution solvent mixture. A solvent such as acetone is able to bridge the properties b e t w e e n two immiscible solvents, such as w a t e r a n d m e t h y l e n e chloride. W h e n such a solvent s y s t e m is utilized, the r e s u l t is to chemically dry the sorbent, removing w a t e r from the pores, a n d a w a y from the surface region of the sorbent. In this m a n n e r , good contact is achieved b e t w e e n the solvent a n d a n a l y t e s (Figure 18). An e x p e r i m e n t was p e r f o r m e d to d e t e r m i n e if there was an i m p a c t on the r e s u l t s w h e n the cartridge drying time was reduced from ten m i n u t e s to 0.5 m i n u t e s prior to an elution with a w a t e r miscible solvent (acetone/ethyl a c e t a t e

/%

H

H

H

\ \

----

H\

E.-'~

Ac

---s~-~ x . ~ ~ ,,O OH i ~H -'~k

H/O

-~~-o~

p

Si'o'S 'o'Si'o-s, P'~ P""o "'i'" o ,s, 9 /\ \

six

Figure 18. Water miscible solvent chemically drying surface and eluting analytes.

53 75:25) [9]. A half m i n u t e was shown to be sufficient to displace bulk w a t e r from the silica bed. There was no significant difference between the two sets of data, each having an average 82% recovery, with a five percent s t a n d a r d deviation. It h a s been shown t h a t the drying times for bonded phases vary significantly, depending on the chain length of the hydrocarbon phase used for bonding, and the type of silica used. For m a t e r i a l t h a t has been bonded with short hydrocarbon chains (e.g., C2), drying times are significantly longer t h a n those for silica bonded with long hydrocarbon chains (e.g., C18). These results are consistent with the idea t h a t a surface such as C2 is very polar clue to easily accessible silanol groups, and t h a t the impact from this more polar surface is to i m p a r t a higher degree of s t r u c t u r e on the w a t e r in the near surface region. The silanol groups on a C18 surface are buried under the long hydrocarbon chains, which makes this bonded phase the least polar of the modified silicas, having the least ability to organize w a t e r in the n e a r surface region.

3.8. Impact of wetting agent A series of experiments were performed to study the impact on analyte retention of increasing the concentration of wetting agent in an aqueous sample [9]. It was shown t h a t the results obtained for EPA Method 525.1 analytes vary for different types of compounds as the concentration of wetting agent is increased. For smaller and more polar compounds, such as dimethylphthalate, diethylphthalate, simazine and atrazine, the recoveries decreased significantly as the concentration of wetting agent increased, averaging 93% recovery at 0.5% isopropanol, and dropping to an average of 12% recovery at a concentration of 20% isopropanol. This behavior is consistent with w h a t would be predicted, since as the m a t r i x becomes more organic, it competes with the surface for analytes t h a t are strongly dependent on hydrophobic interactions for retention, as the m a t r i x - a n a l y t e interactions are enhanced. For compounds t h a t are eluted in the middle of the chromatographic run, which tend to be of moderate size and/or polar character, there was very little impact on recoveries as the concentration of organic solvent was increased from 0.5 to 20%. The average recovery for this group of compounds over the range of concentrations tested was 79%, with a s t a n d a r d deviation of 3%. In the case of the later eluting compounds, which are the larger and more hydrophobic analytes, recoveries improved as the concentration of organic solvent was increased in the sample matrix, the average recovery for the compounds such as di(2-ethylhexyl)phthalate to benzo[g,h,i]perylene increased from 73% at a concentration of 0.5% IPA to 82% at a concentration of 20% IPA. The fact t h a t the recoveries improved for the more hydrophobic analytes at higher concentrations of solvent could not be a t t r i b u t e d strictly to m a i n t a i n i n g an active silica surface, since 0.5% m e t h a n o l is adequate for this purpose [12]. It would be expected t h a t analytes with more hydrophobic character would be well retained on a C18 surface w h e n extracted from an aqueous, or even partially aqueous matrix. As the concentration of organic solvent is increased in the sample, the analytes themselves are better solvated, and can be more easily eluted from a C18 surface.

64 The experiment of increasing the concentration of organic solvent in the sample was repeated, using polychlorinated biphenyls as the probe set and 200 mg of polystyrene divinylbenzene as the sorbent. Polystyrene divinylbenzene differs from bonded silica, in that the surface does not need to be conditioned with an organic solvent to remain active. Therefore, any effect from the addition of organic solvent to the sample must be due to an impact on the analytes rather than on the sorbent. The average recovery of polychlorinated biphenyls from water at a concentration of 2.0 PPB was 88%. A second set of samples was spiked to a concentration of 10.0 PPB. As the samples were extracted, the effluent water was collected and subsequently re-extracted using methylene chloride in a liquidliquid extraction procedure. For the earlier eluting compounds (2-chlorobiphenyl, 2,3-dichlorobiphenyl and 2,4,5-trichlorobiphenyl), the average breakthrough was eleven percent. For the later eluting compounds (2,2',4,4',5,6-hexachlorobiphenyl, 2 ,2', 3, 3', 4, 4', 6-heptachlorobiphenyl and 2,2',3,3',4,5,6,6'-octachlorobiphenyl), the average breakthrough was fifty two percent. It was demonstrated that as the analytes become larger and more hydrophobic in character, the tendency for breaking through the extraction cartridge increased. The PCB solid phase extraction at 10 PPB was repeated using increasing concentrations of isopropyl alcohol as the wetting agent. The phenomenon that was observed was the same as that seen with the EPA Method 525.1 analytes, where the recoveries of more hydrophobic analytes improved as the concentration of IPA was increased. The improvement in recoveries for these analytes was observed for IPA concentrations as high as 30%, where the average recovery was 97%, dropping to 66% at an IPA concentration of 40%. The experiment was repeated using methanol as the wetting agent. It was shown that recoveries for higher molecular weight compounds do not decrease until the concentration of methanol is increased to from 50% to 70%, where the average recovery dropped to from 97% to 55%. As the concentration of less water-soluble compounds is increased, there is a tendency for the analytes to agglomerate. This behavior becomes more apparent for the later eluting analytes, since these are also the compounds that are less soluble in water. The earliest eluter, 2-chlorobiphenyl, has a solubility of 5.9 PPM [4]. A compound such as 2,2',4,4'-tetrachlorobiphenyl, which elutes at an intermediate time, has a solubility of 0.068 PPM, while octachlorobiphenyls have solubilities below 0.001 PPM [5]. The efficiency of solid phase extraction is dependent on the availability of a large surface area of the sorbent, which is provided by very porous material. When the size of an agglomerated analyte exceeds the size of the pores that are available, the agglomeration can no longer diffuse into the pores. Retention is greatly reduced, since the effective surface area is significantly reduced (from 1100 square meters to approximately 0.1 square meters). When an organic solvent is added, the compounds become better solvated, and the formation of agglomerations is reduced. The molecular compounds are then able to diffuse into the pores, and the effective surface area is increased.

55

4. THE A P P L I C A T I O N OF S P E TO E N V I R O N M E N T A L A N A L Y S E S The use of solid phase extraction (SPE) for environmental analyses is a rapidly growing area in analytical chemistry. The challenge in developing SPE procedures is to selectively concentrate the analytes of interest, maximize their recovery and minimize interferences. For the analysis of a broad range of analytes, conditions must be selected to optimize the recoveries of compounds that are quite varied in properties.

4.1. S e l e c t i v e e l u t i o n o f a n a l y t e s for e n v i r o n m e n t a l a p p l i c a t i o n In the previous sections, the impact of the bonded phase on the retention and elution of a broad range of analytes was examined. In the current section, experiments are described in which advantage is taken of the heterogeneous nature of bonded phases to retain analytes of differing hydrophobicity, and elute them selectively based on differences in functional groups. In this section, different sorbents types are layered for the purpose of extending the capacity to retain polar compounds during an elution with a non-polar solvent. The phenomenon being exploited is the selective elution of particular types of analytes. The analytes used are limited to one compound for which the retention mechanism is purely hydrophobic, and one capable of hydrophobic interaction as well as containing a functional group capable of polar interaction (hydrogen bonding with surface silanols). The analytes are the standards suggested by the EPA for the determination of Oil and Grease in Method 1664, which are hexadecane and stearic acid. Total Oil and Grease is operationally defined by EPA Method 1664 as those compounds that can be extracted from a sample of water using hexane as the extraction solvent (identified as hexane extractable material, or HEM). A subfraction of that material is further defined as silica gel treated hexane extractable material, or SGT-HEM. This fraction is the non-polar material, and is represented in the standard by hexadecane. The polar fraction is represented by stearic acid. The EPA method describes a liquid-liquid extraction, where one liter of sample is shaken vigorously with several portions of hexane, totaling about one hundred mL. Residual water is removed from the extract by passing it over solid sodium sulfate. The solvent is then evaporated, the residue is purged with air and then weighed to determine Total Oil & Grease. The residue is then redissolved in hexane, treated with three grams of silica gel to remove polar components, re-evaporated, purged again, and re-weighed. This residue is designated as the SGT-HEM fraction, also referred to as the Total Petroleum Hydrocarbons, or TPH fraction. If the selectivity of the modified silica surface can be exploited, the solid phase extraction can be used as an alternative to the liquid-liquid extraction procedure. Step one in the six-step SPE procedure involves pre-treating the sample. In the case of Oil and Grease, the sample is acidified to a pH between 1.9 and 2.1 to protonate the acid functions of the fatty acid. Methanol was added as a wetting agent at a concentration of 1%. Although isopropyl alcohol can be used in

66 pesticide work, it can not be used for Oil and Grease, since Oil and Grease is determined using a gravimetric finish. IPA has a higher boiling point t h a n methanol, and adds to the final weight of residue. For the same reason, the cartridge is conditioned with methanol. The equilibration step is performed with reagent water, acidified to the same pH as the sample. The sample is then loaded at rates ranging form 30 to 100 mL per minute. It was found that analytes did not break through the extraction cartridge at this loading rate, which was consistent with earlier work with pesticides [2]. After loading the sample, the cartridge is rinsed with water acidified to pH-2, to remove interferences such as inorganic salts. When extracted on a C18 phase, hexadecane is retained by hydrophobic interactions, and stearic acid is retained by both hydrophobic and polar interactions (Figure 19). In solid phase extraction, advantage can be taken of our

~ S j ~

II ' "

\

-

"1

I ~_ _

./

-

/

~S~

o.j \

u

\

u

\

o. u

Si

\

\

/_...0 ~ s,\

/

o i~o

i

Figure ]9. iLnalytes held by multiple versus single retention mechanisms.

ability selectively choose the characteristics of the bonded phase and elution solvents to retain and elute analytes based on differences in functional groups. The analytes can then be selectively eluted from the sorbent by first using hexane to interrupt the hydrophobic interactions to elute the hexadecane. During the hexane elution, the stearic acid continues to be retained in the sorbent through polar interactions between silanol groups on the surface of the silica, and the -OH group on the acid. Then a mixture of hexane and THF (1:1) can be used to elute the stearic acid fraction, where the THF serves to interrupt the polar interactions, and the hexane disrupts the hydrophobic interactions. Each of these fractions can be quantitatively eluted from the sorbent and collected separately. Ninety-six and one hundred thirteen percent recoveries were obtained for the non-polar and polar fractions respectively, from a 20 PPM spike of one liter of sample. In subsequent experiments, the concentration of the spike was increased, but several problems were encountered. Stearic acid began eluting with hexadecane

67 during the hexane extraction. It appeared t h a t the capacity of the sorbent for retaining polar compounds was being exceeded at spikes greater t h a n 40 PPM. Additionally, stearic acid, which is a waxy substance, has a tendency to precipitate from solution as it is spiked into reagent water. W h e n the samples were extracted using an a u t o m a t e d system (Tekmar AutoTrace), the sample lines became plugged coated with insoluble particles of stearic acid. Some samples were extracted m a n u a l l y by attaching a Teflon cap to the sample bottle, to which an extraction cartridge could be fitted. The cartridge was held in a lure tip syringe in a rubber stopper. A -27 inch v a c u u m drew the sample from the sample bottle into a vacuum flask (Figure 20). When the samples were extracted manually, plugging of the extraction cartridge occurred. Recoveries for both fractions, polar and non-polar, were low. At a concentration of forty mg/liter, Total Oil and Grease recovery from the AutoTrace dropped to 79%. A spike of 120 P P M gave a recovery of 64%. A new cartridge was designed for the

<

<

Loosenedcap

Extractioncartridge Stopcock to adjust flow rate To Vacuum Syringe needle

Figure 20. Manual extraction of large volume samples.

58

manual extraction that would prevent the precipitated stearic acid from plugging the bed. A depth filter made from a polyester web material was placed into the cartridge on top of the sorbent bed to collect suspended material. A layer of unmodified silica was placed under the C18 sorbent to increase the capacity of the sorbent bed for retaining polar compounds. A layer of 200 pm C18 particles was placed above the standard 50 pm C18 layer to help prevent plugging as molecular stearic acid is extracted on top of itself. Reagent water spiked with standards ranging in concentrations from 12 to 120 PPM averaged from ninety four to ninety eight percent recovery. The experiment was automated using the AutoTrace. To prevent stearic acid from plugging sample lines, the depth filter was placed at the bottom of the suction line going into the sample bottle instead of stacking it on top of the extraction cartridge. Prior to the drying and elution steps, the depth filter was removed from the suction line, placed under the plunger and attached to the extraction cartridge through a Teflon adapter. With this setup, both the depth filter and the extraction cartridge could be dried and then eluted simultaneously. When a depth filter was included in the automated extraction process, recoveries from a 120 PPM spike of hexadecane and stearic acid (1:1) were 89% and 103% respectively. When the method was automated, using t h e T e k m a r AutoTrace, recoveries of 90% or better were obtained. 4.2. I m p a c t o f p a r t i c u l a t e m a t t e r on solid p h a s e e x t r a c t i o n One of the challenges in solid phase extraction is in sample preparation of those samples that are heavily laden with particulate matter. Such samples are often found in natural water and wastewater systems. Particles can hinder sample extraction due to the tendency towards plugging the cartridge. Cartridge plugging is a phenomenon that has been observed industry wide when performing solid phase extraction on environmental samples. When plugging occurs, sample loading is significantly slowed, or flow may become completely stopped. In order to better understand the problem of cartridge plugging, experiments were designed to determine where the plugging was occurring. The sorbent in an extraction cartridge is held in place by two porous frits, one on the top and one on the bottom of the bed. It would be possible that plugging occurs above the top frit, within the frit itself, or within the sorbent bed. To help make this determination, five, eight and fifteen micrometer particles were loaded onto six mL, 200 milligram polystyrene divinylbenzene sorbent beds. The five micron particles passed right through the bed, and the effluent appeared cloudy. The eight micron particles plugged the flow after loading 500 mL. The fifteen micron particles piled up on top of the frit, while the sample flow remained steady. Although this experiment gave an indication of what size particles contributed to plugging, it still remained to be determined where in the cartridge the plugging was occurring. The experiment was repeated, but the eight micron particles were loaded onto just frits alone. In this manner, it could be determined whether plugging was occurring in the bed or in the frit. When the bed was not

59 present, no plugging occurred, indicating that plugging was in fact occurring in the sorbent bed itself. The next experiment was designed to determine if the particle size of the sorbent bed would have an impact on plugging. Five micrometer particles were loaded on both the polystyrene divinylbenzene and silica phases, which have nominal sorbent particle sizes of 70 and 50 microns, respectively. As in the previous experiment, when the five micron particles were loaded onto the polystyrene divinylbenzene cartridge, the particle passed right through the bed, and no plugging occurred. When the particles were loaded onto the silica cartridge, the bed became plugged. If a hexagonal close-packed arrangement of particle is assumed, it can be calculated that the diameter of a sphere that could fit between 70 micron particles in the sorbent bed is approximately thirteen microns, where the size that would fit between 50 micron particles is approximately eight microns. It must be considered that although the beds contain particles that are nominally 50 and 70 microns, there is a distribution of particle sizes. In addition, sorbent particles are not truly spherical. Despite these limitations, is can be seen from the approximations that it would be much more likely for nominally five micron particles to pass through a bed where the distance between particles is thirteen versus eight microns. It has therefore been demonstrated that the particle size of the sorbent can have a significant impact on the ease with which the extraction bed is plugged. In an earlier experiment it was demonstrated that larger particles would not plug the extraction bed if they could be retained above the frit. Such was the case for the fifteen micron particles that were held above a standard frit (20 micron) of a polystyrene divinylbenzene bed. The standard frits were replaced with ten micron frits. With this new configuration, the eight micron particles which had plugged the sorbent bed in earlier experiments, piled up on top of the new frit, and plugging did not occur. This experiment demonstrates that when small particles are prevented from becoming lodged in the spaces between larger particles, plugging can be avoided. An additional experiment was performed using five micron particles on a ten micron frit, and plugging did occur. A series of experiments were devised to investigate cartridge plugging by samples containing soil having a natural distribution of particle sizes. Soil samples of the Pima County Flood Plain were obtained which had been characterized with respect to their particle size distribution. Five hundred milligrams of soil were loaded onto cartridges containing 200 mg of the polystyrene divinylbenzene, and having 20 and 10 micron frits. Plugging occurred in each cartridge after 600 and 700 mL, respectively. A mechanism was needed through which the particles in the sample could be prevented from piling up on themselves during loading. A polyester web type depth filter had been used successfully to distribute oil and grease in previous the previous section. Experiments were designed to apply the same technology to distribute particulate m a t t e r above the sorbent bed to prevent plugging. Samples were loaded onto cartridges containing an integral depth filter above the sorbent bed.

70 These samples could be loaded in approximately a half an hour (averaging 30 mL per minute), rather than plugging the bed after 600 mL with no depth filter. The particles of soil could be seen to be distributed throughout the depth filter. Since a typical environmental analysis involves the determination of analytes which may be present in samples laden with particulates, it was desired to determine the impact that particulate matter would have on analyte recovery. Samples were spiked with 500 mg of Pima County Flood Plain Soil, and then with polyaromatic hydrocarbons at a concentration of two ppb. A set of five replicate samples were extracted on polystyrene divinylbenzene as described in the experimental section. Recoveries averaged 98%, with an eight percent relative standard deviation.

4.3. Utilization of layered phases for environmental analyses Layered phases can be used for environmental analyses to serve two different purposes. In one case, they can be used to extend the range of analytes that are extracted from the sample. Other use is to improve the selectivity for the analytes of interest.

4.3.1.

Layered columns for extending the range of analytes

The limitation of using a compromise phase such as C8 (as discussed in Section 3.1) is that although the overall recovery for the entire set of analytes is improved, the recoveries for individual components is not optimized. A subset of the EPA 525.1 analytes (the organochlorines) were utilized to determine the impact of layering bonded phases into a single extraction cartridge for the purpose of optimizing the recovery of both the polar and hydrophobic species. In the previous section, a limitation to optimizing analyte recoveries was described as being a two-fold problem. When extracting a broad range of compounds, the small, polar compounds yielded low recoveries on silica modified with short hydrocarbon chains. In the case of large, hydrophobic species, recoveries were low on ISOLUTE| due to poor elution of large compounds from a very hydrophobic surface. From this data it was hypothesized that a C2 phase could be layered over a C18 phase, allowing the larger compounds to be retained on C2, while the more polar compounds which passed through the C2 phase would be retained on the C18. If the larger compounds are sufficiently solvated, they could pass through the C18 phase during the elution step (Figure 21). Initial experiments were designed to determine to what extent a range of non-ionizable compounds would be stopped on each of two phases. The ISOLUTE | C2 phase was stacked in a six mL cartridge above the ISOLUTE | C18 phase. Each cartridge contained one half gram of bonded silica. The cartridges were connected using a Teflon adapter. Samples containing a mixture of organochlorine compounds were extracted. After the samples were loaded onto stacked columns, the cartridges were pulled apart and eluted separately. This allowed the selectivity for individual compounds by each phase to be determined. The recoveries from each of the phases are shown in Table 5.

71

Loading Layered Phases BroadRangeofAnalytes

Eluting Layered Phases StrongSolvent

m

02

C18

C2 C18 y

Figure 21. Loading and eluting layered phases.

Table 5 Recoveries of Organochlorines from ISOLUTE | C18 (EC) - 1 P P B spike C18 Std (EC) a l p h a - BHC 0.943 0.929 b e t a - BHC 0.895 0.961 g a m m a - BHC 0.957 0.967 d e l t a - BHC 1.021 0.970 heptachlor 0.993 0.825 *aldrin 0.987 0.745 heptachlor epoxide 1.002 0.963 endosulfan I 1.019 0.925 *4,4' DDE 1.016 0.705 dieldrin 0.999 0.941 *endrin 0.993 0.673 *4,4" DDD 1.007 0.811 endosulfan II 0.997 0.974 endrin aldehyde 1.005 1.121 *4,4" DDT 0.973 0.796 endosulfan sulfate 0.950 0.885 methoxychlor 0.958 1.532 Average % Recovery: 98 92 *Recoveries below 80%

It can be seen t h a t some of the analytes were well retained on ISOLUTE | C2, while others broke through the ISOLUTE | C2 cartridge and were stopped on the

72 ISOLUTE | C18 sorbent. Subsequently, the analytes were loaded onto a single cartridge containing the two phases layered, with ISOLUTE | C2 on top of ISOLUTE | C18, which could then be eluted simultaneously. For comparison, data was also obtained for analytes extracted from a single phase (Table 6). As was the case for the EPA Method 525.1 analytes in the previous section, 525.1 analytes in the previous section, several of the compounds were strongly r e t a i n e d on an ISOLUTE | C18 surface. For these analytes, recoveries improve on average from 75 to 84% when the extraction is performed on an ISOLUTE | C8 bonded phase. These results can be compared to the data obtained from stacked phases. While the average recovery for those analytes retained on ISOLUTE | C2 was 54%, and t h a t for the others which broke through the ISOLUTE | C2 cartridge and were stopped on the ISOLUTE | C18 sorbent was 43%, the combined recovery was 97%. This represents a 13% improvement in the overall recovery for these analytes extracted from the ISOLUTE | C8 "compromise" phase. W h e n the analytes were eluted from the C2/C18 layered phases, the average recovery was still quite high (97%).

Table 6 Recoveries of organochlorines from Top Cartridge: Bottom Cartridge: Std a l p h a - BHC 1.006 b e t a - BHC 1.009 g a m m a - BHC 1.010 d e l t a - BHC 0.985 heptachlor 0.995 *aldrin 1.006 heptachlor epoxide 1.024 endosulfan I 1.016 *4,4" DDE 0.998 dieldrin 1.010 *endrin 0.994 *4,4" DDD 1.001 endosulfan II 1.036 endrin aldehyde 1.016 *4,4" DDT 1.018 endosulfan sulfate 0.973 methoxychlor 1.007 AVG % Recovery: 101

stacked and layered columns (1 PPB spike) C2 layered over C2 C18(EC) C18(EC) top bottom Sum 0.998 0.971 1.011 1.011 0.990 0.972 1.011 1.011 0.974 0.952 1.000 1.000 0.909 0.945 0.969 0.969 0.985 0.940 0.851 0.073 0.924 0.861 0.877 0.846 0.846 1.005 1.009 0.574 0.469 1.043 1.077 1.149 0.652 0.417 1.069 0.842 0.886 0.856 0.046 0.902 1.004 1.015 0.801 0.207 1.008 1.111 1.098 0.669 0.340 1.009 0.942 0.966 0.898 0.061 0.959 0.866 1.073 0.648 0.372 1.020 0.858 0.932 0.111 0.837 0.948 0.985 0.933 0.849 0.849 0.817 0.882 0.292 0.575 0.867 1.019 1.027 1.115 1.115 96 98 54 43 97

The d a t a provides a solution for optimizing recoveries for a mixture of polar and hydrophobic analytes being extracted from an aqueous matrix. For compounds t h a t are difficult to elute from a very hydrophobic surface, they can be

73 extracted onto a sorbent from which they can be eluted layered over a more hydrophobic phase. These compounds the more hydrophobic phase once they are in a tetrahydrofuran), since they are well solvated, and it favorable for t h e m to r e m a i n in the solvent.

4.3.2.

more easily, which is are not extracted onto strong solvent (e.g., is energetically more

Layered columns for improving selectivity

Samples containing analytes and interferences with differing properties can be retained on different layered phases. As an example, a sorbent can be selected to specifically r e t a i n the interferences on the top layer, while the analytes are retained on the bottom layer. Elution conditions are then selected such t h a t the interferences continue to be retained while the analytes are eluted (Figure 22).

Elution

Loading

Analyte (~ Interferences

0

J

Figure 22. Interference removal.

One of the p r i m a r y interferences in the analysis of e n v i r o n m e n t a l samples are humic substances. If a single phase is used for the extraction of a particular class of pesticides, the humic substances will be extracted and eluted along with the analytes of interest. The acidic character of humic substances can be exploited to r e t a i n these interferences, while the analytes are retained and eluted through hydrophobic interactions. An example of such a column is aminopropyl (NH2) layered on to of C18. The humic substances are retained on the NH2 phase, while the analytes, such as polyaromatic hydrocarbons or organochlorines, are retained on the non-polar phase. During the elution step using an organic solvent, the humics continue to be retained on the top phase, while the analytes of interest are eluted.

74 5. C H A P T E R S U M M A R Y

This chapter has provided a discussion of the basic principles of solid phase extraction, some of the physical and chemical properties of solid phase extraction sorbents, and a detailed description of a typical solid phase extraction procedure. The six steps of an SPE procedure included sample pre-treatment, column solvation, column equilibration, sample loading, interference elution, and analyte elution. In addition, a variety of factors that influence the efficiency and selectivity of an SPE procedure were presented. The silica surface has been shown to be quite heterogeneous in nature, resulting in species that may be retained through multiple interactions. It was shown that elution solvents could be selected to improve recoveries of analytes that are retained through multiple interactions when the composition of the elution solvent was mixed, rather than using pure solvents for the elution. The porous nature of the sorbent has been shown to have an influence on analyte recovery. It was seen that conditions such as sample temperature and residence time of the elution solvent have an impact on analyte recovery. The former can have an influence the distance to which analytes migrate into the pores, while the later dictates whether the analytes have sufficient time to migrate back out of the pores. Another important factor in solid phase extraction includes the chain length of the bonded phase and its influence on selectivity. As the chain length of the bonded phase increased, the retention of smaller, more polar compounds improved. It was also noted that the elution of larger, more hydrophobic compounds became more difficult with increasing chain length. A second impact of chain length was seen in the propensity for the sorbent to retain water. A sorbent that had been modified with a long chain such as C18 was effectively "water proofed", and could be dried quickly (10 minutes for a one gram bed). The C2 phase, having a significant number of easily accessible silanol groups and less hydrocarbon in the pores required significantly longer drying times (25 to 30 minutes). The presence of water and the impact on recoveries was seen to be most significant when the elution is performed with a water immiscible solvent. In this case, water in the pores of the sorbent formed a barrier between the analytes and the elution solvent. When the elution solvent is switched to one that is water miscible, the surface is essentially "chemically" dried, and analyte recoveries are improved. One of the physical parameters of the sorbent that has an influence each step of the solid phase extraction procedure was shown to be particle size distribution. A wide distribution of particle sizes and the presence of fines result in poor flow characteristics through the column. This leads to an increase in drying times, broader retention bands and an increase in required solvent volumes for conditioning, equilibration, interference elution and analyte elution. The addition of a wetting agent was seen to serve two purposes. An organic solvent such as methanol is necessary when loading large volume samples onto a hydrophobic phase to ensure that the conditioned chains remain extended and

75 available for interaction with the analyte. It has also been shown that a wetting agent help to keep the analyte from agglomerating, which prevent migration into the pores. When the pores are unavailable, the surface area of the sorbent is reduced from 550 to 0.1 meters 2. SPE has been shown to have very useful applications to environmental analyses. It has been used as an alternative to liquid-liquid extraction for Oil and Grease determination. Polar and non-polar fractions can be selectively eluted by judicious choice of elution solvent. Solid phase extraction can also be use to elute a broad range of analytes by using a medium chain length non-polar phase such as C8, or layered phases such as C2/C18. The C8 phase resulted in recoveries that were improved overall, but not optimized for each component. The layered column improved recoveries for both the small, polar and large, non-polar species. The application of a depth filter for the extraction of environmental sample was found to be useful in the prevention of column plugging. Loading times were significantly reduced, and analyte recoveries were not negatively impacted.

ACKNOWLEDGEMENTS The authors would like to thank International Sorbent Technology (IST) for their financial and technical support. REFERENCES 1. L.A. Berrueta, B. Gallo and F. Vicente, Chromatographia, 40 (1995) 474. 2. E. Chladek and R.S. Marano, J. Chromatogr. Sci., 22 (1984) 313. 3. C.F. Simpson, Techniques in Liquid Chromatography, John Wiley and Sons, New York, 1982. 4. S.A. Wise and W. E. May, Anal Chem., 55 (1983) 1479. 5. I. Liska, J. Krupcik and P. A. Leclercq, J. High Resolution Chromatogr., 12 (1989) 577. 6. K.C. Van Horne, Sorbent Extraction Technology, Analytichem International, Harbor City, CA 1985. 7. R.W. Frei and U. A. Th. Brinkman, Trends in Analytical Chemistry, 1 (1991) 45. 8. IBID. 9. M.E. Raisglid, Factors Affecting the Selectivity and Efficiency of Solid Phase Extraction, Ph.D. Dissertation, Department of Chemistry, University of Arizona (1997). 10. K. G. Furton and J. Rein, Anal. Chim. Acta, 236 (1990) 99. 11. C. A. Fung Kee Fung, The Behavior of Water at the Modified Silica Interface, Ph.D. Dissertation, Department of Chemistry, University of Arizona (1992). 12. EPA Document #600/4-88/039, Methods for the Determination of Organic Compounds in Drinking Water by Liquid-Solid Extraction and Capillary Column Gas Chromatography/Mass Spectrometry Revision 2.2 (1988).

Adsorption and its Applications in Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski (Editor) 9 1998Elsevier Science B.V. All rights reserved.

77

S e l e c t i v e r e t e n t i o n , r e m o v a l a n d e l u t i o n for a n a l y s i s of h a z a r d o u s c o m p o u n d s i n b i o l o g i c a l f l u i d s to m a i n t a i n h u m a n h e a l t h H. Shintani National Institute of Health Sciences, Ministry of Health and Welfare of Japan, 1-chome, 18-1, Kamiyoga, Setagaya, Tokyo, J a p a n 158

SUMMARY Selective analysis of compound of interest in complicated matrix such as body fluids is extremely difficult. This is because unexpected interference admixtures with almost identical retention time may co-elute with overlapping. It is essential to remove admixtures by appropriate pretreatment for selective analysis. For pretreatment methods, there are several methods, i.e. solid phase extraction (SPE) using column or membrane, dialysis, filtration, ultrafiltration, super fluid critical extraction (SFE) or adsorption using charcoal or other appropriate adsorbent to remove admixtures as well as toxins in case of t r e a t m e n t for h u m a n body. Some of them mentioned above were carried out using a packed column or membrane. In general, membrane type has more capacity t h a n packed column type. At a hospital, artificial dialysis and filtration are used for replacement of part of kidney function (artificial kidney) and several adsorbents are used for replacement of part of liver function (artificial liver). H u m a n liver function is very complicated, so in the current medical treatment status only removal function in liver function can be anyhow attained. These pretreatment methods are also useful for analytical procedure. There have been reported several pretreatment methods for acidic, basic and neutral compounds in biological fluids. The column injectable directly the body fluids is also available in the market, however this column has restriction of concentration of organic solvent in the eluent, thus resulting in poor or no elution of strongly hydrophobic compounds. Artifact formation must also be considered during pretreatment methods, especially when using liquid-liquid extraction because of long period contact with solvent and the compound of interest. Artifact formation may cause lower recovery rate of the compound of interest and also may cause misunderstanding to the researcher, which he/she extracts unexpected highly toxic compound extraction. According to the recent advancement of analytical column fabrication technology, several new columns based on individual idea are now available in the market. Some of them diminish completely residual silanol effect, thus no common ion addition for strong basic or acidic compounds or no ion-pair method was required. Ion-pair

78 method is so popular, but it often causes to shorten column life by retaining and deteriorating with ion paring reagent. These columns are also described by comparing conventional columns in this chapter. 1.

INTRODUCTION

Solid phase extraction (SPE) using packed columns combined with high performance liquid chromatography (HPLC) for analysis of several toxic compounds in blood was described. SPE and liquid-liquid extraction were also compared in terms of recovery efficiency, solvent consumption, handling time, artifact formation and so on. Membrane type SPE for large scale extraction, especially applicable for environmental analysis, will also be described. Membrane type SPE was not successfully applicable to biological fluids pretreatment. This is mostly due to stacking of membrane pore, therefore it was so often restricted to environmental analysis treatment in such a case of analyzing trace residue of pesticide in water. In this chapter, the author will describe major three substances from his recent papers including unpublished results. They are 4,4'methylenedianiline (MDA) from irradiated polyurethane (PU) for sterilization purpose, residue of several toxic compounds including newly identified compounds from dental materials and blood urea and uric acid. The first toxic, carcinogenic and basic compound is MDA [1-7]. As a medical PU, thermosetting PU was mostly used. For example, potting material for connecting dialysis fibers with outer vessel in artificial dialysis equipment was made from thermosetting PU. In fabrication of thermosetting PU, polyol, methylenediisocyanate (MDI) and butanediol were mixed and terminated polymerization with addition of butanol to obtain appropriate molecular weight of thermosetting PU. Residual MDI changed to MDA by hydrolysis. MDA also produced by cleavage at urethane linkage by irradiation for sterilization [1-7]. MDA in blood was analyzed to assess h u m a n risk contacted with degradated thermosetting PU medical devices during artificial dialysis treatment to patients [1-7]. The thermosetting PU medical devices may have a possibility to be degradated upon gamma-ray irradiation for sterilization. MDA was determined by reverse-phase HPLC using an eluent of a mixed solution of ammonium acetate and acetonitrile at a ratio of 713 (v/v) [8-13]. MDA was reported to be unstable upon heating and converted to MDI, therefore HPLC is considered to be superior to gas-liquid chromatography (GLC) for MDA determination. At that time when the author reported MDA determination in 1989 to 1991 [8,11], only conventional endocapped C-18 column was available in the market, thus that kind of column was used for the study. Conventional endcapping procedure remained many untreated residual silanol (around 80%), thus indicating as a result an incomplete endocapping. Detection was by an electrochemical detector (ECD, amperometry detector from Toadenpa Co. in Tokyo) as well as ultraviolet (UV) detector. ECD is around ten times more sensitive and selective than UV

79 detection, however ECD detection was restricted to the compounds with relatively lower oxidation-reduction potential (less than 1000 mV) such as aromatic amine, OH, SH or aliphatic SH or OH such as cystine or sugar, respectively. Baseline of UV detection was fluctuated due to impurities in the eluent, however ECD detection indicated flat baseline (Figure 1) [11]. Detection limit of MDA by UV at 250 nm and by ECD at 900 mV was 150 and 3 ng/ml, respectively [11 ].

0

I

I

I

I

I

I

25

20

15

10

5

0

min Figure 1. HPLC chromatogram of MDA detected by UV at 250 nm for upper and ECD at 900 mV for lower.

Concerning pretreatment of compounds in body fluid, several methods were compared in terms of their efficiency, recovery and other factors. Liquid-liquid extraction was troublesome due to deproteinization followed by centrifugation and condensation. During condensation there may be a possibility of recovery loss due to evaporation. In this sense, SPE was thought to be superior. Manual type SPE and an automated SPE were also compared in terms of reproducibility of recovery rate. Furthermore, artifact formation described lather in detail must also be considered when using liquid-liquid extraction. Blood MDA was satisfactory recovered using C-18, Phenyl and Cyclohexyl SPE columns Figure 2 indicating major retention mechanism may be van der Waals binding as well as ~-~ binding. The recovery of serum MDA from C-l, C-2, C-8 and Silica columns was unsatisfactory (Table 1). Elution was carried out using methanol containing 1 M NH4OH. In the preliminary SPE experiment based on the theoretical consideration, conventional technique and reported procedure for basic compounds in case of SPE treatment was that acidified methanol (methanol containing 1 M HC1) as an eluent was thought to be superior, however the

80

00

of. ~:~

o

X 9

0

r..)

r~

(~

O']

.r,,4

.v..4

r,.)

0 0

0

0

~

Figure 2. HPLC chromatograms of blood MDA after C 18, C 8, Phenyl and Cyclohexyl column solid phase extraction treatment.

Table 1 Addition-recovery experiment of serum M D A by SPE using silica and reverse p h a s e columns Resin

Added M D A (pg)

Found MDA (lag)

Recovery (pg)

Silica

1

0.12

12

C-1

1

0.56

56

C-2

1

O.75

75

C-8

1

0.90

90

C-18

1

1.00

100

Cyclohexyl

1

1.00

100

Phenyl

1

1.00

100

Strong

1

1.00

100

Cation exchange Fifty pl of 21 pg/ml MDA spiked to one ml serum to prepare one ~g/ml serum MDA. This was applied to the conditioned resins. The following procedure was the same as in the text. The amount is the average of 5 specimens and the C.V. was less than 1.4% in every case.

81 e x p e r i m e n t a l result done by the a u t h o r obtained the opposite result in t e r m s of recovery r a t e (Figure 3). The speculated explanation for the m e c h a n i s m will be described in the text. The result suggested the i m p o r t a n c e to consider simultaneously the silanol effect and chemical s t a t u s of compound of interest. SPE was superior to a liquid-liquid extraction due to an u n n e c e s s a r y of

,-.

0 0

,,-;

J

0 ~

~

C',l

S

0 0

o

oO

o

o

o

o

,,<,

Z

b

0 o r,_)

0

0

0

.=

~

o

0=~

0

9 -

ib ~

= .,-

b

0

o

0

o~

b 0

-~0

o

o~

--

E

Figure 3. HPLC chromatograms of MDA treated with different eluents in SPE using C 18 column. HPLC conditions: C 18 column, eluent: methanol and an aqueous solution of 10 mM ammonium acetate at a ratio of 1/1, flow rate: 1.2 ml/min, detection: 254 nm. MDA peak has a retention time of 6.95 min. (1) 105 ng/10 ~tl MDA standard solution, 10 ~tl applied to HPLC, (2) 100 ~tl of 105 ng/10 gl MDA solution applied to C18 resin with !00 mg and 120 gl, respectively, resin weight and void volume, and eluted with 250 gl of methanol, 200 pl collected and 10 pl applied to HPLC. When 100% is recovered, the theoretical concentration of MDA would be 52.5 ng/10 ~tl. (3) The same procedure as above except for the elution, 250 gl of a mixed aqueous solution of 10 mM ammonium acetate and methanol at a ratio of 1/1 (v/v). (4) The same procedure as above except for the elution, 250 pl of methanol containing 1M NH4OH. (5) 50 ~1 of 21 ~tg/ml MDA was added to 1 ml serum to prepare one/ag/ml serum. This was applied to C 18 resin, eluted with 250/al of methanol containing 1M NHaOH. 200 gl collected. Ten gl applied to HPLC. If 100% is recovered, the theoretical concentration would be 50.0 ng/gl. (6) The same procedure as above but omitting MDA, the serum blank.

82 C1, 2, 8, 18, Phenyl, Cyclohexyl and Silica (100 mg resin weight, 120 ~1 void volume) 2 ml methanol 2 ml water appl 0.1 - 10 ppm MDA in serum eluted with 250 pl MeOH/1M NH40H collected 200 pl u

10 ~1 appl to HPLC

Satisfactory recovery rate obtained by the use of C18, phenyl and cyclohexyl columns

Figure 4. SPE procedure for blood MDA.

deproteinization, centrifugation, condensation, greater recovery, less consumption of organic solvent, less experimental time and less possibility of artifact formation. SPE procedure of blood MDA was described in Figure 4. The second toxic compounds were those from polymethyl methacylate (PMMA) dental material [14-18]. PMMA is widely used as the composite resin for the dental plate. In accordance with the current PMMA fabrication, benzoylperoxide (BPO) and N,N-dimethyl p-toluidine (DMPT) were added as the initiator and the stimulator, respectively, for methyl methacrylate (MMA) polymerization. If insufficiently polymerized, MMA monomer, starting compounds of DMPT and BPO exhibit a residue potential. Especially residue of MMA was significant at around 1-2%. When considering 1-2% of toxic compound of MMA will be eluted into patients during dental treatment, this residue should be diminished for patient's sake. As one proposed method was rinsed with hot water as MMA is hydrophilic. Rinsing with organic solvent such as with methanol must be avoided as MMA dental plate was discolored and deformed. As an additional information, newly found compounds, which have not been reported so far, were identified using HPLC-mass spectrometry (MS)-MS. One of newly found toxic compounds was epoxide compound of DMPT. Epoxide compound was quite reactive to DNA, so most probably mutagen and carcinogen as is ethylene oxide. BPO was converted to benzoic acid (BA) in a few seconds when contacting with DMPT or body fluids such as blood or saliva, therefore BA analysis has an identical meaning to BPO determination in body fluids. BA was not originally used for PMMA fabrication, indicating BA is a sort of artifact from BPO when contacting with body fluids. BPO is quite reactive compound and both

83 BPO and BA are cytotoxic [17]. For determining cytotoxicity, s e r u m was added in cell culture medium, therefore BPO transformed to BA immediately, which was at first confirmed by the author. From this result, cytotoxicity data reported as BPO is confirmed not the exact cytotoxicity data of BPO, but the data of BA. The cytotoxicity data of I C 5 0 (~g/ml) of BA and BPO using Balb 3T3 cell was 28.7 and 22, indicating both are almost identical because BPO data was originally from BA. In order to evaluate the risk factor to the recipient exposed to these compounds from dental material, the authors quantitatively analyzed residual a m o u n t in composite resin using blood serum extraction [16-17]. MMA and BPO are unstable upon heating, therefore HPLC is considered to be superior to GLC. Determination was carried out by HPLC combined with SPE using C-18 columns in both columns. The comparison of eluents of these compounds from C-18 SPE column was discussed. The comparison of SPE and liquid-liquid extraction in terms of recovery efficiency was also discussed. The third toxic compound is blood urea and uric acid. Urea is a major uremic toxin, however urea accumulation is a trigger or result of u r e m i a is not well clarified yet. This m e a n s whichever blood urea m a y cause u r e m i a or blood urea may accumulate after u r e m i a promotes is currently uncertain. In h e a l t h y person, protein t e r m i n a l l y metabolites to urea as a final product through urea cycle and exclude urea to urine through kidney exclude function. U r e m i a patients can't successfully exclude urea to urine as kidney function of the patients damages, therefore urea accumulates back to blood. Urea is a final compound in urea cycle, so accumulation of final product will cause feedback inhibition, thus leading to irregular protein metabolism. There are two types of urea. One is free from protein and the other is bound to protein type. It was reported around 3% of total urea will bind to albumin. Free urea plays an i m p o r t a n t role for promoting uremia syndrome, therefore accurate as well as differential analysis of bound from free urea in blood are i m p o r t a n t to a t t a i n precious diagnosis. Current clinical urea analysis is carried out using an a m m o n i u m selective electrode attached immobilized urease onto the electrode to determine a m m o n i u m as urea, which is so-called blood urea nitrogen, BUN. The alternative method for clinical test carried out at the hospital is autoanalyzer to use Indophenol colorimetry. When using these methods, differential analysis of endogenous a m m o n i u m and urea was not attained. Additionally, differential analysis of free from bound urea can't also be attained. Autoanalyzer used in the hospital as a routine analysis determines total urea as BUN by converting urea to a m m o n i u m with urease and react with Indophenol reagent to detect colorimetry with visible detection. Amount of endogenous blood a m m o n i u m can be neglected due to around 1% of total blood urea, which will be within analytical error, however t h a t of urine a m m o n i u m can't be neglected due to more t h a n 10% of urine urea. There have been reported so far several methods for blood urea analysis [19-22]. Urea analysis using capillary electrophoresis (CE) (mostly using micellar electrokinetic chromatography, MECC or MEKC) was also

84 reported [23-24], but separation of urea from blood admixtures was not satisfactory as urea was migrated at around void volume overlapped with hydrophilic blood admixtures, which was identical to urea analysis using C-18 HPLC column [19-22]. Blood urea was poorer retained on C-18 column and also eluted at around void volume overlapped with blood hydrophilic admixtures as being reported already and blood uric acid retained on successfully, but not separated well from blood admixtures of mostly blood proteins [19-22]. That's why simultaneous blood urea analysis with other blood compounds was so difficult by HPLC or CE. Single urea analysis in blood can be satisfactory attained using capillary zone electrophoresis (CZE). However, for simultaneous analysis of urea together with other compounds such as uric acid in complicated matrix such as blood will require complicated equipment system as previously done by the author using post-column Indophenol colorimetry method or precolumn method combined with column switching and immobilized urease column [20-22]. This is the reason severe selection of appropriate column for HPLC analysis is required. The separation mechanism of reverse phase HPLC column and that of MECG is almost identical [19-24]. To be free urea from blood admixtures is the major purpose of blood pretreatment. This is for selective analysis of blood urea free from blood admixtures, thus the author describes mainly for blood urea analysis as follows: what analytical condition (selection of sort of analytical columns) and what kind of pretreatment methods will be most appropriate for attaining selective analysis of blood urea as well as differential analysis of endogenous blood urea from endogenous ammonium. Direct blood injection is applicable to CE and to HPLC, however there always associates a possibility of deterioration of column and detector cell with blood protein or hydrophobic lipid as well as overlapping of compound of interest with blood admixtures [23], therefore it is more appropriate to avoid direct blood injection. Direct blood injectable HPLC column can be currently available in the market, however this column has a restriction of less than 18% of acetonitrile as an eluent, otherwise deproteinization will occur in the column and this may deteriorate column and make shorten column life. This kind of restriction of organic solvent concentration in the eluent indicates highly hydrophobic compounds can't be successfully eluted. When handling body fluids (blood or saliva), it is recommended to remove and isolate admixtures by p r e t r e a t m e n t to attain significant baseline separation free from body fluid admixtures, reproducible chromatograms and prolonged life of analytical column and detection cell. As p r e t r e a t m e n t methods, traditional liquid-liquid extraction, SPE, dialysis, ultrafiltration or supercritical fluid extraction (SFE) were compared in terms of recovery efficiency and separation efficiency. The obtained result was not always restricted to blood urea analysis, but also applied to blood hydrophilic compound analysis. There have not been published papers so far on blood urea analysis using SPE combined with HPLC because author's paper was the first on this subject. In past, only one paper have been published on urea analysis in food

85 using SPE combined with HPLG by Fujiwara et al. [25]. After tracing the reported method, it can be found out that the satisfactory recovery rate can't be attained using their method (less than 80% recovery). It must be recognized that some researcher may satisfy with 80% recovery, but problem is that it is not certain 80% recovery will be reproducible or not when concentration of the compound of interest in matrix will differ. It should be kept in mind that reproducibility is essential and recovery rate may so often differ depending on concentration. That's why the author required 100% recovery using different sample concentration. Other reason is an artifact formation, which will be discussed later in detail. The reason to be considered for the major difference of recovery rate between the author's experiment and that of Fujiwara's experiment might result in different matrix of blood or food, respectively [25]. Therefore, the author studied for pretreatment method to attain 100% recovery for blood urea analysis. Differential analysis of free from bound blood urea can also be attained using ultrafiltration or dialysis for native blood. The amount of bound urea was not significant at around 3% that of total blood urea. When using denatured blood, total urea amount can be determined in ultrafiltrate after ultrafiltration at around 13,000 g for 40-60 minutes. Denaturation was attained by acidification of blood to make bound urea free from protein. Endogenous ammonium in blood and urine was around 1% and 10% of total urea, respectively, therefore the former was within an analytical error. As the latter is beyond the analytical error, determination of endogenous urine ammonium will be required for differential analysis of endogenous urine urea.

Q

S E L E C T I V E R E T E N T I O N , REMOVAL AND E L U T I O N F O R ANALYSIS OF TOXIC C O M P O U N D S TO HUMAN H E A L T H

Blood and saliva used are sampled from the author. The saliva was sampled before breakfast from the author. Most of chemicals excepting compounds synthesized by the author in this chapter were available in the market with several grade. Pretreatment method described in the following text is applicable to analytical procedure as well as clinical treatment for h u m a n health as being mentioned in summary. In this paper, solid phase microextraction was omitted as the hardware of the equipment has still inferiority to be innovated and remained to be improved for attaining reproducible recovery rate.

2.1. B l o o d M D A 2.1.1. A n a l y t i c a l c o n d i t i o n s of M D A AS mentioned in advance MDA is unstable upon heating, therefore HPLC was adopted for analysis combined with ECD detection. The eluent is a mixture of an aqueous solution (3 parts) containing 50 mM ammonium acetate for common ion effect as well as increasing ECD sensitivity and acetonitrile (1 part) by volume

86 ratio. This method was found to improve ECD detection sensitivity, prevent MDA tailing and accelerate MDA elution [8-13]. At that time when this experiment was carried out in 1989 to 1992, no columns completely diminished silanol effect was available. In the current market, innovated and completely diminished silanol effect columns are easily available in the market, so using t h a t kind of column above mentioned experiment was carried out again for detection efficiency comparison as being mentioned later. An addition of ionic compound at around 50 mM is essential for ECD detection even if newly innovated columns free from silanol effect are used. This is not for attaining conventional common ion effect, but for improving ECD detection. If no salt contained in the eluent, no successful ECD detection was done. ECD requires at least 50 mM salt in the eluent for electricity delivery in the eluent. ECD sensitivity increased increasing applied voltage, but simultaneously decreasing selectivity, indicating less t h a n 1000 mV will be most appropriate (Figure 5) [12].

f

r

J I

200

I

250

UV spectrum of MDA

I

300 nm

0.4

I

I

I

I

I

I

I

0.5

0.6

0.7

0.8

0.9

1.0

1.1

Applied voltage vs Ag/AgC1 (V)

Figure 5. UV spectrum of MDA in the upper and the relationship between applied voltage of ECD and response in MDA detection.

Thus, there are two reasons to add 50 mM to eluent. One is for common ion effect [8-13] and the other is to improve ECD detection sensitivity of aromatic amine. When using insufficiently endocapped ODS columns such as Zorbax | ODS or ODS columns from Merck Co. or Toso Co. were used, at least 50 mM addition will be required for common ion effect and simultaneously for the benefit of ECD detection sensitivity. MDA was detected with ECD at 900 mV as well as UV at 250 nm. The latter is for benzene ring absorption in MDA. As mentioned in advance ECD detection was more selective and sensitive. When a m m o n i u m acetate in the eluent contained impurities, the baseline by UV detection was fluctuated, however no fluctuation was observed by ECD, indicating quite selective for detection (Figure 1) [12]. Consequently, ECD was superior to UV detection. The detection limit of MDA by UV and ECD was 150 and 3 ppb

87 (S/N = 3 at the peak height) [12], respectively, indicating the greater sensitivity by ECD detection. ECD detection has inferiority that it can detect only for the compounds with relatively lower (mostly less t h a n 1000 mV) oxidation-reduction voltage potential as mentioned in the introduction. Greater voltage application may cause any damage to the electrode in short period. The conventional endocapped column was treated with methyl, however residual silanol was not completely diminished. It was reported around 80% of silanol still remained after endocapping treatment. This is mostly due to remaining silanol existing in the interior of the pore of silica. As mentioned in advance, due to recent column technology advancement, residual silanol effect have been completely diminished by employing a silicone coating to eliminate the residual silanol effect (e.g., Capcel Pak | from Siseido Co. in Tokyo) or free silanol diminishment technology for column fabrication at high temperature such as L column | from Chemical Inspection Co. (Tokyo) to prevent a residual silanol effect as far as possible. The latter column, therefore, due to its complicated fabrication technology at high temperature, no large scale column is available as it is too expensive. These columns of Capcel Pak | and L column | are completely free from residual silanol and heavy metal impurities in silica. The study of detection limit using these columns indicated at least 10 times lower than that by insufficiently endocapped ODS columns used in previous study due to narrower peak shape and higher peak height. The reason to be considered is due to no residual silanol, no residual heavy metals and even particle size of round silica at around 2 or 3 ~m diameter. This resulted in prevention of basic compound tailing without any addition of common ions for elution as is required in the conventionally endocapped [14-18]. The elution of strong basic compound, N,N-dimethyl-p-toluidine (DMPT), can be successfully eluted without common ion addition [14-16]. 2.1.2. R e s u l t o f l i q u i d - l i q u i d e x t r a c t i o n vs. S P E for b l o o d M D A Liquid-liquid extraction procedures have been reported describing the repeated extraction with n-heptane-isoamyl alcohol (99:1, v/v), diethyl ether or benzene from alkalized serum in order to prevent MDA dissociation, however the reported recovery rate was unsatisfactory (from 70 to 80%) [26-28]. The author can admit this recovery rate even if it is lower as far as the recovery rate has reproducibility, but it will not be certain for reproducible recovery rate in any case as mentioned in advance. That's why the author mentioned unsatisfactory recovery rate. This is not always for lower figures of recovery rate. Moreover, repeated procedure requires time-consuming, loss of recovery, troublesome handling, copious consumption of organic solvents, hazardous to the handling person or condensation resulting in lower recovery rate due to entrapping loss. In the author's preliminary procedure for liquid-liquid extraction, one part of alkalized serum was extracted twice with seven parts of a mixture of chloroform and methanol at a ratio of 3/1 (v/v) and a satisfactory recovery of blood serum MDA was attained (98% in average, n=3) [11]. This procedure was often used for

88 lipid extraction in biochemistry, but this indicated troublesome procedure. Next, the author innovated a more efficient method, in which only two to three parts of acetonitrile was added to one part of blood serum at a volume ratio for deproteinization and MDA extraction. This procedure required only single extraction and showed a satisfactory recovery of blood MDA (99% in average, n=5) (Figure 6) [8-10]. The reason why acetonitrile was thought to be satisfactory is because acetonitrile is an efficient deproteinization reagent as well as an efficient solvent of MDA. Most liquid-liquid extraction procedures for blood were required centrifugation after deproteinization, supernatant condensation of excessive organic solvent and the repeated extraction. The condensation by vacuum evaporation was undesirable due to MDA instability upon heating and loss of recovery [12-13], therefore SPE for blood MDA was studied as an alternative and more reliable pretreatment method.

Serum p r e t r e a t m e n t using l i q u i d - liquid extraction method One volume serum add three volumes CH3CN shaking for 60 min centrifuge at 4000 rpm for 20 min

The supernatant collected

vacuum evaporation at 50~ redissolved in one ml MeOH

20 ~1 applied to HPLC Figure 6. Liquid- liquid extraction method for blood MDA.

In general, SPE is easy to handle, to prepare for an automated system and requires less solvent consumption, deproteinization, centrifugation, and condensation are usually unnecessary. It is more important to keep in mind that one cycle t r e a t m e n t with SPE will be almost identical to several tens of thousands of liquid-liquid extraction treatment. Therefore, SPE t r e a t m e n t of blood MDA was speculated to be superior to liquid-liquid extraction method. The comparison of the eluent selection of several SPE columns was examined [8-9]. It is important that so often condensation procedure is unnecessary as far as the

89 eluent volume was less than the added volume to SPE column. Eluent volume should be at least two to three times more than void volume of SPE column, otherwise no successful elution attained. For example, 100 mg of Bond Elut | C18 SPE resin from Varian Co. (Harbor City, CA) has 120 pl, thus at least 240 to 360 ~1 as an eluent volume will be required for successful elution. In the elution of blood MDA, SPE columns examined were Bond Elut | C-l, C-2, C-8, C-18, Phenyl, Cyclohexyl, Silica, and strong cation-exchange (SCX) for recovery of blood MDA (Table 1) [8-9]. Reverse phase columns (C-l, C-2, G-8, G-18, Phenyl and Cyclohexyl) with a resin weight and void volume of 100 mg and 120 ~1, respectively, were used. These columns were conditioned with 2 ml of methanol and rinsed with 2 ml of water. Thereafter, one ml of serum was applied to the conditioned columns and rinsed with water. The columns MDA retained were eluted with 250 pl of methanol containing 1 M NH4OH. The drain was trapped and 10 pl were applied to HPLC (Figure 4). Conditioning, rinsing and elution were carried out by a vacuum system using Iwaki Co. vacuum pump from Tokyo [8-9]. C-l, C-2, C-8 and Silica columns indicated lower MDA recovery rates, whereas C-18, Phenyl and Cyclohexyl columns produced satisfactory recovery rates (Table 1). The lower recovery rates by C-1, C-2, C-8 and Silica columns and higher recovery rates by C-18, Phenyl and Cyclohexyl columns were speculated that the van der Waals binding and x-x interactions between benzene rings would be major factors for retention of MDA to SPE columns, while the binding of MDA to free silanol was not major due to lower recovery rate by Silica column (12% recovery, Table 1). This was thought to be due to water in serum interfering the combination of MDA to silanol. Recovery rate among C-l, C-2 and C-8 columns increased with increasing hydrophobicity of the columns (56% recovery for C-l, 75% for C-2 and 90% for C-8), which was thought to be due to increase of van der Waals binding capacity [8-9]. The cation exchange column produced a sufficiently recovery rate (Table 1), but requires a more complicated conditioning procedure t h a n a reverse phase column. The inferiorities are acidification of blood to charge MDA positively and centrifugation prior to SPE t r e a t m e n t [8-9]. MDA bound to the reversed phase column was not sufficiently eluted by methanol alone or HPLC eluent (a mixture of 3 parts of 50 mM ammonium acetate solution and 1 part of acetonitrile by volume ratio), but was almost completely eluted by a strongly alkalized methanol solution (a mixture of methanol and 1 M ammonium hydroxide) (Figure 3) [8-9]. The basic MDA bound to the reversed-phase column was normally speculated to elute more favorably after MDA was acidified to charge positively MDA. Based on this speculation, the author initially carried out an elution experiment using an acidified methanol. However, the results were opposite to the speculated intention, indicating the recovery rate was lower than alkalized methanol (recovery rate for methanol alone was 6.8%, acidified methanol was 10.2% and alkalized methanol was 100%, Figure 3).

90 The precise reason why the alkalized methanol was superior to the acidified methanol was not yet well clarified, which was the major subject in this chapter to consider, the author's speculation based on experimental result are: alkalization will depress MDA charge and stimulate silanol dissociation and acidification will be opposite. From this, of which amine charge in MDA or silanol charge will play more essential role for elution. According to the recovery result of the silica column (12% recovery), binding of MDA to silanol will not be significant. Acidified MDA promotes charging amine of MDA, but this does not result in favorable recovery (recovery rate of 10.2%). Therefore, it will be concluded that both amine and silanol will not be essential factors to be considered. Finally, the author considers the satisfactory result by alkalized methanol may be due to common ion effect. If so, elution mechanism is quite simple t h a n ever been speculated, however initial speculation method, which is opposite to common ion effect, was so often reported for recommendable SPE eluents. F u r t h e r speculation is that acidification was insufficient to reduce blood pH due to blood buffer function at pH 7.4, positively charged MDA may partly bind to dissociated residual silanol in SPE column (this was not completely confirmed and this was from lower recovery by acidification) or positively charged MDA may not be sufficiently dissolved in hydrophobic eluent. If the last reason will be correct, why favorable result by alkalized methanol was attained successful result can be clarified. That's the undissociated MDA will be more favorably dissolved in hydrophobic eluent. Thus, the author considers two major reasons for favorable result using alkalized MDA. One is for favorable dissolution to the organic solvent (methanol) eluent of MDA and the other is for common ion effect. These speculation might not be correct, but the experimental result has a significant reproducibility, therefore nobody can deny the importance of the author's finding from the reproducible experimental result. From this speculation when determining optimum elution condition, we should simultaneously take into consideration the following factors: chemical charge situation in the eluent for the compound of interest, chemical and physical properties of the eluent and SPE column characteristics and the behavior of residual silanol. The above indicates many factors must be simultaneously considered for favorable retention and elution.

2.2. Residual toxic c o m p o u n d s in blood from PMMA dental m a t e r i a l s 2.2.1. Analytical c o n d i t i o n s 2.2.1.1. MMA, DMPT and BPO analysis The column for MMA, DMPT and BPO analysis was Capcell Pak | C-18 SG-120 from Siseido Co. in Tokyo. This column was completely endocapped with silicone coating as being mentioned in advance. The eluent was a mixed solution of water and acetonitrile at a ratio of 1/1 (v/v). No common ion was added. The flow rate was 1.2 ml/min. Detection was UV at 235 nm [14-17].

91

2.2.1.2.BA analysis BA analysis was as follows: a Capcell Pak | C-18 AG-120 column was used with an eluent of a mixture of acidified aqueous solution of water and acetonitrile at a ratio of 4/1 (v/v) adjusted to pH 3 with acetic acid. Acidification is for BA depression, otherwise BA will be eluted before void volume. Detection was by UV at 235 nm. The rest of the procedure was identical to MMA, DMPT and BPO analysis.

2.2.1.3. Newly found toxic compound analysis Methanol extract of dental material fabricated at room temperature of Yunifast | from GC Co., Tokyo, was used for unidentified compound analysis. Using the gradient elution of HPLC shown later in this section, many compounds including MMA, DMPT and BPO were eluted (Figure 9). Total peaks eluted were not completely identified, but identified compounds were shown as follows: aniline, N-methyl p-toluidine, p-toluidine, BA, 3-carboxy 4N-methyl amino toluene, 2-carboxy 4-N-methyl amino toluene, 2-carboxy 4-amino toluene (3-amino 6-methylbenzoic acid), 3-carboxy 4-amino toluene (2-amino 5-methylbenzoic acid), 2-hydroxy DMPT, 3-hydroxy DMPT, 2,3-epoxy DMPT, and o and p-N-methyl amino benzoic acid were identified (Figure 10). In this chapter, hydroxy DMPT and epoxy DPMT was discussed in detail for reproducible determination. The hydroxide derivatives of DMPT were synthesized by the author as follows: Each five gram of 2-amino- 5 methyl-phenol or 3-amino- 6 methyl phenol, 10 gram of methyl iodide (methylating reagent) and 5 gram of potassium hydroxide were refluxed for 120 hours with stirring in 100 ml of anhydromethylethyl ketone. After cooling of the reaction mixture, 300 ml of water was added for dissolution and neutralized, thereafter 300 ml of diethylether was added to extract the hydrophobic compounds. Ether layer was separated and evaporated. The residue after evaporation was distilled at vacuum condition. Thus, 3-hydroxy-4-dimethylamino toluene and 2-hydroxy-4dimethylamino toluene, both of which are DMPT hydroxide compounds, were successfully prepared. No methoxy or benzene- methylated compounds were synthesized in this procedure. In place of methyl iodide, the use of dimethylsulphonic acid was one candidate for methylation reagent. As the methylation ability of the latter reagent was so strong that not only amino group, but also aromatic hydroxy group (phenolic OH) was also methylated, therefore methyl iodide was used for selective methylation reaction. Additionally, the 3-carboxy 4-N-methyl amino toluene and 2-carboxy 4-Nmethyl amino toluene were prepared as follows: Each five gram of 3-carboxy 4-amino toluene (2-amino 5-methylbenzoic acid) or 2-carboxy 4-amino toluene (3-amino 6-methylbenzoic acid), five gram of methyl iodide and five gram of potassium hydroxide were refluxed for 120 hours with stirring in 100 ml of anhydrous methylethylketone. Thus, 3-carboxy-4-N-dimethyl amino toluene, 2-carboxy-4-N-dimethyl amino toluene, 2-carboxy 4-N-methylamino toluene and

92 3-carboxy 4-N-methylamino toluene were prepared without production of methoxy compounds. After cooling of reaction mixtures, 300 ml of water was added for dissolution and neutralized, thereafter the mixture was extracted with 300 ml of diethylether. The residue after evaporation was distilled at vacuum condition. Most of 3-carboxy-4-N-dimethyl amino toluene and 2-carboxy-4-Ndimethyl amino toluene were remained in aqueous layer and 2-carboxy 4-Nmethylamino toluene and 3-carboxy 4-N-methylamino toluene were mostly extracted with diethytether. The 3-carboxy-4-N-dimethyl amino toluene and 2-carboxy-4-N-dimethyl amino toluene were eluted later in Capcell Pak | C 18 SG-120 HPLC than 2-carboxy 4-N-methylamino toluene and 3-carboxy 4-N-methylamino toluene. They were chromatographically separated and collected individually using collection scale C-18 column (50x250 mm, linear velocity was identical to that of analytical column). MS fragmentation spectra and HPLC retention time of these synthesized compounds coincide with unknown compounds, therefore the chemical structure of unidentified compounds can be successfully identified. As these compounds have two opposite functional groups of aromatic amine and carboxyl, therefore appropriate SPE procedure is still under research and the reproducible result will be reported in future. The linear gradient elution was carried out using a mixture of 10 mM ammonium acetate/acetonitrile combined with HPLC column of Capcel Pak | C 18 UG 120A (4.6mmx250mm). For 40 min, ratio of 10 mM ammonium acetate/acetonitrile was changed from 9/1 to 1/9. In order to increase sensitivity by MS detection, ammonium acetate was added to the eluent, therefore the addition of ammonium acetate was not for common ion effect, but for increasing MS detection sensitivity. Flow rate was 1 ml/min, detection was by UV at 235 nm and 10 pl of methanol extract of Yunifast | was injected to HPLC of HP 1050 | from Hulett Packard Co. and HPLC was connected with MS of TSQ 7000 | from Finniganmat Co. at atmosphere pressure chemical ionization mode (APCI). The information of mother ion molecular weight and MS fragmentation by HPLC-MSMS mode was obtained and based on these information, chemical structure was successfully identified.

2.2.1.4. Identification of chemical structure of newly found toxic compounds Reproducible separation of hydrophilic compounds was attained from BA to MMA using linear gradient elution. Using this separation method, chemical structure of unidentified compounds eluted from BA to MMA was identified. Using HPLC-MS at APCI mode, only information of mother molecular weight was obtained, therefore HPLC-MS-MS mode was essential to identify chemical structure from chemical fragmentation information. Inferiority of HPLC-MS at APCI mode was that hydrophobic compounds were not sufficiently detected, however unidentified compounds in this case were DMPT derivatives, therefore

93 as they have aromatic successfully detected.

amine

in their chemical structure,

so they were

2.2.1.5. D e t e r m i n a t i o n of u n i d e n t i f i e d c o m p o u n d in saliva Each three sheet of 3•215 cm of Yunifast | was immersed in 10 ml of saliva and the a m o u n t eluted to saliva was determined. P r e t r e a t m e n t of newly identified DMPT derivatives was carried out using SPE C-18 column with an identical m a n n e r to DMPT p r e t r e a t m e n t procedure.

2.2.1.6. SPE p r o c e d u r e of MMA, DMPT and h y d r o x y DMPT in saliva In saliva, BPO was not existed as it is and immediately transformed to BA, therefore BPO in saliva was determined as BA in the following section. There have not been reported on SPE with a satisfactory recovery of MMA, DMPT, BA from BPO and hydroxy DMPT in saliva. Epoxy DMPT was also immediately transformed to hydroxy DMPT in saliva. SPE column used was Bond Elut | C-18 with a void volume and resin weight of 120 ~1 and 1 100 mg, respectively. SPE t r e a t m e n t of MMA, DMPT and hydroxy DMPT was as follows: the C-18 column was conditioned with 2 ml of acetonitrile and 2 ml of 50 mM phosphate buffer at pH 7.5. Thereafter, one ml of saliva was applied to the conditioned column, vacuumed, rinsed with 0.5 ml of 50 mM phosphate buffer at pH 7.5 and eluted with one ml of an alkalized acetonitrile with 50 mM phosphate buffer at pH 8. The drain was trapped and 20 pl were applied to HPLC. Conditioning, rinsing and elution were carried out by a vacuum system identical to the system described in 2.1.2.

2.2.1.7. SPE p r o c e d u r e of BA in saliva As mentioned above, BPO was stable in methanol, but when contacting with saliva or DMPT, BPO was immediately transformed to BA, therefore even though BA was not originally utilized in PMMA fabrication, but BA was existed as a sort of artifact from BPO. SPE procedure of BA in saliva was as follows: depression of ionization of BA was necessary to r e t a i n in reverse-phase columns. Thus, an acetic acid aqueous solution at pH 3 was added to the sample solution at a volume ration of 1/1 and mixed well prior to SPE application and carried out as follows: Bond Elut | C-18 column was conditioned with 2 ml of acetonitrile and 2 ml of acetic acid aqueous solution at pH 3. One ml of saliva was applied to the conditioned column. Saliva was two-fold diluted with an acetic acid aqueous solution at pH 3 prior to the conditioned column application. Thereafter, they were vacuumed, rinsed with 0.5 ml of acetic acid aqueous solution at pH 3 and eluted with one ml of acetonitrile acidified with acetic acid at pH 2.5. The drain was trapped and 20 pl were applied to HPLC. Conditioning, rinsing and elution were carried out by a vacuum system identical to the system described in 2.1.2.

94

2.2.2. R e s u l t o f l i q u i d - l i q u i d e x t r a c t i o n vs. S P E of MMA, DMPT, B P O a n d BA in b l o o d Liquid-liquid extraction was carried out by adding an identical volume of acetonitrile to serum for deproteinization and extraction of MMA and DMPT. The MMA peak showed an insufficient separation from serum a d m i x t u r e s in HPLC and insufficient recoveries of MMA and DMPT (84% and 62% for MMA and DMPT, respectively, n=3) [9,17]. Therefore, an alternative method, SPE, was studied. There have not been reported on SPE with a satisfactory recovery of blood MMA, DMPT or BA. SPE column used was Bond Elut | C-18 with a void volume and resin weight of 120 ~1 and 100 mg, respectively. SPE t r e a t m e n t of blood MMA, DMPT and BPO was already mentioned in analytical section of 2.2.1.6 [17]. The reason of the use of phosphate buffer at pH 7.5 was to depress ionization of DMPT, strong basic compound, but at pH 7.5 ionization of DMPT was incompletely depressed. However, due to characteristics of silica dissolution over pH 8, eluent pH over 8 was not selected. As mentioned in 2.2.1.1, DMPT analysis was successfully a t t a i n e d without using phosphate buffer eluent. This is mostly due to the use of newly innovated column of Capcell P a k | which is completely endocapped. SPE of blood BA from BPO was also already mentioned in analytical section of 2.2.1.7 [17]. Concerning the SPE eluent of BA, acetonitrile, alkalized acetonitrile containing 50 mM sodium hydroxide or acidified acetonitrile adjusted to pH 2.5 with acetic acid were compared for eluting from SPE column. Acetonitrile alone showed an insufficient recovery (80%). Alkalized acetonitrile and acidified acetonitrile indicated 85% or 100% recovery, respectively, therefore acidified acetonitrile was superior to alkalized acetonitrile (Figure 7) [17]. The successful reason using an alkalized eluent for BA elution was mostly due to an identical reason mentioned in SPE for MDA elution. These are favorable dissolution to the SPE eluent and common ion effect. F u r t h e r speculation was t h a t acidified solution was used during conditioning, so alkalinity m a y be suppressed due to acidified circumstances. As being confirmed in Figure 7, no BA detection was in native blood. BPO in blood was also no detection in native blood [9]. In SPE of blood MMA, DMPT and BPO, 50 mM phosphate buffer at pH 7.5 was used for column conditioning. The use of water or more t h a n 50 mM phosphate buffer resulted in a lower recovery [17]. This is because an insufficient depression of DMPT ionization by w a t e r alone and excessive buffer ions at more t h a n 50 mM m a y interfere with DMPT retention on the column. In DMPT elution, alkalized acetonitrile (phosphate buffer at pH 8) was more effective t h a n acetonitrile or acidified acetonitrile due to the identical reason mentioned in SPE for MDA or BA elution. These are favorable dissolution to the eluent and common ion effect (Figure 8). F u r t h e r speculation was t h a t alkalized solution was used during conditioning, so acidity m a y be suppressed due to alkalized circumstances.

95 (a)

(b) Serum blank

BA from BPO (2.2 ppm) in serum

I

,

,

,

,

0

I

5

l

,

min

I

0

,

,

,

,

I

,

5

,

min

Figure 7. SPE of BA from BPO in serum using SPE C18 column: (a) 11 mg of BPO was added to 50 ml of a mixed aqueous solution of water and acetonitrile at a ratio of 20/1.100 gl added to 1900 ~tl of serum to obtain 1 l ppm (~g/ml) serum solution. This was 5-fold diluted with an acetic acid aqueous solution at pH 3. One ml applied to the conditioned c l 8 SPE column with 100% recovery of BA from BPO. SPE and HPLC conditions refer to the text; (b) one ml of 5-fold diluted native serum with an acetic acid aqueous solution at pH 3 was treated and recovered. This applied to HPLC. The result indicates no interference by serum admixtures with the elution of BA.

MDPT (11 ppm) ,

MMA

(a) standard

MDPT

t

(b) serum

(c) serum blank

Figure 8. SPE of MDA and DMPT from SPE C18 column: (a) 1 lmg of MMA and DMPT were added to 50ml of a mixed aqueous solution of water and acetonitrile at a ratio of 20/1. 100~tl added to 1900~tl of water to obtain 1 l ppm (pg/ml) solution. One ml applied to the conditioned SPE C 18 column with 100% recovery. SPE and HPLC condition refer to the text; (b) identical to (a) excepting that 100~tl of solution were added to 19001.d of serum in place of water to obtain 1 l ppm (lag/ml) serum solution. One ml applied to the conditioned C18 SPE column with 100% recovery; (c) one ml of a native serum was treated, recovered and applied to HPLC. The result indicated no interference by serum admixtures with the elution of MMA and DMPT.

95 As being confirmed in Figure 8, no detection of MMA and DMPT was in native blood. Acetonitrile also produced a satisfactory recovery for MMA (neutral compound), but not so much for DMPT (strong basic compound), which will be reasonably understood [17]. The favorable recovery reason for DMPT was due to identical reason for SPE eluention of MDA or BA. As MMA is a neutral compound, so it will not be effected by eluent pH. As being mentioned previously in SPE eluent for basic compound of MDA, it was thought to elute more favorably by treating MDA with an acidic solvent by charging positively the retained MDA for easily removal from the solid resin. However, the experimental result was opposite to the initially speculated result, indicating that the recovery rate with acidified methanol was lower t h a n alkalized methanol. As mentioned in advance, similar phenomena have been observed in SPE elution of DMPT and BA in PMMA dental materials [17]. Favorable results have been produced, both when acidic acetonitrile was used for elution of acidic compound of BA and when alkalized acetonitrile was used for elution of basic compound of DMPT [17]. These results were different from already reported results for the choice of SPE eluent, however it is important that these experimental data have reproducibility, thus it has any sound scientific rationale as being speculated in advance. Sound reason to be considered is common ion effect and favorable dissolution of the compounds of interest to the SPE eluent. This speculation may not be correct, however it is certain that reproducible experimental results were obtained using these SPE elution, which was most important in scientific study. The author considers that any experimental results with reproducibility will be much superior to only speculation without any experimental proof or computer simulation because speculation is speculation and simulation is simulation. Computer simulation cannot predict any interference peak elution with undefined retention time and it will not be useful when determining the compound of interest in complicated matrix such as blood or dirty environmental matters.

2.2.3. R e s u l t of d e n t a l m a t e r i a l a n a l y s i s Figure 9 shows MS MH § (one protonated mother ion) chromatogram (upper) and HPLC gradient chromatogram (lower) of methanol extract detected by UV (235 nm). Figure 10 shows the chemical structure of identified compounds. Concerning SPE procedure, if compounds have carboxy group, they were treated with identical m a n n e r to BA treatment. If the compound of interest has no carboxy group in the chemical structure, identical SPE m a n n e r to MMA, DMPT and BPO will be utilized. When both carboxyl and amine groups exist in the chemical structure as is the case of newly identified compound, further study will be required far appropriate SPE procedure.

97 100 8O 6o 40 20 0 100 -

E+07 1.188

+Q1MS

9

,,

- _

....

elL_

u

A

.

1.0 AU

UV (23 5nm)

8060-

20 2 0

40

i

200

400

600

800

1000

1200

Scan number Figure 9. HPLC-MS MH + chromatogram BA: scan number of 130. Compounds of g): scan numbers of around 200 and 300, compounds of f): scan numbers of 340 and 460, compounds of e): scan numbers of around 520 and around 630 and compounds of a): scan numbers of around 580 and 590. Compound of b): scan number of 620. DMPT: scan number of around 810. MMA: scan number of around 610. 2,3-epoxy DMPT: scan number of around 340. One scan number corresponds to retention time of 2 seconds.

H3C\ /CH 3

H\ /CH 3

N

N

I

~H2

I

COOHl

@@ a)

CH3

b)

HXN/C H3

CH3 ~ H2

c)

d) HXN/C H3

@]-1COOH~~~--COO~t~~--COOH CH3 e) CH3

0

g)

Figure 10. Chemical structure of compounds of a) to g). Compounds a) 2-hydroxy 4dimethylamino toluene; b) N-methyl p-toluidine; c) aniline; d) BA; e) 3-carboxy 4-N-methyl amino toluene and 2-carboxy 4-N-methyl amino toluene; f) 2-carboxy 4-amino toluene (3amino 6-methylbenzoic acid); g) o- and p-N-methyl amino benzoic acids.

98 Neutral compound of MMA and BPO was not affected to pH of eluent, so MMA and BPO elution were used identical procedure to DMPT procedure. The hydroxy DMPT has phenolic OH, but this acidity is weak, therefore this functional group was not effected so much for elution. The SPE procedure of epoxy DMPT will be identical to that of original DMPT.

2.2.3.1. Result of newly identified compounds The newly identified compounds of hydroxy DMPT were recognized to elute from BA to MMA elution. BA elution was confirmed by MH- (mother ion minus one proton). On the contrary, MMA and BPO, which were neutral compound, were hot detected by MS (MH § and MH-) at APCI mode due to less vaporization characteristic. The elution of newly found compounds of hydroxy DMPT was confirmed by MS (MH+), HPLC with UV detection and also confirmed by coincidence of elution time with standard compounds synthesized. Unidentified hydrophilic compounds were determined their chemical structure as hydroxylated derivative of DMPT, 2,3 epoxy DMPT and carboxylated DMPT derivatives from their molecular weight and MS fragmentation pattern as well as coincidence of elution time of standard compounds. The molecular weight of 2 or 3 hydroxylated DMPT and 2,3 epoxy DMPT were 151,151 and 149 daltons and their retention time was 6.7 min, 10 min and 11.5 min, respectively. The retention time of BA, MMA and DMPT was 4.3 min, 20.3 min and 27 min, respectively. The elution of N-methyl-p-toluidine, which was not reported so far, was also confirmed to be eluted just after MMA elution. The retention time of this compound was 20.7 min and that of MMA was 20.3 min. The elution of this compound was confirmed from its fragmentation by MS. These compounds were confirmed from methanol extract of Yunifast | as well as the mixed solution of DMPT and BPO. This possibility indicates that unidentified compounds may be produced from the reaction of DMPT and BPO during PMIViA fabrication. This compound was also treated with identical SPE manner to DMPT t r e a t m e n t of SPE. The 2,3 epoxy DMPT was stable in methanol solution, but when contacted with saliva, the epoxy compound changed immediately to 2 or 3 hydroxy DMPT. It was recognized that BPO was immediately converted to BA when BPO was contacted with saliva or DMPT. As 2,3 epoxy DMPT and BPO were highly reactive, therefore these were speculated to indicate high toxicity. The degree of toxicity was not always parallel to the eluted amount. For example, serum extraction of MMA, DMPT and BA from Yunifast | was 32.04 }~g/g, 66.44 pg/g and 2.3 pg/g, however the cytotoxicity data of IC50 (pg/ml) of MMA, DMPT, BA and BPO using Balb 3T3 cell was 4400,1500, 28.7 and 22, indicating elution amount was not always parallel to cytotoxicity result. It is a problem that epoxide compound is thought to indicate the greatest toxicity, but as this compound will be transformed to hydroxy compound immediately when

99 contacting with saliva or blood, therefore the cytotoxicity test of epoxide compound was not successfully attained. The hydroxy DMPT compounds found in saliva were the total of originally existed in saliva plus those from 2,3 epoxy DMPT, but the differentiation of origin of these compounds in saliva was extremely difficult and the effort for differentiation was meaningless. The eluted amount of 2 hydroxy DMPT and 3 hydroxy DMPT into saliva in successive three days was 10.7 pg/g and 15.8 ~g/g (n=3), respectively. The amount of epoxy DMPT in saliva was not attained due to transformation to hydroxy DMPT immediately when contacting with saliva. The elution of successive three days will be minimum due to putrefaction of saliva for further period immersion, thus real elution amount will be much greater due to much longer period contact with dental material with saliva or blood through teeth. The BPO in saliva was determined as BA at 3.5 ~g/g.

2.3. B l o o d u r e a - p r e t r e a t m e n t a n d a n a l y t i c a l c o n d i t i o n s 2.3.1. U l t r a f i l t r a t i o n Author's blood was sampled for blood urea analysis. Native blood and denatured blood with acid were centrifuged for ultrafiltration at 13,000 rpm (10,000 g) for 40 to 60 minutes and the s u p e r n a t a n t was applied to p r e t r e a t m e n t method. Ultrafiltrator used was Kokusan Co. H-1300 | in Tokyo. Membrane of ultrafiltration is Centrifree | from Amicon Co. made of cellulose with cut-off molecular weight of 10,000 daltons. Native blood and denatured blood are for analysis of free and total urea, respectively. 2.3.2. A u t o m a t e d S P E For p r e t r e a t m e n t method, an automated SPE and a dialysis were compared. SPE of blood was as follows: S u p e r n a t a n t after ultrafiltration was applied to the conventional strong cation exchange column (H type) of Bond Elut | SCX (500 mg of resin weight and 0.6 ml of void volume). The SCX column was conditioned with 3 ml of methanol followed by 3 ml of water at a flow rate of 3 ml/min. One ml of blood was applied to the conditioned SCX column at a flow rate of 0.3 ml/min and rinsed with 1 m! of water at a flow rate of 3 ml/min. The retained urea to the SCX column was eluted with 4 ml of 5% phosphoric acid at the flow rate of 1 ml/min. These procedures were carried out using the automated SPE equipment of BenchMate | from Zymark Co. (Hopkinton, MA) controlled SPE procedures with computer [29]. Automated SPE was much superior to manual type SPE in terms of pressure control, which will significantly affect to variation of recovery rate. 2.3.3. A u t o m a t e d d i a l y s i s Automated dialysis was carried out as follows: The ASTED | and the trace enrichment column (TEC | from Gilson Co. (Villers-le-Bel, France) were used. A polymer-based strong cation-exchange resin column (Na type) was used for TEC |

100 which served for condensing dialysate. The resin weight was 20 mg. The column was conditioned with 1.5 ml of 1 M sulfuric acid followed by 0.9 ml of water at a flow rate of 2 ml/min. The other conditions were as follows: dilutor 1, 0.01% TritonX 100; dilutor 2, 5 mM phosphate buffer (pH 7.4). The cut-off molecular weight for dialysis membrane made of cellulose was 15,000 daltons [29]. 2.3.4. H P L C c o n d i t i o n Urea analysis by conventional strong cation exchange resin column (H type) was carried out as follows: After pretreatment, sample solution was applied to the conventional strong cation exchange resin column of MCI GEL CK 08S | from Toso Co., 4.6• mm, 11-14 pm particle diameter. Other conditions were as follows: eluent, 1 mM HCI solution; flow rate, 1 ml/min; detection, 200 nm, application volume, 20 pl; column temperature, 35~ HPLC equipment and UV detector used were PU-980 | and PU-970 | respectively, from Nihonbunko Co. in Tokyo. The comparison of separation efficiency between ion chromatography use column with smaller capacity and conventional strong cation exchange resin column with greater capacity was studied and the result will be described later

[30]. 2.3.5. MECC c o n d i t i o n of b l o o d u r e a a n a l y s i s MECC analysis was carried out as follows: running buffer constitute, 75 mM sodium dodecyl sulfate (SDS),10 mM hydrogenphosphate, 6 mM tetraborate, pH 9.2; voltage: 25 kV, current: 70 ~1; effective capillary length: 68 cm, inner diameter 75 ~m. MECC equipment is CAPI-3100 | with photodiodearray detector from Otsuka Electronics Co. (Osaka, Japan). 2.3.6. S u p e r c r i t i c a l fluid e x t r a c t i o n of b l o o d u r e a a n a l y s i s As other pretreatment method, supercritical fluid extraction (SFE) will be available [32]. However, this method has a restriction mostly to vaporizable hydrophobic compound extraction. As urea is hydrophilic, so current SFE technique is not effective for urea extraction. If this restriction will be conquered, this pretreatment method for isolation as well as purification will be more appropriate as extract is liquid gas, therefore it is unnecessary for evaporation and condensation. 2.3.7. R e s u l t of a u t o m a t e d S P E of b l o o d u r e a Blood is a complicated matrix, indicating that blood contains many compounds to interfere urea analysis. In order to attain satisfactory separation with sufficient resolution free from blood admixtures, blood urea must be pretreated. The recovery of blood urea from SPE cation exchange column was almost 100%. No or quite lass recovery of urea was attained when using reversed-phase column such as C-18 column [20-22]. This means urea did not retain sufficiently on the reverse phase columns because urea was highly hydrophilic compound. In case of

101

C-18 column, urea was eluted at around void volume without separation from blood hydrophilic admixtures. That's why the author used complicated column switching method combined with urease immobilized column for post column Indophenol colorimetry in his previous study [22]. This method itself at that time was valuable with new information, however it has originally inferiority that the method was not easily applicable to routine analysis due to its complexity, which was a weak point of the author's previous method [22]. At t h a t time when this paper [22] or precolumn method paper [21] were published in 1985-1986, column switching method was carried out manually by calculating the switching time. Recently an automated column switching equipment is available in the market, therefore by applying this innovated automated column switching equipment to the author's previous set-up method, previous manual method will be more easily applicable for routine analysis of urenic toxin determination. SPE procedure for blood urea was carried out by acidified condition as SCX column was used. In order to elute urea successfully, stronger acidity eluent will be appropriate for both benefit of successful elution and transforming to N type of SCX column. In SPE t r e a t m e n t of blood urea, p r e t r e a t m e n t mechanism and column differed from MDA, BA and DMPT used. P r e t r e a t m e n t mechanism and column used for MDA, BA and DMPT procedure were partition mechanism and reverse phase columns. Those for urea are ion exchange mechanism and cation exchange column. This is because acidic eluent was successful for weekly basic compound of urea. As being mentioned in advance, urea can't be successfully treated with reverse phase columns because urea did not retain on them successfully [20-22]. Differential analysis of free from bound urea can be attained by ultrafiltration. Far differential analysis of bound from free urea, ultrafiltration using centrifugation is one method. The alternative method is dialysis. 2.3.8. R e s u l t o f a u t o m a t e d

d i a l y s i s m e t h o d for b l o o d u r e a

Ultrafiltration can be replaceable to dialysis. Using the automated dialysis of ASTED | dialyzate was condensed on the condensed column (TEC | enrichment column). The compounds accumulated on the condensed column (TEC | column, strong cation exchange column with Na type) were eluted with 1 mM H C1 of HPLC eluent, which is identical to SPE t r e a t m e n t based on same reason [29]. HPLC chromatograms after strong cation exchange automated SPE and the automated dialysis were presented in Figures 11 and 12, respectively [29]. Urea was determined using standard addition method (Figure 13) [29]. The cross point to horizontal line (x line) is the endogenous urea amount determined. When compared SPE and dialysis chromatograms (Figures 11,12), urea peak was separated with baseline separation from blood admixtures in both cases. If mentioning in detail, in the chromatogram after SPE t r e a t m e n t (Figure 11) only urea was eluted with sufficient separation from other blood admixtures eluted in void volume, indicating no compounds interfered blood urea analysis. However, in

102

~.

o~.

Ca) o

"~176 k_.___A O ;> -1.20

r

(b) e,i

-1.20

a

O~ - 1.60 ~

-1.60

lake_-

9

-2.00

_

__

.~

.....

-

0.00

-2.00

,

,

i

i

2.00

4.00

6.00

8.00

0.00

Time (min)

,

,

,

J

2.00

4.00

6.00

8.00

Time (min)

Figure 11. HPLC chromatogram after automated SPE treatment (a) blank blood, (b) spiked urea to blood at the concentration of 0.1 mg/ml. 100.00

-

80.00

-

60.00 -

40.00 -

20.00 -

O. O0

- ~

0.00

r

~

1.00

l

~

r

2.00

~

r

3.00

~

l

4.00

~

'

5.00

r

~

6.00

Time (min) Figure 12. HPLC chromatogram of blood after pretreatment with automated dialysis.

case of dialysis (Figure 12), urea can be separated anyhow with baseline separation from blood admixtures, therefore if blood sample, column capacity, column lot, urea a m o u n t or any other factors will differ, there is a possibility t h a t baseline separation in urea analysis m a y not be attained. Inferior point of dialysis was a lower recovery rate of blood urea at around 10% mostly depending on TEC | capacity r a t h e r t h a n dialysis system, so by improving TEC | capacity, this problem will be resolved. However, it will r e m a i n the problem as follows: if TEC | capacity m a y differ, separation efficiency has a possibility to be diminished and peak broadening m a y be more significant,

103 250000

200000 o 150000 o 100000

50000

0

I

0

I

400

I

I

800

I

I

1200

I

I

1600

I

2000

Concentration (~tg/ml) Figure 13. Calibration curve of spiked urea after dialysis treatment with ASTED | thus separation efficiency of chromatogram after dialysis will be required to be improved because dialysis chromatogram of Figure 12 indicated a sufficient base line separation from admixtures, but an insufficient separation of urea from blood admixtures may occur if TEC | capacity may change [29]. Reverse elution is an alternative approach to resolve successfully without diminishing separation efficiency. Automated dialysis and automated SPE equipments, ASTED | and ASTEC | from Gilson Co., respectively, can be connectable to HPLC for on-line analysis. Additionally if an autosampler will be installed to them, the automated pretreatment system can be attained successfully. The hyphenated and automated set-up system with autosampler-automated SPE or automated dialysis and HPLC in combination will currently be available in the market. For the author it is additionally desirable to obtain the on-line autocentrifugation (ultrafiltration) system for differentiation of free from bound type compounds in blood. However, this kind of equipment is not available in the current market. This hyphenated technique will be desirable for routine analysis in clinical analysis.

2.3.9. Result of u r e a a n a l y s i s w i t h c o n v e n t i o n a l s t r o n g c a t i o n e x c h a n g e resin c o l u m n As reported in the cited literatures from 19 to 22 as well as already mentioned in the text in advance, simultaneous uremic toxin analysis of urea, uric acid, creatinine and methylguanidine was carried out with reverse phase column combined with complicated column switching method and online immobilized urease column for post column method with Indophenol colorimetry for urea analysis [19-22]. Among these uremic toxins, only urea did not significantly

104 retain on the reversed-phase HPLC columns [22]. This indicates reverse phase columns are ineffective for pretreatment of blood urea. P r e t r e a t m e n t method of blood urea with sufficient recovery has not been reported so far, so the author considered the appropriate p r e t r e a t m e n t method of blood urea. In the preliminary experiment, the author carried out the experiment using an ion chromatography column with a smaller ion exchange capacity, however the separation of blood urea from admixtures was not satisfactory [30]. For the alternative method, the author considered the use of the conventional strong cation exchange column (H type) with a greater ion exchange capacity far the differential analysis of blood urea from endogenous ammonium. Endogenous blood urea eluted faster than endogenous ammonium in strong cation exchange chromatogram. Urea was detected at 200-210 nm [30]. By using this procedure, a simpler procedure for differential analysis of urea and endogenous ammonium can be attained. Urea can be successfully separated from blood admixtures using the conventional strong cation exchange column (H type).

2.3.10. R e m a i n e d p r o b l e m s a s s o c i a t e d w i t h the use of s t r o n g c a t i o n e x c h a n g e resin c o l u m n As mentioned in 2.3.9, separation of endogenous blood urea from endogenous ammonium and admixtures can be successfully attained using a conventional strong cation exchange resin column. The problem unresolved yet was that the retention time of urea was unstable even though completely conditioned with strong acid to H type for suitably prolonged period [30]. Retention time of urea gradually increased (however, not always constantly), speculating gradually changing to H type in the column. This will be due to existing SOaH functional groups in the interior of pore of silica support. If any deterioration to the column with blood may occur, shorter elution wilt be observed due to a smaller ion exchange capacity. As being mentioned in advance, speculated reason of gradual increase of retention time may be due to a gradual change to H type of the interior functional group in silica pore. This problem was observed in both cases in MCI | gel from Mitsubishi Co. and TSK | gel from Toso Co. This phenomena were not be well reasonably clarified and explained yet. Therefore, if the reader will determine from the peak height of urea, this phenomena may cause a trouble to attain reliable data. When determining from peak area using computer calculation, this problem will be somewhat diminished, but still remains the problem how to set base-line of urea peak. Depending on base line setting, peak area will be significantly differed, especially if the peak indicates significant tailing, which is so often observed in ion exchange column chromatogram for basic compound analysis. Up to here urea analysis using HPLC has been described. In the next section, urea analysis using capillary electrophoresis will be described.

105 2.3.11. S e p a r a t i o n efficiency of MECC analysis for blood urea and uric acid In MECC using SDS as a micellar compound over critical concentration, blood urea migrated at around void volume overlapped with blood admixtures by direct blood injection Figure 14, indicating undesirable separation, there other mode such as CZE will be required. As being mentioned in advance, single urea analysis in any matrix, urea analysis will be attained by CZE, but simultaneous analysis of urea and other hydrophobic compounds such as uric acid will be required MECC mode. MECC mode is identical to reverse phase HPLC mode, therefore failure of urea analysis by MECC is identical to HPLC with C-18 column. Furthermore, blood uric acid analysis by MECC indicated insufficient separation of uric acid peak from blood admixtures mostly blood proteins (Figure 14), indicating inferior to the reversed-phase HPLC (C-18 column HPLC)

0.1

0.08 -

t U

0.06

0.04

0.02

Uric acid aqueous solution -0.02

-0.04

I 2.2

I 2.4

I 2.6

I 2.8

I 3

I 3.2

I 3.4

I 3.6

I 3.8

Time (min) Figure 14. MECC electropherogram of human blood serum U: urea, UA" uric acid. Upper: human serum; lower: standard uric acid aqueous solution.

105 [19-22, 24]. This result was identical to the published paper by Schmutz and Thormann [23]. HPLC can separate blood compounds successfully by changing sort of columns, so in terms of blood urea analysis, HPLC was thought to be superior to MECC for selective blood urea and uric acid analysis. Concerning blood urea analysis, HPLC is thought to be superior to capillary electrophoresis (CE) in terms of appropriate selection from several sort of separation mechanisms. However, CE technology is advancing day by day, therefore this status is not always unchanged. In case of only blood urea analysis as being mentioned in advance, not for the simultaneous analysis of blood urea and uric acid, CZE mode will be thought to be more appropriate [24]. CZE will be superior to conventional isotachophoresis in terms of ion analysis. Inferior points of CE is that capillary column utilized is only bare or coated silica column with or without addition of critical concentration of micelle compound for MECC for the former, indicating less selection of separation mechanisms compared with separation mechanisms applicable to HPLC [24]. Separation time by CE was much shortened due to a greater theoretical plate number. Sample volume of CE was much less than the conventional HPLC excepting capillary HPLC [24]. This will be desirable for clinical analysis because tiny sample volume is much favorable to patients. The most inferior point of CE to be improved is less reproducibility of injection volume, migration time, peak height and other factors affecting to the accurate determination. Less reproducible data will not be well evaluated. When this inferiority will be conquered by improvement with innovated and advanced technology, CE will become a desirable analytical equipment in clinical analysis.

Q

SPE vs. LIQUID-LIQUID EXTRACTION AND THE FORMATION OF ARTIFACT WITH SOLVENT EXTRACTION

Liquid-liquid extraction was a conventional pretreatment method for isolation and purification of the compound of interest in complicated matrix. The inferior points of this method were copious consumption of organic solvent for extraction, requiring further condensation, which may result in loss of recovery during evaporation condensation, time consuming due to repeated extraction or artifact formation from compound of interest during contact with extraction solvent [31-33]. An example of artifact formation during liquid-liquid extraction was as follows: in case of amine compound extraction with methanol or ethylacetate, formaldehyde from methanol causes Mannich reaction with methylol linkage to produce amine oligomers artifact, which is more carcinogen [15-17]. They are artifact and produce lower recovery rate of compound of interest in complicated matrix. Formation of artifacts was prevented by replacing extraction solvent of methanol with ethanol. Liquid-liquid extraction process required vacuum evaporation/condensation and prolonged contact with the extraction solvent, thus it had more possibilities to produce artifact compounds.

107 During a vacuum evaporation/condensation process, compounds of interest were so often vaporized without being successfully trapped and may cause a reduction of recovery rate and thermal decomposition [31-33]. Liquid-liquid extraction required a greater amount of consumption of organic solvents, which was hazardous to chemists. The recovery rate of single t r e a t m e n t of liquid-liquid extraction was less t h a n recovery rate of SPE. Because single t r e a t m e n t of liquidliquid extraction was almost identical to SPE column with one theoretical plate. In general SPE procedure did not require condensation and could be condensed using less amount of eluent t h a n applied sample volume. Concerning artifact formation, when ethyl acetate was used as an extraction solvent, compounds with hydroxyl or an amino group was acetylated, which causes a reduction of recovery rate. Artifact of acetylated compound was generally more mutagenic and toxic, thus this may lead to misunderstanding to the researcher that he/she may consider to extract strongly toxic compounds and of course this will cause lower recovery rate. Analytical chemists should keep in mind well about artifact formation during solvent extraction by reacting solvent with the compound of interest, otherwise he/she may have misunderstanding to extract high toxic compounds in their experiment. In that meaning almost 100% recovery rate with reproducibility is most desirable to avoid misunderstanding, otherwise less recovery may indicate a possibility of artifact formation during extraction. In order to avoid artifact formation, one approach is to replace to appropriate extraction solvent. The other is to use SPE in place of liquid-liquid extraction. SPE will be superior to conventional liquid-liquid extraction in terms of less organic solvent consumption, mostly unnecessary for condensation, less possibility of artifact formation due to shorter contact period with extraction solvent and other benefit already mentioned. Inferiority of SPE is that if SPE procedure is carried out manually, vacuum pressure control was very difficult, which results in less reproducibility of recovery rate. In that meaning, automated SPE such as BenchMate | ASPEC | or RapidTrace | will be recommendable because these SPE equipment procedure are controlled by computer, thus pressure control reproducibility is much superior to manually controlled SPE. Among them, RapidTrace | from Varian Co. will be most recommendable because it is cheapest and handy type with 10 samples t r e a t m e n t automatically in successive run. 4.

CONCLUSIONS

Artifact formation during solvent extraction including SPE is problematic, but this phenomena was so often overlooked. In order to avoid artifact formation to improve recovery rate, appropriate selection of the extraction solvent, which will be inert to the compound of interest, will be essential. In that meaning, conventional liquid-liquid extraction has several inferiority compared with SPE,

108 therefore SPE pretreatment should be seriously considered because SPE was superior to liquid-liquid extraction in terms of lass possibility of artifact formation, less consumption of solvents, less experimental time, greater recovery rate, fewer necessity of condensation, etc. Eluent of reverse phase columns for basic or acidic compounds was different from conventionally reported results for SPE elution. Suitable experimental result of eluent constitute was opposite from the initially speculated eluent. It is more important to recognize that the experimental result has reproducibility. Reproducible result is essential as this indicates any truth in science. Explanation of the experimental result was most probably due to common ion effect, favorable dissolution to the eluent and other speculated reasons mentioned in the text. Pretreatment with several kind of procedures is identical to the treatment done in the hospital. These are artificial dialysis, artificial filtration, ultrafiltration or artificial adsorption. Artificial dialysis supports kidney function and artificial adsorption supports tiny part of liver function, toxin removal. For example, the artificial toxin removal, i.e. birillubin removal with charcoal adsorption or with strong anion exchange resin in patient's blood for treatment, is for applicable to pretreatment of analysis as well as maintaining h u m a n health. When considering these, readers will recognize that adsorption, recovery, isolation or separation were not always for analytical chemistry, but also for clinical treatment for patient's treatment. As newly identified toxic artifacts were found to be produced during PMMA fabrication. These were successfully identified using HPLC-MS-MS at APCI mode. These were derivatives of starting compounds, mostly DMPT and BPO. During reaction of DMPT and BPO, most of newly identified compounds were produced. Epoxy derivative compound of DMPT was further converted to other compounds when contacting with saliva. Some artifact derivative compounds are toxic and they have both aromatic amine and carboxyl functional groups in their chemical structure, therefore appropriate SPE procedure must be further studied. Eluent of reverse phase columns for basic or acidic compounds was different from reported results by other researchers for SPE elution. The author's experimental result indicated opposite from the initially speculated procedure, which was identical to the already reported procedure by other researchers. More important is to recognize that the experimental result has good reproducibility or not. The reproducible experimental results indicate any truth of science which the researcher may overlook in consideration. Explanation of the author's experimental result was mostly due to common ion effect and favorable dissolution to the eluent. This explanation may or may not be true, however experimental result with reproducibility is true and nobody can deny the experimental result. This is the importance of the experimental science. Due to computer advancement, so often experiment has a tendency to be neglected, but experimental result has an important meaning in that sense.

109

REFERENCES

I. H. Shintani, J. Biomater. Appl.,10 (1995) 23. 2. H. Shintani, J. Radiation Steril., 1 (1992) 1 I. 3. H. Shintani in L.E. Elfer (ed.), Ohio Science Workbook: POLYMERS, The Ohio Academy of Science, Ohio,1993, 84. 4. H. Shintani, Radiat. Phys. Chem., 47 (1996) 139. 5. H. Shintani, Polym. Degradation Stabil., 32 (1991) 17. 6. H. Shintani and N. Hirata, Radiat. Phys. Chem., 46 (1995) 377. 7. Shintani, Biomed. Instrument. Technol., 29 (1995) 513. 8. H. Shintani, J. Anal. Toxicol., 15 (1991) 198. 9. H. Shintani, J. Liq. Chromatogr., 15 (1992) 1315. I0. H. Shintani, J. Chromatogr., 600 (1992) 93. II. H. Shintani and A. Nakamura, Fresenius Z. Anal. Chem., 333 (1989) 637. 12. H. Shintani and A. Nakamura, J. Anal. Toxicol., 13 (1989) 354. 13. H. Shintani and A. Nakamura, J. Biomed. Mater. Res., 25 (1991) 1275. 14. H. Shintani, Japan J. Medical Instrumentation, 65 (1995) 249. 15. H. Shintani, Japan J. Medical Instrumentation, 64 (1994) 345. 16. H. Shintani~ J. Liquid Chromatogr. Clin. Anal., 18 (1995) 613. 17. H. Shintani, T. Tsuchiya and A. Nakamura, J. Anal. Toxicol., 17 (1993) 73. 18. H. Shintani, Japan J. Medical Instrumentation, 65 (1995) 486. 19. H. Shintani, A.B. Wojcik, R. Tawa and S. Uchiyama, in: S. Lain and G. Malikin (eds.), Analytical Applications of Immobilized Enzyme Reactors, Blackie Academic & Professional, Glasgow, UK, 1994, 13 I. 20. Shintani and H. Suzuki, in: D.L Wise (ed.), Bioinstrumentation and Biosensors, Marcel Dekker, New York, 199 I, 18 I. 21. H. Shintani and S. Ube, J. Chromatogr., 344 (1985) 145. 22. H. Shintani, J. Chromatogr., 378 (1986) 95. 23. A. Sehmutz and W. Thormann, Electrophoresis, 15 (1994) 51. 24. H. Shintani, in: H. Shintani and J. Polonsky (eds.), Handbook of Capillary Electrophoresis Application, Blackie Academic & Professional, London, UK, 1996, 499. 25. S. Fujiwara, H. Todoroki, H. Ohhashi, J. Toda and M. Terasaki, J. Food. Sci., 55 (1990) 1018. 26. M. Tortoreto, P. Catalani, M. Bianchi, C. Blonda, C. Pantarotto and S. Paglialunga, J. Chromatogr., 262 (1983) 367. 27. J. Cocker, L.C. Brown, H.K. Wilson and K. Rollins, J. Anal. Toxicol., 12

(i 988) 9. 28. 29. 30. 31. 32. 33.

M.C. Bowman, J. Assoc. Off. Anal. Chem., 61 (1978) 1253. H. Shintani, J. Chromatogr. Sci., 34 (1996) 92. H. Shintani, J. Liq. Chromatogr., 17 (1994) 1737. H. Shintani, J. Liquid Chromatogr. Clin. Anal., 18 (1995) 2167. H. Shintani and G. Inoue, Bunseki Kagaku, 43 (1994) 805. H. Shintani, Japan J. Medical Instrumentation, 66 (1996) 414.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

Adsorption in polarographic/voltammetric

environmental

111

analysis

R. Kalvoda UNESCO Laboratory of Environmental Electrochemistry, J. Heyrovsky Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, Dolejskova 3, 182 23 Praha, Czech Republic

The aim of this paper is to show that adsorption of surface active substances from the solution to the electrode surface doesn't mean always only complications in electroanalysis but in contrary, adsorption can yield advantages in realization of height sensitive electroanalytical methods for determination of inorganic and organic compounds of environmental and biological significance. Emphasis is given on application of polarography and voltammetry as well as procedures derived from them in environmental analysis.

1.

PREFACE

- ADSORPTION

PHENOMENA

IN ELECTROANALYSIS

Many chemical compounds mostly of organic origin exhibit surface activity, reflected in electrochemistry by adsorption at the solution-electrode interface. The particles of the dissolved substance may be bound to the surface of the electrode by physical, chemical or electrical forces. Capillary forces that cause adsorption in solution are regarded as physical forces, if the dissolved substance exhibits specific affinity for the surface, these forces may acquire the character of a chemical bond. Electrical forces are involved when the surface is charged. At a given temperature the quantity of adsorbed substance depends on concentration and the concentration dependence is given by an adsorption isotherm [1]. The rate of the formation of the adsorbed layer is affected both by the rate of the actual adsorption of the compound from the solution layer in close contact with the electrode and also by the rate of transport - mostly governed by diffusion - of this compound from the bulk of the solution to the electrode surface. The slower of these processes is the rate controlling step in the formation of the adsorbate. These adsorption process affects thus the properties of the electrode double layer in a measurable manner, which can form the basis for electrochemical analysis of adsorbable surface active substances (SAS) present in the solution. A number of electrochemical techniques can be used for such measurements - many of them

112 are derived from the Heyrovky's polarography. But it should be also mentioned that some components of the solution adsorbed at the electrode surface influence frequently the electrode process which takes places at polarogarphic/ voltammetric measurements in such a way that they shift, deform, split or even eliminate the polarographic wave or voltammetric peak of the analyzed compound. Similarly adsorption can influence improperly or even can act detrimental to measurements - mainly long term ones - with amperometric or potentiometric sensors (e.g. ISEs). The aim of this article is to show that adsorption of surface active compounds from solution to the electrode surface doesn't mean in electroanalysis always a complicating factor but in contrary that adsorption can serve as a mean for very sensitive electrochemical analysis of inorganic and organic compounds. From various methods based on electrochemical adsorption should be mentioned from the historical point of view on the first place the suppression of the polarographic maxima observed by Heyrovsky at his very first applications of polarography in studies of SAS in water - termed polarographic adsorption analysis. Electrocapillary measurements which used Heyrovsky at his fundamental experiments leading to the invention of polarography can be mentioned on this place. Similarly it is with tensammetry developed from polarography or measurements of differential capacity of the electrode double layer. But from the practical point of view - with respect to recent voltammetric methods - it is necessary to emphasize procedures based on adsorptive accumulation of the analyte on the electrode surface: after an preset accumulation period the whole adsorbate is analyzed by voltammetry ("Adsorptive stripping voltammetry - AdSV") or potentiometry ("Adsorptive stripping potentiometry - PSA (according to IUPAC terminology this method should be called "chronopotentiometry"). At present these stripping methods represent the most frequently used electrochemical methods in environmental microanalysis as well as in analysis of physiologically active compounds mainly in pharmacy, agrochemistry and biochemistry. Thus in the following paragraphs first the stripping methods shall be discussed in more details followed by other electroanalytical methods frequently used in adsorption environmental analysis.

0

ADSORPTIVE ACCUMULATION OF SUBSTANCES AT ELECTRODES (VOLTAMMETRIC ADSORPTIVE STRIPPING METHODS)

2.1. I n t r o d u c t i o n Most applications of chemical analysis to environmental protection involve trace determinations, often at a part-per-billion level or lower. Among methods that can satisfy this demand belongs without doubt polarography/voltammetry. But to attain its today's microanalytical quality the original Heyrovsky method had to pass through different improvements and modifications, mainly in the direction of a substantial increase of its sensitivity.

113 In addition to electronic methods which attempted to eliminate the unwonted charging current (but for the charging the electrode to the desired voltage necessary) and thus improve the signal to noise ratio, methods were developed to ensure the lowest detection limit: These are based on accumulation of the analyte on the electrode surface, followed by voltammetric determination of the previously accumulated compound. In this sense the most popular m e t h o d perhaps now regarded as "classic" - is anodic stripping voltammetry (ASV) used mainly in trace analysis of heavy toxic metals. For trace analysis of organic compounds and metal chelates in the ppb and sub ppb concentration region serves another stripping method - adsorptive stripping voltammetry (AdSV) which in contrast to the previous ASV method which makes use of electrolytic accumulation of the metal at the electrode surface, is based this time on adsorptive accumulation of the species on the electrode. As mentioned earlier the amount of substance adsorbed at the electrode is dependent among other parameters on its concentration in solution. Important role plays here not only the rate of the adsorption but also the rate of the diffusion transport of the substance from the solution to the electrode. For AdSV there is interesting only the diffusion controlled adsorption, for which is valid the Koryta [2] equation (1) Fm = 7 . 3 6 . 1 0 - 4 cDlff2t ~

(1)

where C is the concentration of the surface active substance with diffusion coefficient D, t is the time required for complete electrode coverage and Fm is the maximum value of the surface excess in mol'cm -2 of the adsorbed substance for complete coverage of the electrode. The voltammetric peak current is thus roughly proportional to the product of C and t m, when neither of these values is too large. The parameter t is now considered as the duration of the accumulation the accumulation time tacc. If the amount of substance adsorbed on the electrode surface is controlled by the rate of adsorption which is smaller than the diffusion rate, it can be assumed that the concentration of the SAS at the electrode surface equals to its concentration in the solution. Similar conditions are valid for weak adsorption. Neither of these cases can be exploited in AdSV. The problem of adsorptive accumulation was tackled in the Heyrovsky Institute just in the early fiftieths during our studies connected with alternating current oscillographic polarography [1] when we observed that such an accumulation of a depolarizer (like elemental sulphur [ 3 ] , mercurous thiosulphate, poorly soluble inorganic substances, etc.[4]) on a mercury electrode leads to a thousandfold increase in sensitivity of this oscillographic method. Practical applications of voltammetric analysis with adsorptive accumulation of substances at the electrode started only at the ends of the eighties as improved types of hanging mercury drop electrodes become commercially available. (More historical details on AdSV are given in [5]).

114

2.2.

Experimental a r r a n g e m e n t and working conditions

Concerning instrumentation, it should be emphasized that AdSV can be performed, like other voltammetric stripping methods, using a conventional polarograph equipped with a suitable electrode with a constant surface. Most commercially available instruments can be employed both for classical polarographic method - sc. DC polarography and also for pulse methods, especially differential pulse voltammetry and square wave voltammetry. In the choice of a suitable voltammetric method for recording of the curves, it should be mentioned that the measurement of the peak height can be often complicated by unfavorable supporting electrolyte background curves especially at more positive potentials and that in such cases the DC method is preferable. Apparatus with automatic timing of the individual operations is useful for controlling the individual steps in AdSV measurements like accumulation time, solution stirring, duration of rest period, initiation of polarization. A computerized instrument is useful for this purpose. The stripping process can be also controlled by potentiometry: in this case the potential changes caused by reoxidation (or reduction) of the adsorbate by the oxidant present (like 02 or Hg 2§ in the solution or by imposed constant current is measured. This potentiometric stripping analysis (correctly chronopotentiometric stripping) is mostly used for trace metal analysis. This method, in principle similar to the just mentioned Heyrovsky-Forejt oscillopolarography, is out of scope of this article - for information see e.g. [6] and [7] or [8]. Most types of electrodes used in voltammetry can be employed in AdSV provided that a constant, completely reproducible surface can be ensured throughout the whole measurement cycle or better, during a series of measurements. The above requirements on reproducibility can be best satisfied by a hanging mercury drop electrode. The carbon paste electrodes or platinum electrodes are mostly used for measurements, where the adsorbed compound is during the voltammetric scan oxidized, because they can be polarized to more positive potentials than mercury. AdSV can be carried out also with chemically modified electrodes which consist of an electrode material to the surface of which are bound by chemisorption or covalently substances or functional groups that alter their properties mainly in improving the sensitivity and selectivity. The modifier can also be used as a filter preventing passage of interferents to the electrode surface (more on application of chemically modified electrodes in AdSV see in [9]). For such purposes mostly carbon paste electrode containing different resins, ion exchanger, complexing agents, etc. are used. From many examples can be mentioned paper [10] which deals with the determination of Ni 2§ traces at a dimethylglyoxime modified carbon paste electrode. Several publications, e.g. [11] describe analysis of organic compounds using clays, resins, silicone, etc. as modifier. However the use of modified electrodes in routine analysis can be often complicated due to difficulties mostly in the reproduction of the electrode surface and in addition, the accumulation process at these electrodes is far more complex than the reversible adsorption at a mercury electrode. Also the analyte may not

115 be necessarily stripped off from the electrode during the voltammetric record and thus a regeneration step for the electrode is often required. Nevertheless in some cases they can contribute to improvement in selectivity (e.g. by using phospholipid covered electrodes [12]) or perform preconcentration of large hydrophobic cations at polyester-sulfonic acid film coated electrodes: in fact this is a type of ion exchange v o l t a m m e t r y [13]. It should be noted when using paste electrodes t h a t the substance can be also accumulated as a result of solution penetration into the binder during the accumulation period: this may be a combined adsorption-extraction effect or a purely extraction procedure [14]. It is relatively simple to decide w h e t h e r a substance can be determined by using AdSV at a mercury electrode (an analogous procedure is used with other types of electrodes as mentioned below). First the voltammetric behavior of the compound (at a concentration of about 10 .6 mol-1-1) is examined at a hanging mercury drop electrode in different supporting electrolytes using the differential pulse or square wave method for recording the curve. In the found optimum supporting electrolyte, the initial potential is then set to 0 V or -0.1 V vs. SCE, a new mercury drop is formed and the voltage scan towards negative potentials at an applied scan rate of 20 mV" s 1 is immediately started. After the voltammetric curve has been recorded, a new mercury drop is again formed and the same initial potential applied but this time for a period of 60 s in stirred solution. After this accumulation period (tacc), stirring is stopped and the voltage scan r u n as previously after a quiescent period of 10 s. If the surface activity of the examined compound leads to its accumulation, a substantial increase in the peak current is obtained as not only the substance transported to the electrode by diffusion but also the whole a m o u n t of compound adsorbed on the electrode surface is reduced during the voltage scan (Fig. 1.). For oxidizable organic compounds a solid type of working electrode is used in a similar way: the accumulation is studied at 0 V or with an "open circuit" and then the voltammetric curve is recorded toward more positive potentials. With solid electrodes the accumulation step often occurs simply on i m m e r s i n g the electrode in a stirred solution containing the examined compound for a certain tacc. The electrode is then rinsed, cleaned and transferred to a "pure" supporting electrolyte (and connected to the instrument), where the actual voltammetric determination is carried out. This procedure has a certain a d v a n t a g e because the effect of accompanying substances in the sample (e.g. not adsorbable compounds yielding a peak at the same potential as the compound to be analyzed) on the recorded voltammetric curve can be eliminated. However interfering substances can still be adsorbed from the sample during the accumulation period and can sometimes greatly influence the AdSV determination. Interferences can be sometimes avoided by proper choice of accumulation potential and of the supporting electrolyte or changing only its pH value. The same holds for simultaneous determination of two compounds present in the sample.

116

L/

' V'

I

0.3

I

I

0.5

I

I

0.7

-E (V v. SCE)

Figure 1. The effect of preliminary adsorptive accumulation on the peak height in differential pulse voltammetry of the pesticide Dinobutone (5.10 -8 mol'1-1) in B-R buffer, pH 6.1. Dependence on tacc: Curves (1) 0 s, (2) 60 s, (3) 120 s, (4) 180 s. Eacc -0.3 V. (Reprinted from Ref.[28] by courtesy of Marcel Dekker, Inc.

After these preliminary investigations, the most suitable accumulation potential Eacc is found by examining the dependence of the peak current Ip on Eacc while gradually changing Eacc from an original value o f - 0 . 2 V towards more negative values until a m e a s u r a b l e peak begins to appear. If chemisorption participates in the adsorption process, a more positive Eacc value m u s t often be employed, e.g. +0.1 V. In analytical applications it is sometimes useful to employ a more positive Eacc value so t h a t traces of heavy metals are not deposited on the electrode - these substances m a y be present as impurities in the solution. The optimal accumulation time m u s t be also found. The peak high increases linearly with increasing tacc up to a certain value corresponding to a complete coverage of the electrode by adsorbate or to the m o m e n t of a t t a i n e d equilibrium between the compound adsorbed on the electrode surface and present in the bulk of the solution. Complete coverage is a t t a i n e d after shorter tacc in stirred solutions t h a n in unstirred ones. The tacc value at which the limiting Ip value is a t t a i n e d depends also on the sample concentration. The dependence of Ip on the analyte concentration should be linear over a reasonable wide range. (e.g. from 5"10 .7 mol.1-1 to 1"10 .9 mol.l-1). The method of s t a n d a r d additions can be used for quantitative m e a s u r e m e n t s . Three additions of a s t a n d a r d solution are recommended to ensure t h a t the m e a s u r e d Ip values

117 correspond to the linear part of the calibration curve. When the Ip value does not increase linearly during the s t a n d a r d additions, the sample solution m u s t be diluted or a shorter accumulation time employed. Sometimes helps to perform the accumulation in u n s t i r r e d solutions (Fig. 2.)

4

2 1

03

I

I

I

0.5

0.7

0.9

-E (v v. SCE)

Figure 2. Voltammograms of 2.10 -8 M GeO2 in 0.1 M H2SO4 + 0.15 M pyrocatechol after adsorptive accumulation in unstirred solution. Accumulation time : Curve (1) 0, (2) 360 s, (3) 720 S, (4) 1480 s. Scan rate 20 mV. s-1, pulse amplitude 12.5 mV. (Reprinted from Ref. [16] with kind permission of Elsevier Science).

It is recommended t h a t a blank accumulation experiment be carried out in the pure supporting electrolyte, especially for longer tacc values, because surface active impurities in the solution can also be adsorbed on the electrode (and can yield a parasite peak) or can even affect the accumulation process of the substance of interest even due competitive adsorption. Inhibitive effects from such competitive action can be avoided by using shorter tacc times (e.g. from 15 to 30 s).

2.3.

A p p l i c a t i o n s in e n v i r o n m e n t a l analysis

As mentioned, environmental analysis is in fact mostly applied trace analysis and for trace analysis is from electroanalytical methods ideally suited AdSV as their detection limit for electroactive compounds is in the range of 10 -l~ mol'1-1 concentration. However this value can be only achieved under ideal conditions,

118 which in practice are very rare. The m a i n factor limiting the sensitivity and in m a n y cases even the application of this method is the competitive adsorption of certain other surface active substances t h a t m a y be present in the solution to be analyzed. In such cases a decrease in peak height can occur or mainly at high concentrations of interfering substances the peak can be even eliminated. Thus in respect from the practical point of view the detection limit can be expected to be in the concentration range from 1-10 -s to 1"10 .9 mol.1-1. The scope of application of AdSV ranges from metal trace analysis to analysis of organic compounds and in general to environmental, biochemical, medical, pharmaceutical, toxicological and m a n y others applications. 2.3.1. T r a c e m e t a l d e t e r m i n a t i o n s Most published papers on this method are devoted to the metal trace analysis, exploiting the fact t h a t m a n y metal complexes with organic chelating ligands are adsorbable at the electrode. This property can be utilized in adsorptive accumulation of metal chelates on an electrode after which the reduction of the adsorbed compound is performed w h a t is manifested with a peak formation on the voltammetric curve. This procedure permits sensitive d e t e r m i n a t i o n of metal ions t h a t cannot be determined by anodic stripping v o l t a m m e t r y or which are very difficult or even impossible to determine by conventional polarographic or voltammetric methods. Such examples of the last mentioned category are e.g. A1, Be, Sr, Ba, Ca, Mg, Ge, Si, B, - to mention only few of them, e.g. (Fig. 2.) The most applications of metal ions determination deal with analysis of waters, mainly sea water. Competition between the added ligand and n a t u r a l l y occurring complexing m a t e r i a l provides a means of evaluating the complexing ability of sea w a t e r [15]. In general stripping methods are frequently used for metal speciation in waters. Sometimes AdSV can be used to determine a n u m b e r of cations [16] (such as Cu2+), where the positive potential at which adsorption accumulation is carried out prevents the deposition of some ions (e.g. Pb2§ t h a t would interfere in anodic stripping voltammetry. The sensitivity in AdSV is often greater as the metal is not dissolved in mercury like in anodic stripping v o l t a m m e t r y , but r a t h e r a monomolecular complex layer is formed on the electrode surface. The most extensively method used in practice is the nickel ion d e t e r m i n a t i o n at a mercury electrode as Ni-dimethylglyoximate. The AdSV d e t e r m i n a t i o n of nickel can be carried out in various materials such as water, biological materials, foodstuffs, etc. [17] as well as in lipid fractions of biomaterials [18]. The d e t e r m i n a t i o n limit in w a t e r is 1 ~g'1-1 Ni 2§ In toxicological studies it has been found useful to determine nickel (and also lead, cadmium, and m a n y others) in fingernails [19], where the concentration in contaminated persons is about one order of m a g n i t u d e greater t h a n in urine or blood. Also a glassy carbon electrode covered with a mercury film [20] has been used e.g. for determinations of Ni in biological materials, atmospheric dust in various regions, air-borne ash and rain water. With the same complex forming agent can be determined cobalt.

119 U r a n i u m can be determined in water at concentrations from 0.5 ~tg/1 to 0.2 mg/1 by employing a method based on the adsorptive accumulation of its pyrocatechol complex [21, 22]. M a n y papers are devoted to AdSV determination of a l u m i n i u m and beryllium - for both the conventional polarographic method fails. Thus a l u m i n i u m determination in the concentration range from 1-10-5mo1"1-~ to 1-10 .7 mol'l -~ can be determined after binding A1 into an adsorbable complex with alizarin violet N [23] or cupferron [24]. This method was used mostly for determination of a l u m i n i u m in waters. For determination of beryllium in waters in the concentration range from 1-10 .6 mol'1-1 to 1"10 -s can be used AdSV after binding Be into an adsorbable complex with Beryllon II [25] or Berylon III [26]. A detail extensive review on metal determination using AdSV is published in [26].

2.3.2. Determination of organic compounds As a great percentage of organic substances is characterized by their surface activity (increased often by proper choice of the supporting electrolyte), AdSV permits a relative simple study of their electrochemical behavior and enables their d e t e r m i n a t i o n in the concentration range from 1"10 -6 to 1.10 .9 mol-1-1 (Nevertheless even lower detection limits were found like e.g. 2.5"10 1~ mol'l ~ for the pesticide DNOC (2-methyl-4,6- dinitrophenole) [28]). Some examples of basic p a r a m e t e r s for AdSV determination of environmentally i m p o r t a n t organic compounds are given in Tab. 1. The AdSV method can be employed in trace analysis of a variety of organic compounds which exhibit surface active properties. If the given compound contains an electrochemically reducible or oxidable group, the peak current on the voltammetric curve recorded after completion of the accumulation process corresponds to the reduction (or oxidation) of the total q u a n t i t y of the species accumulated at the electrode (and transported to the electrode during the scan). Though chemical analysis or microanalysis is the m a i n object of AdSV, the investigation of electron transfer processes of biologically i m p o r t a n t molecules are the second, not less i m p o r t a n t field in study of physico-chemical interactions in the environment and h u m a n body. As m a n y organic compounds are hazardous ones AdSV enables not only their monitoring in different moieties of the biota, but supports ecotoxicological research in the clarification of their fate, behavior, metabolism and questions connected with their detoxification/liquidation. AdSV helps in the elucidation of the p a t h w a y of the pollutant from the source to m a n or other object of interest and also in clarification of further transformations into other substances along this pathway, e.g. as a result of the interaction among various pollutants, or as a result of metabolism etc. It is also a m e a n for evaluation of the effectiveness of various processes t h a t prevent the formation of pollutants, or t h a t remove those already formed.

120 Table 1 Basic parameters for AdSV of some environmentally interesting organic substances at a mercury electrode Compound

Supporting electrolyte

Nitrobenzene 2-nitrophenol 4-nitrophenol 2,4-dinitrophenol

B-R, B-R, B-R, B-R,

7 5 5 5

1,8- dinitronapthalene 4,8- dinitronapthalene

B-R, pH 8 B-R, pH 8

2,4-dinitro-l-naphtophenol

0.2 M NaOH

pH pH pH pH

EaccN

E p/V

Ref.

-0.20 0.00

-0.55 -0.25

30 31

-0.20 0.00

31 31

-0.20 -0.20

-0.33 -0.20 -0.35 -0.42 -0.36

-0.50

-0.72

30 30

-0.46 30

-0,77 DNOC

B-R, pH 6.1

-0.20

-0.31

28

-0.44 Dinobuton Prometryne Ametryn

B-R, pH 6.1 B-R, pH 3.5 B-R, pH 3.5

Paraquat Atrazin

Ac, pH 4.6 B-R, pH 2.5

Terbutryn

B-R, pH 4

Metamitron Fenchlorazol-ethyl Chlorhexidin

NH 3, pH 9.7

-0.30 -0.70 -0.70 -0.60

-0.46 -1.05 -1.02 -1.13

-0.80 -0,75

-0.83 -0.94 -1.06

B-R, pH 2.5

-0,25

-0.41

32

0.1 M NaOH

-0.10 0.00

-0.50 -1.53 -0.98

33 34

28 28 28 29 32 32

Neutral red

B-R, pH 9.2 Ac, pH 4.7

0.00 0.00

-0.75 -0.22

36

Azobenzene Lauryl sulphonate* Dodecylbenzene-sulphonate*

1 M NaOH 1 M NaOH

-0.70 -0.70

-1.20 -1.20

16 16

Trichlorobiphenyl*

B-R, pH 6.8

-0.40

-1.05

37

Oil products* and crude oil*

1 M NaOH

-0.70

-1.20

38

35

121 To the group of environmentally important and interesting organic compounds belong agrochemicals like pesticides (Fig. 1), herbicides, insecticides, growth stimulators etc., which with their improving food yield serve like good slaves but with their toxic side effect are bad masters. Thus we are interested in the analysis of their residues in waters, soil, foodstuffs, crop, etc. More details can be found in [39]. Another group of harmful compounds consists in general of chemical carcinogens. Here again not only their chemical analysis in various matrices is important but also the correlation of their electrochemical behavior with their genotoxic properties, the study of the mechanism of their interaction with living cells and their general fate in the environment. Among such substances belong derivatives of azobenzene, aminoazobenzene, phenylmethyltriazene, benzidine, acridine and m a n y others - often serving as coloring matters [39,40]. A n u m b e r of substances that cannot be reduced polarographically can be sometimes determined after their derivatization by introduction of a reducible group such as nitroso-, nitro-, etc. An example of derivative formation is the determination of morphine [41], estrone, estradiol and estriol after nitrosation [42]. Another method based on determination of nitrite makes use of diazotation of aniline and coupling of the resulting diazonium salt with azulene (also 1naphthol as coupling agent, nevertheless with longer coupling time, can be used). Aniline and other aromatic amines can be determined at a similar way [43,44]. Another environmentally important compound formaldehyde - after conversion to an ionic derivative with Girard's reagent T - can be accumulated by adsorption on a Nation coated mercury film electrode followed by voltammetric determination in aqueous solution [45]. Caprolactam in waste and n a t u r a l waters can be determined after its condensation with an azobenzyl chloride compound: the product of this condensation which undergoes adsorption on the electrode, was separated from excess reagent and other interfering species by using TLC [46]. Adsorptive accumulation has been used in interaction studies like the binding of one compound to another one adsorbed on the electrode. As example serves the binding of a n t i t u m o r antibiotics with DNA [48] or the interaction of nucleic acids with enzymes [48] or genotoxic substances [49] or damage caused in vitro to DNA by carcinogenic and mutagenic chemicals [50]. This method was thus devised for screening of chemicals with possible mutagenic action. For environmental screening are recommended also DNA modified electrodes mainly for detection of toxic aromatic amines and potentially for elucidating interactions between intercalating pollutants with DNA [51]. Examples of interactions of biomolecules immobilized at the electrode surface with substances from the solution are given also in [47,48 and 52]. In general for toxicological purposes it would be interesting to perform a detailed study of formation of DNA adducts with toxic compounds. From the toxicological point of view interactions occurring in solution such as antigen - antibody reactions can be studied. From such i m m u n o a s s a y s may be mentioned the reactions of h u m a n serum albumin with a n t i h u m a n serum

122 albumin [53], lymphocytic leukemia cells with monoclonial antibodies raised against them [54]. In these cases the peak current of the compound gradually decreases after addition of the other compound into the solution due to binding of both compounds (in fact it means that a new compound is formed and thus the peak current vs. concentration slope changes.) Nevertheless caution is necessary for to exclude the effect of competitive adsorption - as shall be mentioned also later. Electroinactive compounds (it means compounds that are neither reduced nor oxidized at the electrode) yield in contrary to the previous ones only adsorption/desorption peaks in case they are surface active and accumulate on the electrode. The height of these peaks on the voltammetric curve can be measured by means of differential pulse voltammetry (Fig. 3). These m e a s u r e m e n t s are termed adsorptive stripping t e n s a m m e t r y - AdST. (The method of t e n s a m m e t r y is discussed in more details in the next paragraph, where also some examples of AdST m e a s u r e m e n t s are mentioned). The height of tensammetric peaks obtained in AdST are partly dependent on p a r a m e t e r s similar to those governing electrolytic stripping voltammetry, and partly dependent on the surface active properties - mainly the adsorption p a r a m e t e r s of the particular compound. The stronger adsorption, the higher and narrower are the peaks. The adsorption can be influenced by increasing the concentration of the supporting electrolyte, where sometimes salting-out effects of the compound contribute to an increase in the capacitive phenomena. (This is in contradiction to AdSV of electroactive compounds, where the dilution of the supporting electrolyte

0.2V I

I

4

/",u

L/

J - E (V)

Figure 3. AdST peaks of Diesel oil in 5 M KF. Concentration in mg'l-1 9Curve (1) 0, (2) 0.07, (3) 0.20, (4) 0.33, (5) 0.46. Pulse amplitude 50 mV. Accumulated at open circuit for 120 s. Curves recorded from-1.10 V (vs. SCE). Reprinted from [90] with kind permission of the copyright owner.

123 often leads to an increase of the signal). The adsorptive stripping method can be used only in cases where the compound yields a well developed tensammetric peak already without previous accumulation at concentration of 10 .5 to 10-6mo1"1-1. Such compounds have usually an adsorption coefficient of about 10 .3 mol'l -~ or higher in the respective supporting electrolyte and corresponds roughly to the inverse value of the concentration for a 50 percent coverage of the electrode by the adsorbed substance. This AdST measurements are finding most applications in determination of surface active substances in water like undefined surface active substances, detergents, tensides, polyethylene glycols and petroleum components (Fig. 3) - in Tab. 1 see such compounds labeled with *. Some examples are mentioned e.g. in [5, 36, 55, 89, 90]. In general the scope of compounds which can be determined or studied by AdSV is very broad ranging from the just above mentioned groups of compounds to cancerostatics, vitamins, hormones, antibiotics, pharmaceuticals, food additive dyes, biochemicals, DNA, detergents, crude or motor oils and emulsions. Some representatives from the endless list of compounds which can be analyzed by AdSV are given in Tab. 1.

2.3.3. AdSV m e a s u r e m e n t s in f l o w i n g s y s t e m s ,,Modern analytical chemistry makes wide use of measurements in flowing liquids and the importance of this techniques is steadily increasing" ...is stated in [56]. The combination of the effect of adsorption of the analyte on the electrode surface with the medium exchange principle led to the application of AdSV in flow through systems. Here accumulation at a given potential is carried out during the interval when the carrier solution with the injected sample flows through the detector. This interval thus defines the tacc value. When the sample plug leaves the detector, the stripping process is started either without interrupting the flow or after stopping the flow. The later is usually necessary when a peristaltic pump is used, because the pulses in the carrier stream produce large current oscillations. The use of an isocratic pump, on the other hand, permits the m e a s u r e m e n t of the stripping curve without stopping the flow, as a constant flow is ensured under these conditions. The detectors used are mostly commercial ones like mercury, mercury film, carbon or carbon paste electrodes, that are often employed for electrochemical detection in HPLC. Concerning the flow rate - if this is slow (below 0.5 ml min -1) and the sample volume is small (less than 1 ml), dispersion of the sample plug is limited. The passage of fresh, sample free carrier solution through the detector unit during the reduction or oxidation step ensures electrochemical stripping of the analyte into pure electrolyte with no electrochemical interfering compounds. The application of AdSV in flow through systems improves the selectivity and sensitivity of the determination, simplifies the analytical procedure and increases the sample throughput. Stripping analysis can be on line combined also with inductively coupled plasma-atomic emission spectrometry (ICP-AES) and with inductively coupled plasma-mass spectrometry

124 (ICP-MS), mainly for enhancement of sensitivity (e.g. some radioactive isotopes present in the environment at ppt levels), elimination of detrimental matrix effects, speciation of elements and multielement determinations [57]. Some examples of application of AdSV in flowing systems are given in [58]. In paper [59] an automated system for on-line monitoring of traces of uranium is described, the same arrangement can be used for measurements of other metal ions. 2.4. C r i t i c a l a s s e s s m e n t of t h e A d S V m e t h o d The most important feature of the adsorptive stripping method is its sensitivity. Frequently it is stated that the detection limit for the determination of organic compounds is in the range of 10 1~ mol'1-1 concentration (in the case of metal chelates the DL is about at 10 .9 mol-1-1 and electroinactive compounds is the DL at values of 10 -s mol'l-1). However this values can be obtained mostly only under ideal conditions. One of the most serious complication in the use of AdSV and AdST is mostly the presence of other surface active compounds in the solution examined: competitive adsorption usually occurs, decreasing the height of the peak or even at higher concentrations suppressing the signal. Such undesirable surfactants present in the sample can contribute to full coverage of the electrode surface when using long accumulation and can thus hinder or prevent the application of the method. For instance the detection limit for the pesticide DNOC was found in B-R buffer of pH 6.1 be equal to 5-10 -1~ mol'1-1 [28]. If this supporting electrolyte was mixed in the ratio 1:1 with mountains river water (from a rivulet serving as very good drinking water source), this limit was at the concentration of 8.10 .9 mol'1-1, in case the supporting electrolyte was mixed with polluted river water (Vltava river, near Charles Bridge, Prague), the detection limit dropped to a value of 1"10s mol'1-1. Similar example was described with the pesticide Ametryne in the Rhine water [60]. Interfering effects depend also on the nature of both the analyzed and interfering substance and on their concentration ratio as it was shown in paper [30]. In general, the inhibitive effect of accompanying adsorbable substances can be suppressed by using short tacc values (this holds mainly for weekly adsorbable compounds which need longer accumulation). On the other hand, interferences can be separated e.g. by gel chromatography on Sephadex (Pharmacia Uppsala) [30], by ultrafiltration [61] or different types of extraction methods. Among the most used separation techniques (mainly for isolation of organic compounds from different body fluids) extraction with diethylether, can be mentioned. More details see in [62]. In connection with the determination of metals bound to chelates, it should be mentioned that interferences from surface active compounds and other organic compounds can be prevented (or better said must be prevented) by prior destruction of them by irradiation of the sample with UV in presence of hydrogen peroxide. Only after this operation the ligand solution should be added to the

125 sample. Serious complications in metal determination can occur also due to competitive adsorption of the ligand, the concentration of which must be in excess. Such difficulties arise mainly if the peak potential of the ligand is very close to the peak potential of the metal chelate. Often the conditions are not so ideal as in the mentioned example of nickel determination with dimethylglyoxime. Concluding the AdSV method leads to a great improvement in sensitivity of polarography/voltammetry for determination of surface active organic compounds. As many organic substances possess such properties AdSV has found extensive applications. The method can be employed for concentrations of 1-200 ~g'1-1 (and sometimes from 0.1 ~g'l-1). This sensitivity in determination of organic compounds is similar to that found for metal ions by the anodic stripping method and thus corresponds to a considerable extension of voltammetry in organic trace analysis. This mode of analysis is extended also to many metals which form with complexing agents adsorbable complexes. AdSV thus permits sensitive determination mainly of ions of metals that are difficult or impossible to determine by ASV. It can be concluded, that in general the ASV and the here described AdSV represent at present the most used modes of polarographic/voltammetric analysis.

1

MISCELLANEOUS ELECTROANALYTICAL METHODS FOR D E T E R M I N A T I O N OF S U R F A C E ACTIVE S U B S T A N C E S

Beside the just in detail discussed adsorptive stripping method some other electroanalytical methods are used for determination of SAS in aqueous solutions. These compounds, which represent a large part of dissolved organic matter in natural waters have to be monitored both in production of potable and treatment of sewage waters as the presence of SAS in waters is or can be harmful for living organism. Time-consuming two-phase titration procedures and direct photometric determination are the most widely used methods, nevertheless some electrochemical procedures can be more simple in use [63]. Among procedures enabling the study of the concentration dependent surface activity of compounds present at the solution/mercury electrode interface methods derived from polarography play an important role. One of such polarographic procedures is based on the fact that SAS suppress the polarographic maxima [1] or give rise to adsorption/desorption peaks - sc. tensammetric peaks - on the polarographic curve [64]. On the other hand SAS can influence in some respect the polarographic base line of the "pure " supporting electrolyte (and thus the course or changes of the charging current) [65 and 66]. Among methods for recording the interface activity belong also electrocapillary measurements [67] as will be shown in the chapter 3.3 of this article. All these measurements are important in environmental chemistry.

126

3.1. S u p p r e s s i o n of the p o l a r o g r a p h i c m a x i m a J u s t at the very beginning of polarography at some m e a s u r e m e n t s an abnormal increase of the polarographic current on the rising portion of the polarographic curve to values several times greater t h a n the usual limiting current followed by a discontinuously fall to the normal current value was observed: thus a sharp current peak - the sc. polarographic m a x i m u m on the curve is formed (Fig. 4). Two types of polarographic m a x i m a exist: the m a x i m u m described in the above lines is of the sc. first kind and is mostly observed in solutions of low ionic strength. M a x i m a of the sc. second kind a p p e a r in the limiting current region, are usually rounded and do not fall discontinuously to the normal value of the limiting current. Both types of m a x i m a are caused by increased t r a n s p o r t of the analyte by s t r e a m i n g of the solution around the drop of the dropping mercury electrode and they decrease in presence of SAS in the solution. The detection limit for m a x i m a corresponds to about 0.01 mg'1-1, the m a x i m a of the second kind are often one order of m a g n i t u d e more

-E Figure 4. Polarographic maxima at water quality examination: To 5 ml of 0.01 M KC1 added: curve (1) 5 ml of distilled water, curve (2) 5 ml of tap water. Dropping mercury electrode, on air. Curves recorded from 0 V. Reprinted from [88] with kind permission of the copyright owner.

sensitive. Problems are encountered in the practical utilization of these m a x i m a because of their lack of specificity and dependence on a great m a n y p a r a m e t e r s such as the composition of the supporting electrolyte, p a r a m e t e r s of the mercury dropping electrode, etc. Due the mentioned unspecificity the m a x i m a give mostly only a general picture of surface activity of the sample w h a t is i m p o r t a n t from the e n v i r o n m e n t a l point of view. The most applications are in the field of control of w a t e r purity on SAS (Fig. 4). They are based on the suppression of the

127 polarographic maximum of oxygen or the Hg 2§ ion [68-70], e.g. sea water containing SAS is evaluated by comparison with the calibration curve for a "synthetic" sample of sea water containing Triton X-100 as standard : suitable dilution of the sea water sample (polluted by industrial or city wastes, crude oil products) by distilled water yields the surface activity equivalent to Triton X-100. (Experiments performed in the vicinity of the shore of the Adriatic Sea had a surface activity equivalent to 0.2 to 5 mg-1-1 Triton X-100). The degree of suppression of the polarographic maximum thus corresponds basically to the overall content of SAS in the water sample [71]. The method has been used also for determination of uncharged tensides and anionic tensides in waste waters [72] (e.g. from large laundry facilities). Improved method for the determination of surfactants in fresh potable waters based on suppression of polarographic oxygen maximum is given in paper [73]. According the author this method should be useful in the studies of natural waters and their pretreatment as well as in protection of deep-well waters from pollution e.g. with crude oil or bitumens. The method can be applied in the concentration range from 0.01 to l m g of surfactants in 1 1 of water. Maxima of the second kind has been used for determination of pyrogenes (dead microorganisms and product of microbial metabolism) in distilled water [74]. This method can serve also for monitoring of SAS in the preparation of pyrrogene free water. 3.2. T e n s a m m e t r y The influence of SAS on the shape of the polarographic curve has been already described by Heyrovsky. Namely the capacity of the electrode double layer and thus the charging current necessary to charge the electrode to the desired potential depends on the character of substances present in the solution. At adsorption of SAS on the electrode surface the capacity of the electrode double layer mostly decreases and thus also the charging current (Fig. 5). The capacity of the electrode double layer depends also on the electrode potential - in such case we are speaking about the "differential capacity" which decreases in the potential region where the given compound is adsorbed. At the adsorption and desorption potentials is thus a sudden change in the differential capacity reflected in sharp maxima on the curve of the potential dependence of the differential capacity or of a parameter which is a function of it e.g. the charging current. In addition to impedance bridge measurements which is the most accurate mean for estimations of differential capacity value of the electrode double layer, various methods derived from polarography can be used for measurements of parameters (like the peak current of the tensammetric peak) which are only proportional to the mentioned differential capacity. To those methods belongs also differential pulse polarography/voltammetry. Such measurements are termed tensammetry.

128

-0.01 -1.52

E (V v.SCE)

Figure 5. Tensammetric curve of 2.10 l M cyclohexanol in acetate buffer pH 5.4. a - denotes the curve of the pure base electrolyte. Recorded by means of AC polarography. (Reprinted from Ref. [66] with kind permission from Elsevier Science).

It should be mentioned t h a t the peak height in t e n s a m m e t r y depends on the concentration only over a range of one or one and half order of concentrations reaching at higher concentrations a limiting value. The curve depicting the dependence of the peak height on the concentration corresponds roughly to the shape of the adsorption isotherm. The potential of this peak also depends on the concentration of the substance in solution. The height and shape of the t e n s a m m e t r i c peaks can be affected by traces of some other SAS present in solution. Competitive adsorption causes mostly difficulties in the analysis of mixtures of surface active substances and seldom a different peak for each component is obtained : quantitative evaluation becomes difficult or impossible in such cases and some separation procedures are then necessary. In adsorptive stripping t e n s a m m e t r y such possibility brings often the proper choice of the accumulation potential [75]. Another possibility yield chemometric methods like the combination of the method of calibration area with the s t a n d a r d addition method [76]. T e n s a m m e t r i c methods are mostly used for evaluation of waters on the presence of SAS. Such a method for the determination of undefined surface active substances in distilled, potable and u n t r e a t e d waters is described in ref [77]: in stirred solutions the SAS is first accumulated at the surface of the HMDE (at a potential o f - 0 . 6 V). The ratio of the decrease of the charging current recorded after this accumulation to the current value obtained in only "pure" solutions is t h e n proportional to the concentration of the SAS (in the range from 10 to 300 micrograms per liter - Triton X-100 or Na-dodecylsulphate serves as model or

129 standard SAS for comparative measurements). Waters with higher concentrations of SAS can be analyzed without previous accumulation on the electrode. (Prior to analysis water samples should be acidified with perchloric acid to a concentration of 0.1 mol'll). The determination of polyethylenglycols in river water is described in [78]. In the same paper a method for determination of polyethylene glycols in mixtures of metabolic products of the biodegradation of non-ionic surfactants in the control of aquatic environments is discussed. From the theoretical point of view t e n s a m m e t r y and mainly t e n s a m m e t r y with accumulation (AdST) is suitable for studies of adsorption phenomena, adsorption characteristics, diffusion conditions, and can be combined also with electrocapillary studies as shown in [79]. Structural studies of tensides as well as stability studies of their adsorbed layers at the electrode surface are described in papers [80,81]. Compression of the electrode/solution interface resulting in an accumulation of surfactants in the interface with oversaturation of the surface by the adsorbed species, questions of reorientation of the adsorbate ev. followed by breaking up of the adsorbed layer are mentioned in [82], describing in detail a new method of s. c. Compression Accumulation Techniques. The change of the peak current value with concentration - reflected in the Ip/C slope - can be exploited for studies of structural changes of biomolecules and changes occurring in mixtures of compounds as it was demonstrated for mineral oils in aqueous solutions [38]. Such slope changes can be employed in chemometric evaluation of mixtures. A detailed critical assessment of t e n s a m m e t r y "in day to day routine analysis" can be found in [60]. 3.3.

Electrocapillary measurements

Measurement of electrocapillarity is one of the oldest methods for studying surface activity of substances in solution. The adsorption of SAS on mercury changes its surface tension and this change is reflected in a characteristic m a n n e r in the shape of the electrocapillary curve, which shows the effect of the surface tension of mercury on its potential (Fig. 6) [1,67]. Under conditions of adsorption equilibrium, the decrease in electrocapillarity at constant potential or charge is a function of the concentration of the SAS in solution. The set of electrocapillary curves yields the adsorption p a r a m e t e r s that can be used to evaluate the interfacial and electrochemical properties of the system. A great deal of information can be obtained directly from these curves, such as the adsorption and desorption potential values, zero charge potential, the potential of maximum adsorption. A shift in the zero charge potential to negative or positive values indicates the presence of SAS with cationic or anionic character, respectively, in the supporting electrolyte solution. Electrocapillary data can be obtained e.g. by measuring the drop time of a dropping mercury electrode present in the examined solution. Common electrocapillary m e a s u r e m e n t s permit the study of SAS at concentrations down to levels of 10 .5 mol'1-1 [83]. For measurements at lower concentrations of SAS methods were developed where it

130

is worked in stirred solutions using a dropping mercury electrode with an extremely long drop time, at which the test substance is accumulated [84]. Under such conditions the detection limit is shifted to a concentration of 10 .7 to 10 -s mol-1 1. On these principles determinations of various SAS such as t e t r a a l k y l a m m o n i u m , dextran and crude oil products in w a t e r are based [84,85,89,90]. The detection limit lies in the concentration range from 10 to 100 micrograms per liter.

94 _

/

,N

/z

90

84 I

0

I

I

0.6

I

I

1.0 - E (V v.SCE)

Figure 6. Electrocapillary curves of petroleum standard (0.22 g.m -3) in 0.l M KC1. Curve (1) only base electrolyte, (2) after addition of petroleum, without convective adsorptive accumulation, (3) as (2) but with convective adsorptive accumulation. Reprinted from [89] with kind permission of the copyright owner.

The surface tension has a hyperbolic dependence on concentration and is linear at a sufficiently low concentration range. Electrocapillary m e a s u r e m e n t s can quantitatively express the electrosorption activity in solution in t e r m s of an expression for the change of the surface tension indicating the overall effect of the presence of SAS in solution [89].

131 3.4. O t h e r e l e c t r o c h e m i c a l m e t h o d s For completeness should be added that in electrochemical analysis of surface active compounds also some other methods like potentiometry, conductometry and various amperometric techniques are employed, and eventually used in titrimetric methods. Information in this respect can be found in the article "Electrochemistry and the environment" in [86 and 87]. It should be emphasized that in theoretical electrochemical studies measurements of differential capacity are frequently used.

4.

CONCLUSION

In this article adsorption from two points of view is discussed: in the first one adsorption is regarded as a tool for attaining substantially increased sensitivity in voltammetry based on previous adsorptive accumulation of the analyte at the electrode. This ultratrace analytical method belongs today among the most frequently used and cited electroanalytical methods. On the other hand the second aspect followed in this article is the discussion of possibilities for determination of SAS in environmental chemistry using electroanalytical methods. However it is necessary to be aware that methods discussed in previous chapters are subjects for interferences by other compounds able to be adsorbed on the electrode and are thus unspecific: therefore these methods in routine analysis should be used only with care. As an exhaustive treatment of the problem "adsorption phenomena in electroanalysis" can not be given, only some of the principal possibilities of adsorptive phenomena in electrochemistry to the achievement of a cleaner environment are listed. It should be emphasized that electrochemistry has its place not only in solving analytical problems - mainly in monitoring the polluted environment - but also in repairing some of the unfavorable consequences of industrial and other activities - that means in removal or destruction of pollutants. Here the combination of electrochemistry with adsorption has again its important position like in environmental monitoring.

REFERENCES

I. J. Heyrovsky and J. Kuta, Principles of Polarography, Academic Press, New York, 1965. 2. J. Koryta, Coll. Czech. Chem. Commun., 18 (1953) 206. 3. R. Kalvoda, Coll. Czech. Chem. Commun., 21 (1956)852. 4. R. Kalvoda, Chem. Listy, 54 (1960) 1265. 5. R. Kalvoda and M. Kopanica, Pure and Appl. Chem., 61 (1989) 97. 6. J. Wang, Stripping Analysis, VCH Publishers, Inc., Deerfield Beach, 1985. 7. Chi Chua, D. Jagner and L. Renman, Talanta, 35 (1988) 597.

132 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

J.M. Estela, C. Tomas, A. Cladera and V. Cerda, Critical Reviews in Analytical Chemistry, 25 (1995) 91. D . W . M . Arrigan, Analyst, 119 (1994) 1953. R. P. Baldwin, J. K. Christensen and L. Kryger, Anal. Chem., 58 (1986) 1790. Ch. Faller, A. Meyer and G. Henze, Fresenius J. Anal. Chem., 35 (1996) 279. I. R. Miller and L. Doll, Bioelectrochem. Bioenerg., 24 (1990) 129. J. Wang and Z. Li, J. Electroanal. Chem., 266 (1989) 287. J. Wang and B. A. Freiha, Anal. Chem., 56 (1989). C. M. G. van den Berg and J. R. Donat, Anal. Chim. Acta, 257 (1992) 281. R. Kalvoda, Anal. Chim. Acta, 138 (1982) 11. B. Pihlar, P. Valenta and H. W. Nfirnberg, Z. Anal. Chem., 307 (1981) 337. V. Gemmez-Colos and R. Neeb, Z. Anal. Chem., 327 (1987) 547. B. Gammelgaard and J. R. Andersen, Analyst, 110 (1985) 1197. H. Braun and M. Metzger, Z. Anal. Chem., 318 (1984) 321. N. K. Lam, R. Kalvoda and M. Kopanica, Anal. Chim. Acta, 154 (1983) 79. C. M. G. van den Berg and Z. Q. Huang, Anal. Chim. Acta, 164 (1984) 209. V. Stara and M. Kopanica, Coll. Czech. Chem. Commun., 54 (1989) 370. A. Dostal, H. B. Duong and R. Kalvoda, Chem. Listy, 86 (1992) 847. A. Dostal and R. Kalvoda, Chem. Listy, 86 (1992) 380. Ch. Sun, J. Wang, W. Hu, T. Xie and W. Jin, Anal. Chim. Acta, 259 (1992) 319. M. G. Paneli and A. Voulgaropoulos, Electroanalysis, 5 (1993) 355. H. Benadikova and R. Kalvoda, Anal. Lett., 17 (1984) 1519. R. Kalvoda, Proc. of Conference on Trace Analysis of Organic Compounds in Waters, Pardubice, 1984, 166. R. Kalvoda, Anal. Chim. Acta, 162 (1984) 197. J. Barek, H. Ebertova, V. Mejstrik and J. Zima, Coll. Czech. Chem. Commun., 59 (1994) 1761. J. Skopalova and M. Kotoucek, F. J. Anal. Chem., 351 (1995) 650. A. Meyer and G. Henze, F. J. Anal. Chem., 350 (1994) 150. P. Gratteri, S. Pinzauti, E. Laporto, G. Papeschi, V. Cavrini and G. Santoni, Analyst, 116 (1991) 723. G. Xu, J. J. O'Dea, A. Mahoney and J. G. Osteryoung, Anal. Chem., 66 (1994) 808. M. S. Ibrahim, Z. A. Ahmed, M. M. Kamal and Y. M. Temerk, Coll. Czech. Chem. Commun., 60 (1995) 65. N. K. Lam and M. Kopanica, Anal. Chim. Acta, 161 (1984) 315. M. Kovac, R. Kalvoda, L. Novotny and A. Berka, Electroanalysis, 5 (1993) 171. K. Stulik and R. Kalvoda (eds.), Electrochemistry for Environmental Protection, UNESCO Venice Office, Venice, 1996. Y. Castrillejo, R. Pardo, E. Barrado and P. S. Batanero, Electroanalysis, 2 (1990) 553. R. Kalvoda and M. Kopanica, not yet published.

133 42. Z. Zhaou, S. Hu, Q. He and Y. Yan, Proc. 2nd Beijing Conf. and Exhibition on Instrumental Analysis, Beijing, (1987) 1199. 43. A. G. Fogg and R. M. Alonso, Analyst, 113 (1988) 1337. 44. J. A. Rodriguez and A. A. Barros, Talanta, 42 (1995) 915. 45. Wing Hong Chan and Hao Huang, Analyst, 121 (1996) 1727. 46. Z. Tocksteinova and M. Kopanica, Anal. Chim. Acta, 191 (1987) 77. 47. J. M. Plambeck and J. W. Lown, J. Electrochem. Soc., 131 (1986) 359. 48. E. Palecek and I. Postbiegelova, J. Electroanal. Chem., 214 (1986) 359. 49. J. M. Sequaris, Electrochemical analysis of nucleic acids, in: R. M. Smyth and J. G. Vos (eds.) Electrochemistry, Sensors and Analysis, Elsevier, Amsterdam, 1986, 191. 50. J. M. Sequaris, H. W. Nfirnberg and P. Walenta, Toxicol. Environ. Chem., 10 (~ 9s5) s3. 51. J. Wang, G. Rivas, D. Luo, F. S. Valera and N. Donthe, Anal. Chem., 68 (1996) 4365. 52. E. Palecek, Anal. Biochem., 170 (1988) 42. 53. J. R. Flores and M. R. Smyth, J. Electroanal. Chem., 233 (1987) 317. 54. B. Lannon, R. O'Kenedy, Jr. Flores, J. J. Fennelly and M. R. Smyth, Electroanalysis, 1 (1989) 431. 55. R. Kalvoda, Pure and Appl. Chem., 59 (1987) 715. 56. K. Stulik and V. Pacakova, Electroanalytical measurements in flowing liquids, E. Horwood, Chichester, 1987. 57. F. Zhou, Electroanalysis, 8 (1996) 855. 58. R. Kalvoda, Adsorptive stripping voltammetry in trace analysis, in: A. Ivaska, A. Lewenstam and R. Sara (eds.), Contemporary Electroanalytical Chemistry, Plenum, New York, 1990, 403. 59. J. Wang, R. Setiadji, L. Chen, J. Lu and S. G. Morton, Electroanalysis, 4 (1992) 161. 60. P. M. Bersier and J. Bersier, Applied Polarography and Voltammetry in Dayto-Day Environmental Analysis, in: A. Ivaska, A. Lewenstam and R. Sara (eds.), Contemporary Electroanalytical Chemistry, Plenum, New York, 1990, 109. 61. Z. Tocksteinova and R. Kalvoda, Chem. Listy, 82 (1988) 1209. 62. R. Kalvoda, F. J. Anal. Chem., 349 (1994) 565. 63. P. M. Bersier, Chem. Listy, 89 (1995) 742. 64. B. Breyer and H. H. Bauer, Alternating Current Polarography and Tensammetry, Interscience, New York, 1963. 65. J. Heyrovsky, F. Sorm and J. Forejt, Coll. Czech. Chem. Commun., 12 (1947) 11. 66. R. Kalvoda, J. Electroanal. Chem., 4 (1960) 314. 67. H. Jehring, Elektrosorptionsanalyse, Akademieverlag, Berlin, 1975. 68. T. Zvonaric, V. Zutic and M. Branica, Thalassia Jugoslavica, 9 (1973) 65. 69. B. Cosovic, V. Zutic and Z. Kozarac, Croat. Chem. Acta, 50 (1977) 229. 70. V. Zutic, B. Cosovic and Z. Kozarac, J. Electroanal. Chem., 78 (1977) 113.

134 71. Z. Kozarac, T. Zvonovic, V. Zutic and B. Cosovic, Thalassia Jugoslavica, 13 (1977) 109. 72. Z. Kozarac, V. Zutic and B. Cosovic, Tenside Detergents, 13 (1976) 260. 73. E. Bednarkiewicz, Electroanalysis, 3 (1991) 839. 74. K. Micka and R. Kalvoda, Czechoslov. Pharm., 5 (1956) 138. 75. A. Szymanski and Z. Lukaszewski, Anal. Chim. Acta, 281 (1993) 443. 76. J. Krista, N. Novotny, M. Drevinek, M. Kopanica and T. Navratil, Proc. Conf. on Analysis of Organic Pollutants, Komorni Lhotka, Czech Rep., 1996, 112, 133. 77. E. Bednarkiewicz, M. Donten and Z. Kublik, J. Electroanal. Chem., 127 (1981) 241. 78. A. Szymanski and Z. Lukaszewski, Analyst, 121 (1996) 1897. 79. Z. Lukaszewski, M. K. Pawlak, The behaviour of oxyethylatedalcohols under conditions of tensammetry with adsorptive accumulation on HMDE, in: M. R. Smyth and J. G. Vos (eds.) Electrochemistry, Sensors and Analysis, Elsevier, Amsterdam, 1986, 119. 80. M. K. Pawlak, Coll. Czech. Chem. Commun., 56 (1991) 78. 81. Z. Lukaszewski, H. Batycka and W. Zembrzuski, Anal. Chim. Acta, 175 (1985) 55. 82. L. Novotny, Electroanalysis, 8 (1996) 135. 83. B. Behr and M. Drogowska, J. Electroanal. Chem., 82 (1977) 317. 84. L. Novotny and I. Smoler, J. Electroanal. Chem., 146 (1983) 183. 85. L. Novotny, I. Smoler and J. Kuta, Coll. Czech. Chem. Commun., 48 (1983) 964. 86. R. Rajeshwar, I. G. Ibanez and G. M. Swain, J. Appl. Chem., 24 (1994) 1077. 87. M. Gerlache, J. M. Kauffmann, G. Quarin, J. C. Vire, G. A. Bryant and J. M. Talbot, Talanta 43 (1996) 507. 88. J. Heyrovsky and P. Zuman, Uvod do prakticke polarografie (Practical polarography - Introduction), Academia, Prague, 1964. 89. L. Novotny and R. Kalvoda, Coll. Czech. Chem. Commun., 51 (1986) 1595. 90. R. Kalvoda and L. Novotny, Coll. Czech. Chem. Commun., 51 (1986) 1587.

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski(Editor) 9 1998 Elsevier Science B.V. All rights reserved.

135

Resolved and unresolved questions of analysis of surfactants in the aquatic environment Z.Lukaszewski Institute of Chemistry, Technical University of Poznan, ul Piotrowo 3, PL-60-965 Poznan, Poland

1. A B S T R A C T

214 references concerning the analysis of surfactants in the aquatic environment have been reviewed. Methods for the analysis of anionic and cationic surfactants are critically discussed, but closest attention is paid to non-ionic surfactants, as this type of surfactant gives rise to the most serious unresolved questions of analysis. Typical levels of concentration of ionic and non-ionic surfactants in raw and treated sewage, surface water and river sediments are given. Recently developed tensammetric techniques, useful for the analysis of non-ionic surfactants and their metabolites, are broadly described. The following unresolved questions related to anionic surfactants are emphasised: specific determination of different classes of anionic surfactants (alkyl sulphates, alkylether sulphates), analysis of metabolites of anionic and non-ionic surfactants, and the problem of 'pseudosurfactants'. In the case of cationic surfactants, the question of inconclusive results of measurements of aquatic environment samples is selected as the most serious problem. The need to develop analytical tools for the investigation of interactions between cationic and anionic surfactants in the aquatic environment is also emphasised. The following unresolved questions are distinguished for non-ionic surfactants: selection of method for determining the total concentration of non-ionic surfactants capable of approval in interlaboratory tests, selection of the s t a n d a r d surfactant representative for the mixture of non-ionic surfactants in the aquatic environment, and the development of methods for determining non-ionic surfactants having less t h a n 5 oxyethylene subunits, as well as those having more t h a n 30 subunits. The development of methods for the specific determination of such classes of ethoxylates as oxyethylated amines, oxyethylated fatty acids, as well as methods for the specific determination of newly introduced non-ethoxylate non-ionic surfactants such as alkyl polyglucosides and metabolites of non-ionic surfactants is suggested. Methods for the trace analysis of non-ionic surfactants also require development as do methods suitable for controlling biodegradation at a realistic level of concentration. The necessity for the future development of detectors for HPLC

136 and FIA of oxyethylated alcohols and in the automation of the analysis of nonionic and cationic surfactants is emphasised. 2. I N T R O D U C T I O N

Because of their surface active properties surfactants are widely used in detergent formulations as well as in wetting agents, emulsifiers and dispersants. They are also frequently used as chemical reagents [1]. However, the m a i n application which may have an influence on the aquatic environment, is in the m a n u f a c t u r e of laundry detergents and cleaning agents. Basically, the mass of surfactants used in laundering and cleaning is directed to sewage, and then to sewage t r e a t m e n t plants or, simply directly to the aquatic environment. This creates a massive s t r e a m of synthetic organic carbon directed to surface w a t e r and may be the largest s t r e a m of synthetic organic carbon charging the aquatic environment [2]. Only efficient biodegradation of surfactants m a y reduce their influence on the environment to an acceptable minimum. Therefore, the total concentration of every type of surfactant and their major metabolites should be efficiently controlled in both raw and treated sewage as well as in surface and aquifer water. Deposition of surfactants and their metabolites in river sediments should also be kept under control. The other task of surfactant analysis is the development of efficient methods of control during biodegradation tests. This work aims to show which problems of control are resolved and which r e m a i n unresolved. Unfortunately, the list of unresolved problems in the analytical control of surfactants in the aquatic environment is much longer t h a n the list of resolved ones. The reasons for the unsatisfactory state of analysis are the complexity of w a t e r matrices, the supermulticomponent character of the analysed mixture and the strong surface active properties of surfactants. Humic and fulvic acids, as well as carboxylic acids, may occur in surface w a t e r as its n a t u r a l components along with antropogenic phenols, lignin-sulphonates and petroleum products [3]. The mixture of surfactants is also extremely complex. Four m a i n types of surfactants may be distinguished: anionic, cationic, non-ionic and amphoteric once. In terms of output the leading type is anionic surfactants, which constitute 59% [4] or 51% [5] of the total output of surfactants. The non-ionic surfactant share was e s t i m a t e d at 33% [4] or 37% [5] and the output of cationic surfactants at 7% [4] or 10% [5], with amphoteric one, at 1% [4] or 2% [5]. The ratio of output of different types and classes of surfactants may serve as a highly approximate m e a s u r e of the ratio of these surfactants in raw sewage and surface w a t e r if the analytical data is not available. However, this ratio may change due to the differing biodegradability of different classes of surfactants. The aquatic environment should be rich in barely biodegradable surfactants, while the concentration of easily biodegraded surfactants should be lower t h a n it might be, based on the level of output of these classes. Non-ionic surfactants represent a much higher level of complexity t h a n other types of surfactants. Several h u n d r e d

137 of individual substances m a y occur in a mixture of non-ionic surfactants. The main reason for such a high n u m b e r of individual compounds in the mixture is the polidispersity of ethoxylates. This complexity makes the analysis of non-ionic surfactants much more difficult t h a n in other types of surfactants. This is why much more attention is paid in this study to non-ionic surfactant analysis t h a n to ionic surfactants. The new t e n s a m m e t r i c techniques in the analysis of non-ionic surfactants, developed in the author's laboratory, are broadly described in contrast to the other techniques and methods.

3. INITIAL S T E P S OF SURFACTANT ANALYSIS 3.1. P r o b l e m s related to the strong a d s o r p t i v e ability of s u r f a c t a n t s Apart from general problems connected with the initial steps of analysis, surfactants provide several additional problems related to their strong adsorptive ability and generally fast biodegradability. The strong adsorptive accumulation of surfactants on the liquid-air surface should be t a k e n into account. A sample of surface w a t e r or sewage collected from the water surface or sewage is much more concentrated t h a n the bulk of the liquid. Thus, the sample m u s t be t a k e n from below the w a t e r or sewage surface (usually at a depth of 1 m). An adsorptive preconcentration of surfactants will take a place on the surface of sampling vessels or containers used for storage samples. 55% of Triton X-100 from 100 gg spike was lost on the polyethylene container during storage, and 33% on the glass container [6]. Generally, this loss is governed by the adsorption isotherm of the surfactant on the container material and should be negligible for the high concentration range of surfactants (sewage) but very significant for the lower concentration range (surface water). To prevent this adsorptive loss of surfactants on the surface of the container, it should be conditioned with an additional portion of sampled water before filling. Another problem related to the adsorption of surfactants on the surface of vessels is the 'memory' of the vessel. Surfactants adsorbed on the surface of the vessel in one m e a s u r e m e n t may be released during the next m e a s u r e m e n t (if a new concentration is lower t h a n the previous one). To prevent this effect, vessels should be carefully cleaned between m e a s u r e m e n t s ; methanol may be recommended for this purpose. A very serious error may be the use of surfactant-containing detergents for cleaning the vessels. Despite residual surfactants, polymer containers may contain plasticiser, which may be washed out by the sample and be a hindrance in the further stages of the analysis. P a r t of the total a m o u n t of surfactants contained in the sample occurs in the form of adsorbed on solid particles of sewage or surface water. Though the problem will be further considered in detail, it is necessary to stress t h a t the total flux of surfactants in a river or in sewage consists of two components: soluble and adsorbed on particles [7,8]. In order to separate soluble surfactants from those adsorbed on particles, the filtration of samples is frequently used. However, the

138 filter used may also exhibit a certain adsorptive ability, which can not be ignored. This ability is governed by adsorptive equilibrium and kinetic factors. The initial volume of the filtrate has a diminished concentration due to the adsorptive loss. This volume depends on the concentration of surfactants in the sample, the adsorptive ability of the filter and the filtration rate. These factors decide how much volume of the filtrate should be discarded. In the case of sewage the first 10 - 20 ml should be discarded, and the first 200 ml in the case of surface water. The adsorption of surfactants also takes place on the polymer tubings which may be used at different stages of the analysis. The consequence of this effect is the loss of surfactants when the first sample is processed, and the 'memory' of the tubing during the use of subsequent solutions. The effect strongly limits the application of polymer tubings in the analysis and is also a serious barrier to the application of the flow injection analysis for surfactants.

3.2. P r e s e r v a t i o n of s a m p l e s Many surfactants in the aquatic environment undergo relatively fast biodegradation. This especially concerns samples rich in microfauna such as raw and biologically treated sewage as well as surface water. Even a period of several hours may cause a serious drop in surfactant concentration [9]. Therefore, immediate and effective preservation of the aquatic environment samples is necessary. Usually 1% formaline is successfully used [9,10]. Other preservants were found to be ineffective (chloroform) or only effective for several days storage (mercury(II) and copper(II) [9]). Refrigerating the samples as an isolated measure against biogegradation was also ineffective. Some barely biodegradable surfactants such as oxyethylated alkylphenols may be more easily preserved [10]. Even samples preserved with 1% formaline may exhibit a loss of surfactant when stored for more than one year [11]. Surfactants contained in the samples may be stored over a period of several years in the form of dry ethyl acetate extracts containing chloroform. Because the surfactant concentration in sewage may vary throughout the day, samples over the 24-hour period are taken to give the average value for a specific day. However, the biodegradation of surfactants in the collection tank is still in process and the sample may contain less surfactants. In the opinion of the author, this problem should be carefully analysed. 4. SEPARATION OF SURFACTANTS AND THEIR METABOLITES None of the methods used for the determination of surfactants in the aquatic environment samples is used without the preliminary separation and preconcentration of surfactants. In the majority of cases, a complex multistage separation scheme is required prior to the determination itself. Liquid-liquid extraction is a frequently used separation technique. Anionic surfactants are separated by extraction with chloroform in the form of ion-pairs with cationic methylene blue [12-16]. Similarly, cationic surfactants are extracted with

139 chloroform as the ion-pair with anionic disulphine blue [17-22]. Non-ionic surfactants are separated from the water matrices by extraction with ethyl acetate. Only ethoxylates having less than 30 oxyethylene subunits are separated this way [6]. Sequential extraction with ethyl acetate and chloroform is used for the simultaneous separation and determination of non-ionic surfactants ethoxylates, having more than 30 oxyethylene subunits and poly(ethylene glycols) (PEG) [24-26]. A specific form of extraction technique is the gas-stripping of surfactants, where the surfactants are transferred to the ethyl acetate layer [27]. Non-ionic surfactants [27], anionic [28] and even cationic surfactants [18] may be separated this way. Gas-stripping was also applied for the separation of linear alkylbenzene sulphonate (LAS) from sea water [29]. Though a nitrogen stream is used in this technique to transport surfactants from the water phase to the organic phase, equilibrium between the water and organic phases must 'govern' the system i.e. the same factor as in liquid-liquid extraction. Thus the same degree of separation should be achieved both in gas-stripping and liquid-liquid extraction, using the same solvent and ratio of phases. Certainly, these two techniques may differ in the kinetics of extraction. It is worth mentioning that the gas-stripping technique is very convenient for processing large volumes of the water sample. The longchain ethoxylates are lost in both gas-stripping and liquid-liquid extraction with ethyl acetate [6,30]. It is surprising that some surfactants are located on the boundary of phases during gas-stripping separation, and they disappear from both the water and organic phases [6]. This effect, negative from the point of view of successful separation, seems to be a consequence of the surfactant structure. A very interesting extractive technique is the combination of the exhaustive steam distillation and solvent extraction developed by Giger and Ahel [31] for the separation of metabolites of oxyethylated alkylphenols. Recently, supercritical fluid extraction was applied for the separation of cationic surfactants from dried activated sludge [32]. Ion-exchange resins are usually used for clean-up procedures as an intermediate stage in the procedural chain. Strong anion-exchange resin, applied as a solid-phase extraction cartridge SAX, is used as an intermediate stage in the determination of LAS by gas-chromatography or HPLC [29,33-39]. Strong anionexchange resins are also used in the analysis of surfactants to split ion-pairs of cationic surfactants-to-be-determined with LAS added at the initial stage of analysis [19,20,40]. Cationic surfactants are freed at this stage. The mixed-bad ion-exchange column is used to clean up the sample containing non-ionic surfactants-to-be-determined in order to stop interference of ionic surfactants

[10,41]. Solid-phase extraction is used in separation schemes for the specific determination of LAS or, the determination of non-ionic surfactants. A C18 or C8 reverse phase cartridge [29,38,42] or graphitised carbon black cartridge [43,44] is used for the preconcentration of LAS. Macroreticular adsorbent Amberlite XAD-8 was also used for the extreme preconcentration of anionic surfactants [45]. An

140 adsorption of non-ionic surfactants on macroreticular XAD-2 is recommended at the initial stage for the CTAS procedure, or for the specific determination of oxyethylated alkylphenols or oxyethylated alcohols [10,23]. Effective preconcentration of non-ionic surfactants in Teflon tubes is worth further study [46]. 5. I O N I C S U R F A C T A N T S

5.1.

Anionic surfactants

Anionic surfactants are a leading group of surfactants in terms of manufacture. 59% of the total production of surfactants are surfactants of this type [4]. LAS, alkyloxyethyleneether sulphates (alkylether sulphates) and alkyl sulphates belong to the major classes of anionic surfactants. The ratio of these classes is approximately 7.5 : 4 : 3. [47], although substantial regional differences may occur in the use of different classes [4,5,47,48]. The vast majority of LAS and alkyl sulphates is used in laundry detergents while alkylether sulphates are mostly used in shampoo and foaming detergent formulations [49]. 5.1.1. N o n - s p e c i f i c m e t h o d s of a n a l y s i s 5.1.1.1. The MBAS m e t h o d The main tool for the determination of the total concentration of anionic surfactants in the environmental samples is the Methylene Blue Active Substances method [12,13]. The results of this non-specific determination are known as the MBAS concentration, and the principle of the procedure is the formation of ion-pairs between anionic surfactants and cationic methylene blue. This ion-pair is extracted with chloroform and the concentration is validated spectrophotometrically ()~= 650 nm). The detection limit of the method was evaluated at 20 gg l-I [36]. Humic substances, phenols and inorganic chlorides, nitrates and thiocyanates were identified as giving positive errors while quaternary ammonium salts and fatty amines gave negative errors in the MBAS analysis [14]. Where the concentration of anionic surfactants is below the capacity of the simple extraction procedure, a gas-stripping preconcentration to ethyl acetate in a stream of nitrogen is recommended [28]. In such a case the water phase should contain a high concentration of sodium chloride. XAD-8 macroreticular resin was also used for the extreme preconcentration [45]. The MBAS procedure is commonly used for the determination of anionic surfactants in surface water, raw sewage, processing liquors, effluents of sewage treatment plants as well as in biodegradation studies, and is an internationally recommended method [14,15,16]. New methods combining the extraction with spectrophotometry are in the process of development [50-54], but none have found a wide application. Several attempts at achieving a spectrophotometric determination of anionic surfactants in the flow-injection systems are worth mentioning [55-58]. The first

141 attempt was based on liquid-liquid extraction realised in the flow-injection system [55], but more recent works [56-58] used solvatochromism i.e. a spectrum shift caused by anionic surfactants. Despite a very high rate of measurements of the order of one sample per minute, the recent applicability of these methods to the environmental samples is limited, due to the too high a range of concentration measured by these methods. 5.1.1.2. T i t r a t i o n m e t h o d s The two-phase titration[59-68] and potentiometric titration [69-82] methods have a degree of potential for the determination of anionic surfactants in the environmental samples, although the optimal range of their application corresponds to a higher range of concentration more characteristic of detergent formulations than the environmental samples. Both methods work within a similar range of concentration and in some recent studies they are used alternatively [68]. Both methods are based on the formation of ion-pairs of anionic and cationic surfactants. The two-phase titration method, also called titration, by Epton [59], is performed in a heterogeneous water/chloroform mixture in the presence of cationic dye. Initially, the coloured ion-pair of anionic surfactant(s) and dye is dissolved in the chloroform layer. Colourless strong cationic surfactant is used as a titrant. The end point of the titration is determined visually. At this point the dye replaced by the cationic titrant passes to the water layer. The cationic surfactant Hyamine 1622 (p-tert-octylphenoxyethoxydimethylbenzyl ammonium chloride) is frequently used as a titrant but cetyltrimethylammonium bromide, lauryltrimethylammonium bromide, lauryldimethylalkyl(C12_14)ammonium chloride and tetraphenylphosphonium chloride were also applied. Though many coloured cationic dyes, including methylene blue, may be used as indicators, the mixed indicator consisting of cationic dimidium bromide and anionic disulphine blue VN is recommended [60-62]. The lowest concentration using two-phase titration was reported by Tsubouchi et al [63-65]. Using tetraphenylphosphonium chloride titrant and a tetrabromophtalein ethyl ester indicator they detected 12.5 pg in the sample. In the potentiometrically indicated precipitation titration a cationic titrant forms a precipitate with the determined anionic surfactant-to-be-determined. An excess portion of the titrant causes a jump in potential of an ion-selective electrode indicated as an end point on the titration curve. Usually a liquid membrane or a coated wire electrode was used as well as Hyamine, cetyltrimethyl ammonium bromide, cetylpiridinium bromide or cetyldibenzylammonium chloride as titrants [68-82]. The range of hundreds of pg and a single mg in the sample is the most suitable for titration [69,70]. On the other hand, the linear response range for direct potentiometric measurements was determined as 1• .7 - l• [70]. The range of concentration of the twophase and potentiometric titration used shows that these methods may be useful in monitoring raw sewage and in biodegradation tests but hardly useful for

142 monitoring effluents of sewage treatment plants and surface water. Nowadays, the MBAS or LAS specific HPLC methods are usually used for these purposes. 5.1.2. Specific m e t h o d s of analysis Linear alkylbenzene sulphonates and their metabolites sulphophenyl carboxylates are the only specifically determined anionic surfactants. This is due to LAS being the main anionic surfactant as well as the availability of analytical methods. LAS constitutes 50-60% of the total production of anionic surfactants and approximately 25% of the total production of surfactants in Western Europe [5,47,49]. In contrast to the other anionic surfactants, LAS contains the phenyl ring which makes chromatographic detection easier. Worth mentioning is the gas-chromatography method for the specific determination of aklyether sulphates developed by Neubecker [83] and the determination of the fraction of organic sulphates (alkyl sulphates and alkylether sulphates) developed by Oba et al [84], though not widely used. The detection and specific determination of sodium dodecyl sulphate, LAS and branched-chain alkylbenzene sulphonate may be done using carbon-13 nuclear magnetic resonance spectrometry [45]. However, a three hundred litre water sample is required for processing prior to the determination itself.

5.1.2.1. Specific determination of linear alkylbenzene sulphonate LAS are determined in different environmental samples using gas chromatography(GC) or high performance liquid chromatography (HPLC). Most of the GC schemes required desulphonation of LAS [33,85-95]. This was usually done by hydrolysis of LAS in acidic aqueous solution. The corresponding mixture of alkybenzenes which was produced was separated (e.g. by liquid-liquid extraction to heptane [95]) and determined by GC. Orthophosphoric acid for hydrolysis and a flame-ionisation detector were often used [92-94]. The derivatization of preliminary separated LAS is another option [96,97]. The LAS ion pair formed with tetrabutylammonium cation transforms into its butyl esters in the injection port of the chromatograph (300 ~ A hyphened GC-MS technique was also successfully applied [98-100]. Since the mid-eighties HPLC rather than GC has been used for the determination of LAS [29,33-39,42-44,101-105]. The determination requires several separation stages, although the first paper concerning the subject by Nakae et al [101], predicted no pretreatment. A C18 or Cs reverse phase solid phase extraction cartridge was often used for initial separation [29,38,39,42]. However, preconcentration by evaporation of a water sample was also utilised [33,36]. Gas stripping separation of LAS from sea water samples to ethyl acetate constituted another preconcentration approach [29]. Ion exchange separation is usually used as a subsequent stage [29,33-39], as well as the second adsorption into the C18 or C8 reverse phase cartridge for the final cleaning of the sample. Preconcentration and cleaning may be achieved in a single stage using a graphitized carbon black cartridge [43,44]. Specific homologues and isomers of

143 LAS, s e p a r a t e d on a C18 or Cs reverse phase chromatographic column are detected using a fluorimetric detector (excitation at ~.= 230 or 225 nm and emission at ~.= 290 or 295 nm) [33,34,36,43,44,101] or UV detector (k= 225 nm) [39]. Different versions of HPLC procedures show a detection limit of the order of 10 pg 11 [34,36] or 0.3 ~g in the sample [39] i.e. considerably better t h a n GC procedures. The possibility of recording changes of homologue distribution of LAS is an i m p o r t a n t a d v a n t a g e of the HPLC procedures. Apart from LAS, its main biodegradation intermediates, sulphophenyl carboxylates, m a y also be determined by the HPLC [106].

5.1.3. C h a r a c t e r i s t i c c o n c e n t r a t i o n s in the a q u a t i c e n v i r o n m e n t Concentration of anionic surfactants in components of the aquatic environment may be expressed both as the total concentration determined by the MBAS procedure, and the concentration of LAS, which is the m a i n anionic surfactant. The LAS results are equivalent to 55-93% of the MBAS results, in the case of raw sewage, 18-53%, in the case of biologically treated sewage, 11-50% in the case of river water and 2-38%, in the case of river sediment [33]. Raw sewage concentration of certain surfactants depends mainly on their use and the consumption of w a t e r by the population. The concentration of anionic surfactants (as determined by MBAS) in raw sewage depends very much on the country. In the USA, C a n a d a and Switzerland the concentration of anionic surfactants in raw sewage was estimated at 2 - 5 mg 1-1 [107-109] while in Germany, UK and Spain, at 4 - 21 mg 1-1 [107,108,110,111]. The concentration of LAS in the final effluents of numerous sewage t r e a t m e n t plants using activated sludge was d e t e r m i n e d at 0.02 - 1.0 mg 11, but typically at 0.05 - 0.10 mg 1-1 [107]. Higher results were observed in Spain (0.04 - 0.40 mg 1-1, typically approximately 0.15 mg 1-1 [111]) and in the UK (0.1-0.3 mg 1-1 [94]), which corresponds to a usually higher concentration of anionic surfactants in the raw sewage of these countries [94,111]. Concentration of LAS adsorbed on activated sludge was determined at 0.09 - 0.86 g kg -1 dry m a t t e r [107]. Apart from LAS, sulphophenyl carboxylates, being the main soluble metabolite of LAS biodegradation [112] were monitored in the raw sewage and in the effluent of a lagoon t r e a t m e n t plant [113]. A concentration of sulphophenyl carboxylates in the order of 1 mg 11 was found both in the raw and in the t r e a t e d sewage. It is necessary to stress t h a t the concentration of sulphophenyl carboxylates in the effluent is twice higher than the residual concentration of LAS. The most detailed published data for the concentration of anionic surfactants in surface w a t e r concerns four rivers in west Germany: Rhine, Ruhr, Main and Neckar [114-118]. Very well documented MBAS based monitoring clearly shows long term tendencies and seasonal variations in the concentration of anionic surfactants. The level of 400-600 ~g 1-1 , characteristic for the period before 1964, dropped to a level basically below 100 ~g 1-1 in the eighties and still exhibits a slight tendency to go lower [116]. The main reason for such a radically improving situation in relation to anionic surfactants was the production ban of barely

144 biodegradable tetrapropylene derived alkylbenzene sulphonate and its replacement with the more environmentally friendly LAS. The concentration of LAS in these rivers was determined at 9 - 35 ~g 1-1 which constitutes of 28% of the MBAS result. In the UK, the average concentration in 1982 as determined by the MBAS method from 35 measuring points was 150 ~g 1-1 , and the LAS concentration was 26% of the MBAS result [94]. Several American rivers showed a level of LAS concentration of 10-40 ~g 11 [118]. The Warta River in Poznan (Poland), monitored over 7 years using the MBAS procedure, showed a relatively stable level of approximately 200 ~g 1-1 [120]. Another tendency in the monitoring results of anionic surfactants is a distinctly higher winter level t h a n s u m m e r level [119,120]. The results concerning anionic surfactants in sea water were rarely published and only concerned the LAS concentration. This is due to a better detection limit of the HPLC method for LAS t h a n the MBAS. The concentration of LAS found in Tokyo Bay ranged from 0.8 to 30 ~g 11 [121] while in the North Sea, in the Scheldt River estuary on the Dutch coast, concentration of LAS ranged from 0.5 to 9.4 ~g 11 , showing a gradual decrease towards the open sea [29]. Ground water from 500 m and 3000 m wells located in the vicinity of sewage infiltration ponds contained 0.4 - 2.5 mg 1-1 of anionic surfactants as determined by the MBAS [45]. Most of these surfactants were identified as branched-chain alkylbenzene sulphonates. A report concerning tap water at an unspecified location in J a p a n showed a concentration of LAS of 71 ~g 1-1 [121]. Wickbold [122] however, by applying a preconcentration from 10 litres of tap water in Dfisseldorf, G e r m a n y did not detect either LAS nor its metabolites. The degree of biodegradation of anionic surfactants in sewage t r e a t m e n t plants using activated sludge was estimated at 93-98%, while LAS was biodegraded in 95 - 99.5% [107]. In sewage t r e a t m e n t plants using biological filters the degree of biodegradation of anionic surfactants is lower and was estimated as a 85-88% reduction in concentration of anionic surfactants (determined by the MBAS method) or a 73-91% reduction in LAS concentration. It is necessary to add that under model conditions, anionic surfactants determined using the MBAS method were reduced 89-100%.

5.1.4. U n r e s o l v e d questions in the analysis of anionic surfactants Four main unresolved questions may be specified: i. the specific determination of different classes of anionic surfactants (alkyl sulphates, alkylether sulphates), ii. the analysis of metabolites of anionic surfactants iii. partially biodegraded non-ionic surfactants of anionic nature iv. 'pseudosurfactants' At least two additional classes of anionic surfactants should be monitored in the aquatic environment: alkyl sulphates and alkylether sulphates. Though LAS remains the main surfactant in terms of output at approximately half of the demand for anionic surfactants and a quarter of total surfactant demand, the two

145 classes also represent a significant participation in the total consumption of surfactants; alkylether sulphates constitute approximately 20 - 25% and alkyl sulphates, 10-20% [5,47]. Rapidly increasing consumption of these classes of surfactants, especially of alkyl sulphates, is the novel tendency and therefore their percentage will be increasing [48]. On the other hand, only LAS has been monitored in the aquatic environment. The major reason for this is the availability of established methods for the analysis of LAS and a complete lack of methods for the specific determination of the other classes of anionic surfactants. Because of this the development of methods for the determination of alkyl sulphates and alkylether sulphates in the aquatic environment is a great challenge for analysts. An even more difficult unresolved problem is the analysis of major metabolites of anionic surfactants in the aquatic environment. Sulphophenyl carboxylates being the main metabolite of LAS, are already, though rarely monitored. An alternative method for their determination is needed as well as more monitored sewage treatment plants and other components of the aquatic environment. The adaptation and introduction of monitoring methods for the determination of fatty alcohols (the major product of linear alkyl sulphates biodegradation and one of metabolites of biodegradation of alkylether sulphates) is required as well as methods for the determination of oxyethylated fatty alcohols having a very short oxyethylene chain (1-3 oxyethylene subunits). The latter are possible metabolites of alkylether sulphates. Both fatty alcohols and oxyethylated fatty alcohols having very short oxyethylene chains may also be metabolites of non-ionic oxyethylated fatty alcohols. Surfactants of anionic nature may be formed from non-ionic surfactants as a result of their biodegradation. The alkyl chain of non-ionic surfactants may undergo enzymatic co-oxidation: C H3-CH 2-(CH2),- CH 2-(O- C H 2-CH2)m-OH

5 HOOC-CH2-(CH2)n-CH2-(O-CH2-CH2)m-OH Thus non-ionic surfactants become anionic. Enzymatic co-oxidation of the oxyethylene chain also leads to the formation of compounds of anionic nature: CH3-CH2-(CH2)~-CH2-(O-CH2-CH2)(m_I)- O-CH2-CH2-OH

5 CH3-CH2-(CH2)n-CH2-(O-CH2-CH2)(m_I)-O-CH2-COOH These metabolites, having an anionic nature should behave similarly to anionic surfactants in separation schemes. They may interfere with the nonspecific determination of anionic surfactants. The method for their separation and

146 determination seems, to some extent, to be similar to those for anionic surfactants. The development of methods for the analysis of the metabolites is a serious analytical challenge. The total concentration of anionic surfactants as determined by the MBAS method may be overestimated due to the presence of 'pseudosurfactants' i.e. natural substances giving analytical signals in the MBAS method [114]. Humic substances, phenols and inorganic chlorides, nitrates and thiocyanates were identified as giving positive results in the MBAS analysis [14]. However, according to Osburn [95] many unidentified substances remain, particularly in river water and in the final effluents of sewage treatment plants. Although this opinion was expressed ten years ago, it still seems to be valid. The hypothesis concerning the existence of 'pseudosurfactants' was formulated on the basis of a comparison of the MBAS results and a specific determination of LAS results [33,114,123]. The LAS/MBAS ratio was found to be high for raw sewage (55 - 95% acc. to [33] or 71% acc. to [125]) and roughly reflected the value expected on the basis of the consumption of anionic surfactants. On the other hand the LAS/MBAS ratio in biologically treated sewage and in river water was much lower than expected. In the biologically treated sewage, the ratio was estimated at 18 - 53% [33] or 12 - 25% [125]. The most piecemeal data of the LAS/MBAS ratio was obtained for river water. 11-50% was estimated by Matthhijs and De Henau[33], 24-30%, by Waters and Garrigan [123] and an average 28% by Gerike et al [125]. Very detailed investigation of the River Rhine and its main tributaries in 1979 showed that the LAS/MBAS ratio was 43-70% in the upper part of the river and gradually decreased to 15% at the Dutch border [114]. An additional aspect in the discussion concerning 'pseudosurfactants' may be the fact that the MBAS procedure used together with the gas-stripping separation stage gave lower results than the simple MBAS version [126]. The lower the concentration level of anionic surfactants, the higher the difference between the results of the simple version of the MBAS and the MBAS combined with the gasstripping separation. This difference may partially correspond to the concentration of 'pseudosurfactants'. All these considerations concerning the 'pseudosurfactants' question did not take the role of alkylether sulphates and alkyl sulphates in the MBAS result into account. A more precise consideration of the problem should be based on a comparison of the total concentration of LAS, alkylether sulphates and alkyl sulphates with the MBAS result. Therefore the development of methods for the specific determination of alkylether sulphates and alkyl sulphates is such an important and challenging task for the analyst. Anionic metabolites of non-ionic surfactants may also be one of the components of 'pseudosurfactants'. 5.2. C a t i o n i c s u r f a c t a n t s Only 7% of the total production of surfactants is cationic surfactants [4]. Due to their textile softening, antistatic and bacteriostatic properties they are used broadly as fabric rinse conditioners. Quaternary ammonium salts are used

147 mainly though alkylpyridinium salts also play some role. Dihardenedtallowdimethylammonium chloride has been the major cationic surfactant [40]. Hardenedtallow is the name of a complex alkylchain which is represented by a Cls - alkylchain in 65% and by a C16 - alkyl chain in 30%. Undoubtedly, the main chemical species of dihardenedtallowdimethylammonium chloride is distearyldimethyl ammonium chloride. 5.2.1. N o n - s p e c i f i c m e t h o d s of d e t e r m i n a t i o n 5.2.1.1. The D B A S m e t h o d Nowadays, the Disulphine Blue Active Substances method is the main tool for the determination of cationic surfactants [19]. The results of this determination are known as the DBAS concentration. The principle of the method is based on the formation of a coloured ion-pair comprising a surfactant-to-be-determined and the anionic dye disulphine blue VN150 (also called Acid Blue 1')[17]. The ionpair is extracted with chloroform and the concentration is validated spectrophotometrically ()~ = 628 or 625 nm). In these terms the DBAS method is very similar to the MBSA method. However, contrary to the MBAS method, the DBAS requires a relatively sophisticated separation procedure prior to the determination itself [19]. The main stages consist in the evaporation of the analysed sample, extraction of ion-pairs from the residue with methanol and ionexchange removal of anionic surfactants. An excess of anionic surfactants in the initial sample is required. The method shows a recovery of 93-97% [19]. A slightly modified version of the method published by Osburn [20]. Kunkel [18] proposed the gas-stripping separation of cationic surfactants in the presence of an excess of anionic surfactants prior to the further stages of DBAS. Before disulphine blue has been introduced into the analytical routine, the other anionic dyes such as Orange II [127] or picric acid [128] were used. 5.2.1.2. O t h e r m e t h o d s Similarly as in the case of anionic surfactants, the two-phase titration and potentiometric titration have some potential for the analysis of environmental samples, especially those having a relatively high concentration of cationic surfactants such as sewage, processing liquors and solutions in biodegradation studies. The principles of these methods were described in paragraph '5.1.1.2. Titration methods'. The titrant usually used is sodium tetraphenylborate [64,69,82,129-133] although lauryl sulphate [134,135], tetradecyl sulphate [134] and picric acid [136,137] were also used. In the two-phase titration the same indicators as in the titration of anionic surfactants such as the mixed indicator consisting of dimidium bromide and disulphine blue VN [60,61] or tetrabromophenolophtalein ethyl ester [64] were recommended. In the latter case [64] the amount of cationic surfactants to of the order of 10 gg in the sample may be determined. However, the presence of anionic surfactants interferes with the determination. In the potentiometric titration, liquid membrane ion selective electrodes [82,129,131,133,135-137] were used. They frequently contained the

148 tetraphenylborate anion in the membrane [129,131,133]. Coated wire electrodes were also used for this purpose [69,130]. The physico-chemical limitation of the potentiometric titration is the relatively high solubility of the ion-par precipitates formed during titration.

5.2.2. Specific m e t h o d s of analysis 5.2.2.1. Specific determination of d i s t e a r y l d i m e t h y l a m m o n i u m cation Distearyldimethylammonium cation is the main objective of the specific determination of cationic surfactants, as it is the main manufactured surfactant of the cationic type. Previously thin-layer chromatography was used for this purpose [20,138-141], but more recently, HPLC determination combined with an adequate separation scheme predominates [142-145]. Osburn [20] developed the separation scheme using the final solution as used for the DBAS determination. The ion-pair of cationic surfactants with disulphine blue (previously used for the DBAS determination) is split on the cation exchange resin and cationic surfactants are eluted with hydrochloric methanol and separated and visualised on the silica gel G plate. A spot of distearyldimethylammonium cation is used for indication and semiquantitative determination. The HPLC procedure for the determination of distearyldimethylammonium cation basically uses the same separation scheme as the DBAS method [40]. The water or sewage sample is evaporated and the residue is washed with hydrochloric methanol solution. After methanol evaporation the residue is treated with LAS anion to form an ion-pair and is extracted with chloroform. This residue after chloroform evaporation is transferred into the anion exchange resin column to remove LAS and other anionic surfactants. Cationic surfactants contained in the methanol effluent of the column are put in the HPLC apparatus after evaporation and liquid/liquid back extraction from chloroform. The HPLC peak of distearyldimethylammonium cation may be recorded using the amino/cyano bounded silica column and detection by conductometry. The mixture of chloroform, methanol and glacial acetic acid is used as the mobile phase. A very similar scheme may be used for the determination of distearyldimethylammonium cation in environmental solid samples such as dried sludge [40]. The detection limit of the procedure was estimated at 2.5 gg 1-1 or 0.5 gg g-1 of solid matrix. Supercritical fluid extraction may be used as the first stage of the above described separation scheme for the analysis of dried sewage sludge [32]. It is worth mentioning research which demonstrates the possibility of analysing cationic and anionic surfactants simultaneously by utilising field desorption/collisionally activated dissociation mass spectrometry with simultaneous ion detection [146]. Only 5 pg of sample is required; however, only model samples were measured.

149

5.2.3. C h a r a c t e r i s t i c c o n c e n t r a t i o n s in the a q u a t i c e n v i r o n m e n t The available data is r a t h e r piecemeal and concerns the USA, UK, Germany and the Benelux countries [20,40,140,141,149]. The total concentration of cationic surfactants as determined by the DBAS method in raw sewage varied from 0.7 to 2.4 mg 1-~ [20,40,141] while in the final effluents of sewage t r e a t m e n t plants, from 0.03 to 0,27 mg 1-1. The DBAS concentration in river water was reported to vary from 4 gg 1-1 to 42 gg 1-1 [20,141,147,148] and the influence of the effluent of sewage t r e a t m e n t plant on the level of the DBAS concentration in river water was clearly shown [141]. Wasted dry activated sludge is reported to contain 0.4 0.5 gg g-~ [20] or 3.3 - 3.7 gg g-1 [40,141] while in river water sediments, from 0.02 to 0.14 gg g-1 [20,40,149]. Cationic surfactants are removed 72 to 98% in activated sludge t r e a t m e n t in sewage t r e a t m e n t plants [141]. Distearyldimethylammonium chloride consisted of 74% of the DBAS result in raw sewage [141], from 34% to 86% in final effluents [40,141], from 43% to 63% in river water, 60% in river sediments [40] and 81% in wasted sludge [141]. It is significant that the ratio of the MBAS (anionic surfactants) to DBAS was estimated as roughly equal 2.5 : 1 [149]. 5.2.4. U n r e s o l v e d q u e s t i o n s in the a n a l y s i s of c a t i o n i c s u r f a c t a n t s The following unresolved questions may be specified: i. the inconlusive results of m e a s u r e m e n t s of aquatic environment samples, ii. the development of an analytical tool for the investigation of interactions between cationic and anionic surfactants in the aquatic environment, iii. the automation of measurements. One of the main problems is to build up a more conclusive pattern of concentration levels in the components of the aquatic environment. The available results are not numerous and are very divergent. The reason may either be very divergent levels of concentration in the analysed samples or imperfections in the analytical methods. Generally, more data is needed. Interlaboratory tests would be useful in checking the analytical methods and the performance level of laboratories which analyse cationic surfactants. Precipitation of cationic surfactants by anionic surfactants is relatively well investigated under the model conditions characteristic for titration. However, the aquatic environment is much more complicated. It characterises both the excess of man-made anionic surfactants as well as numerous n a t u r a l substances of anionic nature, such as fulvic and humic acids or fatty acids. The precipitation of cationic surfactants under such complicated conditions may differ strongly from the model conditions. The analytical tools for the investigation of such kinds of complex systems are needed. A sophisticated separation scheme requires either the DBAS method or the specific determination of distearyldimethylammonium cation by the HPLC making the automation of m e a s u r e m e n t s impossible. The most troublesome stage

150 of these procedures in terms of the possibility of automation is the evaporation of the sample and the leaching of the residue with methanol. Alternative methods for the separation of cationic surfactants from water samples should be developed to make automation possible. 6. NON-IONIC S U R F A C T A N T S Non-ionic surfactants come second in terms of output after anionic surfactants. The yearly output of non-ionic surfactants is slightly above the half yearly output of anionic surfactants [4,5]. However, non-ionic surfactants display a greater variety of chemical structures and number of chemical compounds. Approximately one thousand individual chemical substances may be found in a mixture of non-ionic surfactant in the aquatic environment. The vast majority of non-ionic surfactants are ethoxylates i.e. products of the reaction of oxyethylene oxide with a reagent having a hydrophobic part and active hydrogen such as fatty alcohols, fatty acids, fatty amines, alkylphenols etc. An oxyethylene chain is formed and is the hydrophilic part of the molecule. The length of an oxyethylene chain may be controlled and in this way the ratio of the hydrophobic and hydrophilic part of the molecule may be tailored. A great variety of surfactants having different a hydrophobic-liophobic balance may also be easily tailored. On the other hand, the oxyethylation process leads to polydispersal products. Therefore, in reality, every commercial ethoxylate is a mixture consisting of several dozens of chemical substances. The analytical control of non-ionic surfactants is much more difficult than anionic or cationic surfactants [150-152]. Gas chromatography is usually applied for the determination of homologue distribution in commercial products. This distribution is one of the fundamental characteristics of non-ionic surfactants. Conventional methods are still used to determine the concentration non-ionic surfactants in commercial formulations such as extraction with chloroform, evaporation of organic solvent and gravimetric determination of dissolved nonionic surfactants [153]. Recently, titration of non-ionic surfactants by sodium tetraphenylborate in the presence of barium (II) ions with endpoint detection by ion-selective electrode is being introduced [154-159]. The methods for the determination of non-ionic surfactants in aquatic environment components (sewage, surface water) are usually very complex and complicated. 6.1. N o n s p e c i f i c m e t h o d s for t h e d e t e r m i n a t i o n of n o n - i o n i c s u r f a c t a n t s 6.1.1. The c h o i c e of s t a n d a r d error The general problem of nonspecific methods for the determination of non-ionic surfactants is the choice of standard. The background of this problem is the extremely high number of individual substances which may constitute the mixture of non-ionic surfactants. Approximately one thousand individual chemical substances may be present in the mixture, due mainly to the

151

r/l

~

/

=

/

concentration Figure 1A. Surfactant calibration graphs with large dispersion

concentration

Figure lB. Surfactant calibration graphs with narrow dispersion

polidispersity of ethoxylates. Even if the composition of this mixture were known, the reasonable choice of a standard would be a difficult task. Unfortunately, this composition is unknown and only speculations made on the basis of the manufacture of different classes of non-ionic surfactants may be done. Each component of the mixture characterises its own calibration graph. The bunch of calibration graphs of the mixture components may be less or more focused (see Figure 1A and 1B). The arbitrary choice of one standard substance gives rise to the error called the 'choice of standard error' [6]. If the calibration graph of the selected standard has a slope lower than the surfactant-to-be-determined (see Figure 2A), the value determined on the basis of the analytical signal and the calibration graph of standard is overestimated. The error of underestimation will be made if the calibration graph of the standard has a slope higher than the calibration graph of a surfactant-to-be-determined (see Figure 2B). In the case of the mixture, the best opportunity is the selection of a standard with the calibration graph in the middle of the bunch of calibration graphs of the mixture components, where errors compensate each other. However, such a choice requires that components of the mixture and their ratio be known. This is not yet a realistic task for the case of the mixture of non-ionic surfactants in the aquatic environment. However, the choice of standard error may be minimised by the use of the analytical method which characterises a more focused bunch of calibration graphs of particular non-ionic surfactants.

152

l

t

true

I

concentration

determined concentration

I

st

on

~0

~

o~.,~

r op,,~

o~..~

e~

concentration Figure 2A. The standard choice error: overestimation un- unknown surfactant-tobe-determined

concentration Figure 2B. As 2A: underestimation error st - chosen arbitrary standard

6.1.2. The BiAS m e t h o d The Bismuth Active Substances method (BIAS) is recommended in Europe for the determination of non-ionic surfactants in aquatic environment components [27,41,148]. This method is also known as the Wickbold method or the method with modified Dragendorff reagent. A selective step of the BIAS method is the precipitation of the orange coloured compound of ethoxylates with Dragendorff reagent in the presence of barium(II) ions. A pseudocation is formed between ethoxylates and barium(II) ion. This pseudocation reacts with the anionic complex of bismuth(III) and iodides (called Dragendorff reagent). Only ethoxylates having more than 4 oxyethylene subunits form the precipitate. The first semiquantitative method using this reaction was developed by B~irger [161]. The concentration of non-ionic surfactants was evaluated on the basis of the height of the precipitate layer in special tubes with capillaries in the lowest part. The present version of the BiAS method was developed by Wickbold [27]. A gasstripping separation of non-ionic surfactants from a water matrix to an ethyl acetate layer was introduced prior to the determination itself, both for an effective preconcentration as well as for the separation of non-ionic surfactants from impurities in the water matrix. Evaporation of the organic solvent was followed by the precipitation of ethoxylates with modified Dragendorff reagent. The concentration of non-ionic surfactants is indirectly determined by the determination of bismuth(III) in the precipitate. Basically, the gas-stripping technique in this particular case is a special extraction technique adopted for extraction of large volumes of samples. However, the partition coefficient of

153 surfactant still 'governs' the water/ethyl acetate system as in the case of the liquid/liquid extraction. The water sample is processed in a special column with a frit bottom. Prior to processing the sample should be filtered. Sodium chloride and sodium hydrogencarbonate are added to the sample to adjust ionic power and pH and a nitrogen stream is passed through it during the separation. Gasstripping separation is repeated twice [27] or four times [41] in order to achieve a satisfactory degree of separation. The time-consuming stage of the procedure is the evaporation of ethyl acetate i.e. 400 - 800 ml. The residue is dissolved in methanol and treated with barium(II) and Dragendorff reagent. Bismuth(III) is determined in the precipitate by potentiometric titration [27] or spectrophotometrically by measurement of the absorbance of bismuth(III)-EDTA complex in UV range [41]. A non-filtered water sample should be treated according to Waters, and an additional step of cleaning the sample in order to remove ionic surfactants is necessary. The difference in Wickbold's approach (a filtered water sample) and Waters (a non-filtered sample) comes down to the amount of non-ionic surfactants adsorbed on particles of surface water [8]. The BiAS method is considered as a method for the determination of the total concentration of non-ionic surfactants. In reality only ethoxylates having 5 - 30 oxyethylene subunits are determined by the method. Non-ionic surfactants of non-ethoxylate structure (alkylamides, alkylpolyglucosides, etc.) are not determined by this method. Moreover, ethoxylates having 1-4 oxyethylene subunits are not precipitated by the modified Dragendorff reagent. Ethoxylates having more than 30 oxyethylene subunits are lost during gas-stripping separation. The BiAS procedure is recommended for the analysis of samples containing 100 to 1000 ~g. This range corresponds to the amount of non-ionic surfactants in raw and treated sewage as well as in samples used in biodegradation tests. If there are less than 100 ~g of non-ionic surfactants in the sample the BiAS method does not work properly. Interlaboratory experiments with the participation of prominent laboratories in Germany, UK and Belgium showed satisfactory results for samples of raw sewage, but totally unsatisfactory results for biologically treated sewage [162]. For example, the results for one of these samples were scattered over the range of 70 - 390 ~g 1-1. Additional problems concern the determination of non-ionic surfactants in river water by the BiAS procedure. In relatively non-polluted river water the concentration of non-ionic surfactants is below 100 ~g 11. Therefore several litres of water must by processed by gas-stripping in order to obtain a sample containing more t h a n 100 ~g 1-1. This large volume of water sample generates further problems. Filtration of such a sample, especially if it contains high concentrations of particles is a cumbersome task. 800 ml of ethyl acetate must be used during gas-stripping separation which then needs to be evaporated. The main reason for the poor detection level, and unsatisfactory results of interlaboratory experiments was the loss of the surfactant-to-be-determined during washing of the precipitate containing surfactants with glacial acetic acid [163]. This stage of the classical BiAS procedure is necessary in order to

154 remove excessive bismuth(III) from the precipitate and the filter, though simultaneously a part of the precipitate is lost. Because the washing is performed on the filter, the loss is kinetic dependent and the duration of washing may be an important factor. This may well be the main reason for the poor interlaboratory tests results. A serious problem in the BiAS procedure is calibration. The total concentration of non-ionic surfactants should be calculated using a single standard and an error connected with this selection (the standard choice error) was discussed elsewhere. 6.1.3. T h e CTAS method The Cobalt Thiocyanate Active Substances method (CTAS) is recommended in the USA [14] for the determination of non-ionic surfactants in environmental samples. CTAS is considered as equivalent to the BiAS method and is an extraction-spectrophotometric method. Ethoxylates having more than 5 oxyethylene subunits form coloured compounds with an anionic complex of cobalt(II) and thiocyanates. This coloured compound is extracted to dichloromethane and spectrophotometric measurement is performed. To eliminate interferences, a relatively complicated separation procedure is recommended [23]. This consists of adsorption of non-ionic surfactants from water or sewage samples on the macroreticular resin Amberlite XAD-2, a removal of fatty material with petroleum ether, elution of adsorbed non-ionic surfactants by an ethyl ether/methanol mixture, evaporation of solvents, extraction of non-ionic surfactants to ethyl acetate, evaporation of solvent again and treatment of the residue dissolved in methanol on an ion exchange resin column with the mixed bed. After methanol evaporation, the residue is processed according to the CTAS extraction-spectrophotometric procedure, briefly described above. More recently, the gas-stripping separation procedure is recommended, performed just as in the BiAS method, instead of the sophisticated separation scheme described above [14]. The 'spectrum' of ethoxylates determined by the CTAS method is roughly the same as that of the BiAS method i.e. those having 5 - 30 oxyethylene subunits. The detection limit of the CTAS is about 50 ~g 1-1 [36].

6.1.4. The GC-hydrogen bromide cleavage method This is based on the cleavage of the ethoxylene chain with hydrogen bromide and gas-chromatographic determination of dibromoethane, the product of the cleavage [164-168]. The final version of the method, developed by Wee [169] and recommended by Matthijs and Hennes [36] is preceded by a relatively complicated separation scheme. Surfactants are separated from a water matrix by gas-stripping to ethyl acetate or by liquid-liquid extraction to chloroform. After evaporation of the organic solvent the residue is dissolved in methanol and passed through an ion-exchange resin column. The eluent is evaporated, the residue dissolved in chloroform and passed through a silica gel column. The eluent is washed with methanol, after evaporation, and is treated with hydrogen

155 bromide and the dibromoethane formed is determined by GC. The detection limit of the method is evaluated at 10 ~g 11 [36].

6.1.5. Other m e t h o d s An extraction-spectrophotometric method similar to the CTAS has been developed by Favretto et al [170]. Pseudocation is formed between ethoxylates and potassium(I) cation which forms the ion-pair with picrate anion. The ion-pair is extracted to dichloromethane and its absorbance is measured by spectrophotometry at k = 378 nm. Another alternative is the precipitation of ethoxylates with phosphorous-molibdic acid in the presence of barium(II). Pseudocation of ethoxylates and barium (II) forms a precipitate with phosphorous-molibdic anion. The non-precipitated excess of molybdenum is determined by atomic absorption spectrometry [171,172]. Gas-stripping separation is used for initial separation and preconcentration. A similar method has been developed Winkler and Buhl [173,174], where removal of albumenic substances from the water matrix was applied instead of gas-stripping and three alternative techniques for the determination of molybdenum: X-ray fluorimetry, atomic absorption spectrometry and polarography, were used. All three gave similar results. 6.1.6. T e n s a m m e t r i c t e c h n i q u e s

6.1.6.1. The adsorptive stripping t e n s a m m e t r y The role of tensammetric techniques in the determination of non-ionic surfactants in the aquatic environment is growing. The most frequently used tensammetric technique is alternating current voltammetry. However, the first method developed which took into account the complexity of aquatic environment samples, was the Kalousek Commutator technique [175]. Earlier works did not consider the problem of matrix effects and therefore they may be classified as very interesting heuristic models. A very important factor in the development of the recent version of t e n s a m m e t r y has been the introduction of the hanging mercury drop electrode (HMDE). Adsorptive preconcentration of surfactants on the surface of HMDE has enabled the determination of concentrations of surfactants to two orders of magnitude lower than previously [176-181]. This technique is called the adsorptive stripping t e n s a m m e t r y (AdST). The additional advantage of the use of HMDE is the ability to differentiate the adsorption of different substances by a change in the preconcentration potential. The main disadvantage of the AdST is the complex behaviour of mixtures of surfactants [182-184]. This factor makes direct application of the AdST to aquatic environment samples very difficult [182-184]. However, the AdST may be very useful in those systems where a single surfactant predominates. The sources of error of the classical BiAS procedure were clearly shown by the use of the AdST and Triton X-100 as the model surfactant [163]. In this way the further successful modification of this procedure was achieved. Strong adsorption of nonionic surfactants on different materials was shown by the use of the AdST as a

156 tool [185,186]. The AdST has been shown to be a useful auxiliary tool in biodegradation studies [187]. 6.1.6.2. The i n d i r e c t t e n s a m m e t r i c t e c h n i q u e and i n d i r e c t t e n s a m m e t r i c method

The indirect tensammetric technique (ITT) is the technique specified for the determination of non-ionic surfactants in the aquatic environment samples and is therefore much better adapted to this task than the majority of other tensammetric techniques: The lowering of the tensammetric peak of a monitoring substance due to competitive adsorption of surfactants-to-be-determined is the analytical signal in the ITT (see Fig. 3). The monitoring substance must be a weaker surfactant than the surfactants-to-be-determined. Usually ethyl acetate is used as the monitor [188]; it is a weaker surfactant than non-ionic surfactants, but stronger than anionic. Therefore, anionic surfactants do not interfere with the determination of non-ionic surfactants by the ITT, with ethyl acetate as the monitor [189]. Another important advantage of the ITT is the approximate additivity of the analytical signals of particular mixture components [190]. The ITT combined with a simple extraction procedure is called the indirect tensammetric method (ITM). The only separation stages are filtration and extraction of non-ionic surfactants to ethyl acetate [188]. An aliquot of ethyl acetate phase may be introduced into the measuring cell, dissolved in the base electrolyte and the tensammetric measurement performed. In the case of a lower concentration of non-ionic surfactants an all ethyl acetate phase may be

- --analytical signa7 - - I

il potential Figure 3A. Tensammetric peak of the monitoring substance alone.

'

-

potential Figure 3B. Tensammetric peak of the monitoring substance in the presence of the surfactant-to-be-determined.

157 evaporated and the residue dissolved in 1.5 ml of ethyl acetate, the optimal a m o u n t needed as the monitor for performance of the ITT m e a s u r e m e n t . The ITT m a y be used for the d e t e r m i n a t i o n of non-ionic surfactants in sewage, effluents of sewage t r e a t m e n t plants and surface w a t e r [191]. The detection limit of the ITM was determined at 1.5 pg in the sample i.e. almost two orders of magnitude better t h a n in the classical BiAS method [191]. Ethoxylates having 1 - 30 oxyethylene subunits m a y be determined by the ITM i.e. the broader 'spectrum' t h a n in the BiAS or CTAS, which are only able to determine ethoxylates having 5 - 30 oxyethylene subunits. However, the other substances extracted by ethyl acetate are d e t e r m i n e d using the ITM e.g. free fatty alcohols [188]. Fortunately, no substantial differences were evidenced when river w a t e r or sewage samples were determined simultaneously by the ITM and classical BiAS [191]. Due to a much better detection limit t h a n the classical BiAS or CTAS, the ITM requires a substantially lower volume of w a t e r sample to be processed. Usually 100 - 200 ml is extracted by 50 ml of ethyl acetate used in two portions. There is no need for time-consuming gas-stripping with large volumes of w a t e r and ethyl acetate. This advantage is especially seen in the case of samples with low concentrations of non-ionic surfactants. An additional advantage of the ITM in comparison to the classical BiAS or CTAS is a considerably narrower dispersion of slopes of calibration graphs of different non-ionic surfactants. Therefore the ITM is substantially less sensitive to the error of choice of s t a n d a r d (see section 6.1.1.) and d e m o n s t r a t e s an a d v a n t a g e in the case of the d e t e r m i n a t i o n of a non-ionic surfactant mixture of u n k n o w n composition [6]. The disadvantage of the ITM is interference of chlorophyll extracted from water plants [6]. This drawback may be overcome by the filtration of the sample prior to extraction. Monitoring of nonionic surfactants in the River W a r t a in Poznan was established in 1990 on the basis of the ITM m e a s u r e m e n t s [120]. Seven years experience in the use of the ITM for this purpose d e m o n s t r a t e s t h a t this method m a y be recommended for the routine control of non-ionic surfactants in river w a t e r [191].

6.1.6.3. The m e t h o d combining BiAS separation procedure with the indirect t e n s a m m e t r i c technique (BiAS-ITT) The BiAS-ITT is more selective t h a n the ITM due to the additional step of selective precipitation of ethoxylates with modified Dragendorff reagent [192,193]. Chlorophyll interference can be removed this way. Non-ionic surfactants are determined in the final stage of the BiAS-ITT method instead of bismuth(III) as in the classical version of the BIAS. Washing with glacial acetic acid is omitted in the BiAS-ITT, compared with the classical version, because it is at this stage t h a t serious loss occurs. The washing m a y be omitted because nonionic surfactants are determined instead of bismuth(III) and its excess in the precipitate does not hinder the determination. The detection limit of the BIASITT was determined as equivalent to 1.5 ~g in the sample i.e. equivalent to the ITM, though the BiAS-ITT is more complicated and required more labour. Only ethoxylates having 4 - 30 oxyethylene subunits can be determined by the method

158 i.e. similar to the classical BIAS. The BiAS-ITT appeared to be the method capable of competing with the ITM. Several problems unresolvable by the other methods, including the ITM, were resolved using the BiAS-ITT. Two methods for the determination of non-ionic surfactants adsorbed on particles of river water were developed [194]. A total flux of non-ionic surfactants in river water can be determined this way as well as its components: soluble non-ionic surfactants and those adsorbed on particles. The method for the determination of non-ionic surfactants adsorbed on activated sludge in a dynamic biodegradation test according to the OECD Confirmatory Test was developed [195]. A slightly modified version of the BiAS-ITT was developed for the determination of nonionic surfactants in the presence of hydrocarbons [196]. Hydrocarbons adsorbed on the precipitate formed by ethoxylates with modified Dragendorff reagent can be removed by washing with iso-octane. This step does not cause a loss of surfactant. On the other hand, unremoved hydrocarbons extract ethoxylates causing a negative error. The BiAS-ITT can be successfully applied to the determination of primary biodegradation in the OECD Confirmatory Test [197]. The ITM causes very serious errors under these conditions due to the presence of free fatty alcohol being the metabolite of biodegradation.

6.2. Specific m e t h o d s for the determination of non-ionic surfactants In contrast to the general meaning of this term, where a 'specific signal' usually means the signal of a single chemical species, in surfactant analysis, because of the complexity of subject the specific determination may mean the determination supplying more information than the 'total concentration' e.g. a concentration of specific group of surfactants. 6.2.1. Oxyethylated alkylphenols Most attention has been paid to the determination of this class of non-ionic surfactants because of their poor biodegradability. Additionally, due to the presence of the phenyl ring in the molecules, it was an easier task t h a n in the case of other classes of non-ionic surfactants. The presence of the phenyl ring facilitated the use of UV-spectrophotometric or fluorimetric detection. Only oxyethylated alkylphenols among non-ionic surfactants posses this property. Ahel and Giger [198] developed an HPLC method with UV detection for the specific determination of oxyethylated alkylphenols. Gas-stripping separation of non-ionic surfactants according to Wickbold [27] was the first stage of separation and an aluminium oxide column for additional cleaning of the sample was used. The method was used for the determination of oxyethylated alkylphenols in different environmental samples [199,200]. Application of fluorimetric instead of UVspectrophotometric detection substantially improved the detection limit of the method [10,201-203]. Kubeck and Naylor [10] applied another separation scheme to the HPLC determination of oxyethylated alkylphenols. The sample was cleaned on the ion-exchange column with the mixed bed, and in the solid phase extraction column with a Cls reverse phase. Hot methanol was used to wash the

159 determined oxyethylated alkylphenols from the column, prior to the separation on the HPLC column.

6.2.2. O x y e t h y l a t e d a l c o h o l s The success of the specific determination of oxyethylated alkylphenols stimulated attempts to develop a similar specific method for the determination of oxyethylated alcohols, which are the main group of non-ionic surfactants in terms of output [5,47]. In order to solve the detection problem, Allen and Linder [204] derivatized oxyethylated alcohols with phenylisocyanate, converting them to the form detectable in UV-spectrophotometry. The preliminary separation scheme was partially adapted from the previous schemes concerning oxyethylated alkylphenols. The sequential application of gas-stripping, ion exchange resin and aluminium oxide columns comprises one of the versions of the separation scheme. Another scheme developed by Schmitt et al [23] proposes a sequential application of the column with macroreticular adsorbent Ameberlite XAD-2, liquid-liquid extraction with ethyl acetate and ion-exchange treatment prior to the derivatization with phenylisocyanate. HPLC separation with the normal phase (~Bond-apakNHD leads to a separation of the treated mixture in accord with the length of the oxyethylene chain while separatien on the reverse phase column (~Bond-apak Cas), accords with the length of alkyl chain [23]. The method applied for calculation of the results requires a presumption of the length of the oxyethylene chain. The results obtained by using the HPLC method with derivatization are much lower than those for the same samples obtained by the CTAS [124] or the classical BiAS [162] methods. On the basis of these results and taking into account that oxyethylated alcohols are the main class of non-ionic surfactants in terms of output, the opinion was expressed that the CTAS or BiAS methods lead to a serious overestimation of non-ionic surfactants [124,162]. Other authors suggest that the considered HPLC method leads to considerably lower results for complex environmental samples [206]. It is worth mentioning that the results of the CTAS [205] and classical BIAS [191] methods yield similar results to ITM and BiAS-ITT. Hence, the HPLC method with derivatization using phenylisocyanate in its current version seems to be erroneous. 6.2.3. Specific d e t e r m i n a t i o n of non-ionic s u r f a c t a n t s h a v i n g m o r e than 30 o x y e t h y l e n e s u b u n i t s and o x y e t h y l a t e d a l c o h o l s h a v i n g the C16-1s alkyl c h a i n with m u l t i - m o n i t o r i n d i r e c t t e n s a m m e t r i c technique Recently, two tensammetric methods for the specific determination of specific groups of non-ionic surfactants have been developed using the so called multimonitor indirect tensammetric technique[25,26,207]. There is the method for the determination of non-ionic surfactants having more than 30 oxyethylene subunits [25]. This group was not determined either by CTAS or by the BiAS procedures. The ITM and BiAS-ITT methods were also unable to determine this group of nonionic surfactants. Therefore these surfactants, including the important barely

160 biodegradable oxyethylene-oxypropylene block copolymers, were thoroughly uncontrolled. The developed method combines an adequate separation scheme with the newly developed tensammetric technique which improves the specific detection of surfactants. Non-ionic surfactants having less than 30 oxyethylene subunits, which hinder the determination, are first separated from the water or sewage sample by extraction w{th ethyl acetate. The sequential extraction of the same sample with chloroform gives the fraction containing poly(ethylene glycols) (PEG) and non-ionic surfactants having more than 30 oxyethylene subunits. To remove the influence of anionic surfactants, which also hinder the determination, ethoxylates and PEG of the chloroform fraction are precipitated with the modified Dragendorff reagent. The 'total concentration' of surfactants of the 'chloroform fraction' is determined using the ITT with an ethyl acetate monitor, while the concentration of non-ionic surfactants having more than 30 oxyethylene subunits is determined using the second monitor tetrabutylammonium bromide (TBAB) [25]. The adsorptive ability of TBAB is weaker than non-ionic surfactants having more than 30 oxyethylene subunits but stronger than PEG. Therefore, only nonionic surfactants having more than 30 oxyethylene subunits can replace the TBAB monitor molecules on the electrode surface, which is apparent from the lowering of the TBAB tensammetric peak. The multi-monitor indirect tensammetric technique is also useful in the specific determination of oxyethylated alcohols having the C16-18 alkyl chain [26,207]. No special separation scheme is required. The same separation scheme is used as in the BiAS-ITT. The final stage of the BiAS-ITT is the determination of the total concentration of non-ionic surfactants, using the first monitor, ethyl acetate. Simply the introduction of the second monitor, TBAB and performance of the next tensammetric measurement creates the signal useful for the determination of oxyethylated alcohols having the C16-18 alkyl chain. These surfactants exhibit the strongest adsorptive ability among non-ionic surfactants separated by extraction with ethyl acetate. Oxyethylated alcohols having the C16-18 alkyl chain are a stronger surfactant than TBAB, while the other oxyethylated alcohols, oxyethylated alkylphenols, oxyethylated amines, etc. are weaker. Therefore oxyethylated alcohols having the C16-18 alkyl chain replace TBAB monitor molecules on the electrode surface, creating a decrease of the TBAB tensammetric peak, while the other non-ionic surfactants in the ethyl acetate fraction, do not. 6.2.4. M e t a b o l i t e s Extensive model studies were performed for every class of surfactants to investigate the degree of biodegradation and the optional pathways of this process [150]. However, metabolites are rarely monitored in the aquatic environment components. The reason may be the more complex composition of the aquatic environment samples than the samples from the biodegradation tests.

161

6.2.4.1. Alkyphenol, short chain alkylphenol polyethoxylates and their carboxylated biotransformation products Free alkylphenol and short chain alkylphenol polyethoxylates having 1 or 2 oxyethylene subunits are considered as the main metabolites of the biodegradation of oxyethylated alkylphenols. They are formed as a result of the enzymatic hydrolysis of the oxyethylene chain pathway [150] according to reaction:. enzyme CH3-CH2-(CH2)n-CH2--CGH4-O-CH2-CH2-(O-CH2-CH2)m-(O-CH2-CH2)-OH CH3-CH2-(CH2)n-CH2--C6H4-O-CH2-CH2-(O-CH2-CH2)m-OH + HO-CH2-CH2-OH The gradual shortening of the oxyethylene chain finally leads to the formation of alkylphenol monoethoxylate and alkylphenol diethoxylate. At this stage the biodegradation is slowed, causing the accumulation of these compounds. Further biotransformation leads to the formation of carboxylated compounds such as alkyphenol acetic acid and alkylphenolethoxy acetic acid [208,209]. Ahel and Giger developed the HPLC method for the determination of free alkylphenol and alkylphenol monoethoxylate and alkylphenol diethoxylate with an adequate separation scheme [31]. An exhaustive steam distilation/solvent extraction procedure was used to preconcentrate and separate the determined compounds. Quantitative determination was performed by the normal phase aminosilica column, using UV-spectrophotometric detection (k = 277 nm). The detection limit was evaluated as equivalent to 0.5 ~g 1-1. The further improvement of the method [208] was connected with the application of spectrofluorimetric detection (excitation at ~.= 277 nm and emission at ~.= 300 nm), which resulted in a better detection limit (10 ng l-l). The method was applied to aquatic environment samples [2,199,200] and is widely accepted. Alkylphenoxy carboxylic acids were separated by extraction with chloroform from acidified water samples [208]. The metabolites contained in the residue after evaporation of solvent were derivatized in the form of corresponding methyl esthers. HPLC determination with a spectrofluorimetric detection was applied (excitation at ~.= 277 nm, emission at ~.= 300 nm) and the detection limit was estimated as eqivalent to 100 ng 1-1. The solid phase extraction scheme for nonylphenoxycarboxylic acids was also proposed [209]. 6.2.4.2. P o l y ( e t h y l e n e glycols) Poly(ethylene glycols) (PEG) are the main products of the biodegradation of ethoxylates following the central fission pathway [150]. The reaction of oxyethylated alcohol may be an example of this pathway:

162 enzyme

5 CH3-CH2-(CH2)n-CH2- O-CH2-CH2-(O-CH2-CH2)m-OH

5 CH~-CH2-(CHDn-CH2-OH + HO-CH2-CH2-(O-CH2-CHDm-OH The other product of the biodegradation of ethoxylates, according to this pathway, is fatty alcohol. Oxyethylated alcohols having n-alkyl moiety frequently undergo this pathway. It is worth mentioning that this class predominates in the output of non-ionic surfactants [5,47]. Recently, two tensammetric methods for the determination of PEG have been developed [24,210]. The simplest applied is the modified ITM [24]. The first stage of the procedure is the extraction of nonionic surfactants, which hinder the determination. The sequential extraction to chloroform causes the separation of PEG from the water matrix. PEG are determined using the ITT. However, ethoxylates having more t h a n 30 oxyethylene subunits are also determined this way [211 ]. The other tensammetric method for the determination of PEG is the modified BiAS-ITT [210-212]. The separation of PEG by sequential extraction with ethyl acetate and then chloroform is identical to the method described above. In the next stage, PEG and ethoxylates having more than 30 oxyethylene subunits are precipitated with the modified Dragendorff reagent. Anionic surfactants are in this way effectively separated. The total concentration of PEG and ethoxylates having more than 30 oxyethylene subunits can be determined in the dissolved precipitate by the ITT using ethyl acetate as the monitor [210-212]. To distinguish PEG and ethoxylates having more than 30 oxyethylene subunits the second monitor, TBAB is applied [26,212]. In this way the concentration of ethoxylates having more than 30 oxyethylene subunits can be determined. The concentration of PEG is calculated by the difference. 6.3. C h a r a c t e r i s t i c c o n c e n t r a t i o n s in t h e a q u a t i c e n v i r o n m e n t Concentrations of non-ionic surfactants obtained using the BIAS, CTAS and ITM methods can be roughly considered as equivalent to each other. However, it is necessary to stress that each of these methods yield results for surfactants having less than 30 oxyethylene units. Surfactants having more than 30 oxyethylene units are uncontrolled. The results available in the literature concerning the concentration of non-ionic surfactants in raw sewage, biologically treated waste water, river and lake water are given in Table 1. The differences in the levels of non-ionic surfactant concentrations in raw sewage were quite understandable due to the divergent levels of consumption of non-ionic surfactants and water, as well as population density. This concentration varied within the range of 1 to 10 mg 1-1. Basically biologically treated waste water contains from 0.1 to 0.5 mg 1-1 of non-ionic surfactants. In the majority of cases

163 Table 1 Characteristic concentrations of non-ionic surfactants (as d e t e r m i n e d by the BiAS, CTAS, ITM or similar methods) in the aquatic environment samples ( m g l 1) ill

Sample

i

i

Country

Concentration

Reference

USA Germany

1.62 3.1 - 8.5 0.2-1.2 0.9-2.3 1.0-3.9 5.0- 5.5

[124] [162] [149] [141] [141] [25,26]

0.13 -0.18 0.19 - 0.23 0.21-0.24 0.09-0.17 0.2- 0.5 0.33 - 0.38 0.55

[124] [162] [149] [141] [141] [25,26]

i

Raw sewage

UK Poland Treated w a s t e w a t e r USA Germany

UK Poland

[191]

River w a t e r Germany

0.02 - 0.06

[114]

0.02-0.25

[149]

Poland-Silesia Israel

0.02 - 0.06 0.02-0.20 0.15- 1.3 7.8

[120] [191] [173,1741 [213]

Poland

0.02 - 0.045

[214]

Poland

Lake w a t e r

the concentration of non-ionic surfactants in river w a t e r varied between 20 and 250 ~g 11. The lower value seems r a t h e r to be the result of the detection limit of the BiAS and CTAS methods t h a n the real lowest concentration. In some cases [173,174,213] the reported concentrations were more characteristic of raw sewage t h a n river w a t e r due to extreme pollution. In the only reported case, lake water had a concentration of non-ionic surfactants similar to the lower range of river water concentrations [214]. M e a s u r e m e n t s with the application of the multimonitor BiAS-ITT supplied additional information concerning the aquatic e n v i r o n m e n t which was not

164

available using the other methods. These results are shown in Table 2. The results concerning the concentration of non-ionic surfactants having more t h a n 30 oxyethylene subunits show that an additional 16% of non-ionic surfactants is present in raw sewage, 3 - 18%, in biologically treated sewage and 21%, in river water. The representativeness of these values is limited because they were obtained from a limited number of sewage treatment plants (2) and river water samples (4) and they therefore represent the potential of this approach rather, than a typical ratio of different fractions of non-ionic surfactants in the aquatic environment. Table 2 also shows the concentration of oxyethylated alcohols having C16-18 alkyl. It is apparent that this group of non-ionic surfactants is the major component of the mixture of non-ionic surfactants and constitutes 55 - 65% of non-ionic surfactants in raw sewage, 40 - 55% in biologically treated sewage and 43 - 48% in river water. Obviously, these proportions were characteristic of only a part of Poland as well as the corresponding period of experiments (1996). Poly(ethylene glycols), as the metabolites of the biodegradation of oxyethylated alcohols according to the central fission pathway, were present in raw sewage at a negligible concentration (see Table 2), while their concentration in biologically treated sewage was in the same range as the residual concentration of non-ionic surfactants. This increase in concentration of PEG strongly supports the central fission pathway. PEG in the River Warta constituted approximately 10% of nonionic surfactants over the period of measurement (1996). The concentrations of the metabolites of oxyethylated alkylphenols are shown in Table 3. The relatively low concentration of alkylphenol and short chain alkylphenol ethoxylates in raw sewage is apparent but with relatively high concentrations of alkylphenol mono- and di-oxyethylates in the biologically treated sewage. This confirms the enzymatic hydrolysis of the oxyethylene chain biodegradation pathway in the case of oxyethylated alkylphenols. A relatively high concentration of free alkylphenol and alkylphenol mono- and di-oxyethylates in the river water, and free alkylphenol in the ground water demonstrate that focusing attention on these compounds in the aquatic environment is necessary. Alkylphenol acetic acid and alkylphenoxy acetic acid are products of a deeper biotransformation of oxyethylated alkylphenols. Their relatively high concentrations in river and ground waters support the thesis concerning the difficult biodegradation of oxyethylated alkylphenols having short-chains, and their accumulation in the aquatic environment.

165

Table 2 Concentration of non-ionic surfactants having more than 30 oxyethylene subunits (EO), oxyethylated alcohols having C16-18 alkyl and poly(ethylene glycols) in the aquatic environment samples (mg 1-1 ) Sample/fraction/metabolite Raw sewage non-ionic surfactants having 4 - 30 EO (BiAS-ITT 1-monitor) oxyethylated alcohols having C16-18 alkyl (BiAS-ITT 2-monitors)

(%) non-ionic surfactants having > 30 EO (BiAS-ITT 2-monitors) real total concentration of non-ionic surfactants (4- 30 EO + >30EO) poly(ethylene glycols) Treated sewage* non-ionic surfactants having 4 - 30 EO (BiAS-ITT 1-monitor) oxyethylated alcohols having C16-~8 alkyl (BiAS-ITT 2-monitors)

(%) non-ionic surfactants having > 30 EO (BiAS-ITT 2-monitors) real total concentration of non-ionic surfactants (4- 30 EO + >30EO) poly(ethylene glycols) River water* non-ionic surfactants having 4 - 30 EO (BiAS-ITT 1-monitor) oxyethylated alcohols having C16-18 alkyl (BiAS-ITT 2-monitors)

(%)

*

Concentration

Reference [25,26,212]

5.0- 5.1 2.9 - 3.2 55 -65 0.81 - 0.82 5.8-6.0 0.04 - 0.13 [25,26,212] 0.33 -0.38 0.15 -0.18 4 0 - 55 0.01 - 0.06 0.39 0.29 - 0.35 [25,26,212] 0.064 - 0.082 0.028 - 0.036 43 -48

non-ionic surfactants having > 30 EO 0.013 - 0.018 (BiAS-ITT 2-monitors) real total concentration of non-ionic surfactants 0.073 - 0.100 (4 - 30 EO + >30EO) poly(ethylene glycols) 0.006 - 0.011 sewage treatment plants at 'Mosina' and 'Pleszew', ** - the River Warta, Poznan -

166 Table 3 Concentrations of metabolites of oxyethylated alkylphenols in the aquatic environment samples (lag 11 ) or (lag g-1 ,) Metabolite/sample Alkylphenol raw sewage treated wastewater river water** ground water activated sludge Alkylphenol monooxylate raw sewage treated wastewater river water** ground water activated sludge Alkylphenol dioxylate raw sewage treated wastewater river water** ground water activated sludge Alkylphenol acetic acid river water** ground water Alkylphenoxyethoxy acetic acid river water** ground water * - dry weight ** - the River Glatt, Switzerland

Concentration

Reference

14 6-14 <0.5 - 2 0.7 - 26 0.1 -33 12.8

[31] [31] [31] [208] [208] [31]

18 29-63 0.5- 15 2-20 0.1- 1.7 76

[31] [31] [31] [208] [208] [31]

18 42- 72 0.5- 14 0.8- 21 <0.1 61

[31] [31] [31]

[208] [208] [31]

8.4 - 2 0 . 1 2.8- 3.0

[208] [208]

20.6- 28.7 12.7- 16.3

[208] [208]

167

6.4. U n r e s o l v e d q u e s t i o n s of a n a l y s i s of n o n - i o n i c s u r f a c t a n t s The list of unresolved questions in the analysis of non-ionic surfactants is much longer than in the cases of the other types of surfactants. The following major unresolved questions may be specified: i. the selection of the method for the determination of the total concentration of non-ionic surfactants capable of approval in interlaboratory tests, ii. the selection of a standard surfactant representative for the mixture of non-ionic surfactants in the aquatic environment, iii. the determination of non-ionic surfactants having less than 5 oxyethylene subunits, iv. the determination of non-ionic surfactants having more than 30 oxyethylene subunits, v. the specific determination of other classes of ethoxylates (oxyethylated amines, oxyethylated fatty acids, etc.) vi. the specific determination of newly introduced non-ethoxylate non-ionic surfactants (e.g. alkyl polyglucosides), vii. the trace analysis of non-ionic surfactants, viii. the development of a method suitable for the control of biodegradation at a realistic level of concentration, ix. methods for the analysis of metabolites of non-ionic surfactants, x. detectors for HPLC and FIA measurements, xi. automation of analysis of non-ionic surfactants. Though the BiAS and CTAS methods are considered as established and equivalent to each other, they operate optimally only at a range of concentrations higher than several hundreds pg in the sample. Attempts to perform interlaboratory tests with several biologically treated samples showed results too piecemeal to be accepted [162]. The sources of errors of the classical BiAS method were found [163]. Therefore the method capable of passing the interlaboratory tests at the level of 20 - 200 ~g in the sample is yet to be found. The choice of standard is still arbitrary due to the complete lack of knowledge concerning the composition of the mixture of non-ionic surfactants in real samples of the aquatic environment. The other drawback of the BiAS and CTAS methods is their analytical response to ethoxylates having only 5 - 30 oxyethylene subunits. The control of concentration of ethoxylates fraction having 1 - 4 oxyethylene subunits has become more important recently due to the tendency of using oxyethylated alcohols which have shorter oxyethylene chains [48]. The other factor enhancing the importance of this fraction may be the gradual shortening of oxyethylene chains in the aquatic environment due to the enzymatic hydrolysis of the oxyethylene chain biodegradation pathway [150]. Because of both reasons the control of the concentration of the fraction of ethoxylates having 1 - 4 oxyethylene subunits is yet to be found. The long chain ethoxylates are represented by barely

168 biodegradable oxyethylene oxypropylene block copolymers. This fraction is also uncontrolled. The recently developed method for the determination of this fraction [25,26] supplied the unique information concerning this fraction in the aquatic environment. However, the data should be considerably supplemented to have a clear vision of the role of this fraction in the aquatic environment and checked by another method which needs to be to developed. Alkylpolyglucosides are being introduced into commercial use as more environmentally friendly surfactants than the other groups of non-ionic surfactants. However, this opinion is based on the model biodegradation studies. No method exists to determine polyalkylglucosides in the aquatic environment components. Despite the excellent results of the model investigations, real measurements of concentration of these surfactants in the aquatic environment are needed, to finally confirm a positive opinion concerning the biodegradability of polyglucosides. To check the absence, or control the level in tap and aquifer water, methods for the trace analysis of non-ionic surfactants are required. It may be presumed that only extremely high concentrations of non-ionic surfactants in tap water would be detectable by the methods currently used. Therefore a new method aimed at considerably lower concentrations of non-ionic surfactants should be developed. The presence of non-ionic surfactants in aquifer water would be the measure of the migration of surface water pollution to the aquifer. More advanced analytical tools should be useful in the investigation of biodegradation processes running at a realistic level of concentration. Most biodegradation experiments were performed at a much higher level of concentration than that expected in the aquatic environment samples. The main reason for such a high level of concentration was the fact that the lower concentrations would be undetectable. However, unrealistically high concentrations affect the biodegradation process, considerably modifying it [187]. Therefore, more advanced analytical methods may provide the opportunity for reconsidering these investigations at a realistic level of concentration of non-ionic surfactants. The analysis of metabolites of non-ionic surfactant biodegradation may fill the gap between the pattern of the process observed by primary biodegradation and the pattern of the process exhibited by the monitoring of the total organic carbon or biological oxygen demand reduction. The methods for the analysis of metabolites of oxyethylated alkylphenols are already relatively well established. The exception, however, is the method for the determination of alkylphenoxyethoxy carboxylates which reqiures support by more data and checking by more laboratories. The metabolites of the biodegradation of oxyethylated alcohols correspond to a different pathway than oxyethylated alkylphenols. Though the method for the determination of PEG was developed and its utility demonstrated, it needs to be checked by other laboratories. If PEG is the main metabolite of the biodegradation of oxyethylated alcohols, the other

169 metabolites of the same pathway, such as free fatty alcohols, short chain PEG and free ethylene glycol should also be controlled. The HPLC and flow-injection analysis have the potential to solve most of the analytical problems concerning non-ionic surfactant analysis, provided the detection problem is solved. Unfortunately, oxyethylated alcohols, being the main group of non-ionic surfactants, do not have chromophoric groups which might be used for spectrophotometric or fluorimetric detection. The solution of the problem of detection may open the way for the full automation of measurements and in this way reduce the costs of analysis. REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 16.

17. 18. 19. 20. 21. 22.

E.Pramauro and E.Pelezetti,, "Surfactants in Analytical Chemistry. Applications of Organised Amphiphilic Media", Elsevier, 1996. P.H.Brunner, S.Capri, A.Marcomini and W.Giger, Water Res., 22(1988)1465. J.R.Dojlido,"Chemistry of Surface Water"(pol), Wydawnictwo Ekonomia i Srodowisko, Bialystok, 1995. M.Gerlache, J.M.Kauffmann, G.Quarin, J.C.Vire, G.A.Bryant and J.M. Talbot, Talanta, 43(1996)507. L.Noll, Tenside Surf. Det., 28(1991) 90. Z.Lukaszewski andA.Szymanski, Mikrochim.Acta, 125(1996)185. E.Matthijs and H.De Henau, Tenside Surf. Det., 22 (1985)299. B.Wyrwas, A.Szymanski and Z.Lukaszewski, Anal.Chim.Acta., 331(1996)131. A.Szymanski, Z Swit and Z.Lukaszewski, Anal.Chim. Acta, 311(1995)31. E.Kubeck and C.G. Naylor, J.Am.Oil Chem. Soc., 67(1990)400. Szymanski and Z. Lukaszewski, ,,Preparation of Environmental Samples of Non-Ionic Surfactants for Long Term Storage", International Congress on Analytical Chemistry, Moscow, June, 1997. J.Longwell and W.D.Menice, Analyst, 80(1955)167. D.C.Abbot, Analyst, 87(1962)286. American Public Health Association, Standard Methods for the Examination of Water and Wastewater, 18th Edition, 1992, 5540 Surfactants. Bestimmung der methylenblauaktiven Substanz - Verfahren DIN 38409. Polish Standard PN-85-C-04550/02 ,,Determination of synthetic anionic surfactants by colorimetric method with Methylene Blue", Polish Committee for Standradization, Measures and Quality Control, 1985. H.K.Biswas and B.M.Mandal, Anal.Chem., 44(1972)1636. E.Kunkel, ,,Die Analytik der Tenside", R.Wickbold (ed.), Hills, Marl, 1976, p. 102. J.Waters and W.Kupfer, Anal.Chim.Acta, 85(1976)241. Q.W.Osburn, J.Am.Oil Chem. Soc., 59(1982)453. H.Klotz, Tenside Surf. Det., 24(1987)370. H.Hellmann, Fresenius Z.Anal.Chem., 315(1983)425.

170 23. T.M.Schmitt, M.C.Allen, D.K.Brain, K.F.Guin, D.E.Lemmel and Q.W.Osburn, J.Am.Oil Chem. Soc., 67(1990)103. 24. A.Szymanski and Z.Lukaszewski, Analyst, 121(1996)1897. 25. Z.Lukaszewski, A.Szymanski and B.Wyrwas, ,,Non-Ionic Surfactants in Surface Water Beyond the Routine Determination. Schemes for Their Separation and Determination", International Congress on Analytical Chemistry, Moscow, June, 1997. 26. Z.Lukaszewski and A.Szymanski, ,,The Indirect Two-Monitor Tensammetric Technique: A Tool For Selective Detection in Tensammetry". Deauville Conference: 5th Symposium on Analytical Sciences, Nice, June, 1997. 27. R.Wickbold,Tenside Deterg., 9(1972)173. 28. R.Wickbold, Tenside, 8(1971)61. 29. E.Matthijs and M.Stalmans, Tenside Surf. Det., 30(1993)29. 30. H.Block, W.Gniewkowski and W.Baltes, Tenside Surf.Det., 24(1987)160. 31. M.Ahel and W.Giger, Anal.Chem., 57(1985)1577. 32. D.G.P.A.Breen, J.M.Horner, K.D.Barfle, A.A.Clifford, J.Waters and J.G.Lawrence, Wat.Res., 30(1996)476. 33. E.Matthijs and H.De Henau, Tenside Surf. Det., 24(1987)193. 34. M.A.Castles, B.L.Moore and S.R.Ward, Anal.Chem., 61(1989)2534. 35. A.Moreno, J.Ferrer and J.L.Berna, Tenside Surf. Det., 27(1990)312. 36. E.Matthijs and E.C.Hennes, Tenside Surf. Det., 28(1991)22. 37. J.Sanchez Leal, M.T.Garcia, J.Ferrer and C.Bongoechea, Tenside Surf. Det., 31(1994)253. 38. P.SchSberl, H.Klotz, R.Spilker and L.Nitschkke, Tenside Surf. Det., 31(1994)243. 39. D.Prats, F.Ruiz, B.Vazquez and M.Rodriguez-Pastor, Wat.Res., 31(1997)1925. 40. P.Gerike, H.Klotz, J.G.A.Kooijman, E.Matthijs and J.Waters, Wat.Res., 28(1994)147. 41. J.Waters, J.T.Garrigan and A.M.Paulson, Water Res., 20(1986)247. 42. A.Marcomini, S.Capri and W.Giger, J. Chromat., 403(1987)243. 43. A.Di Corcia, M.Marchetti, A.Marcomini and R.Samperi, Anal.Chem., 63(1990)1179. 44. C.Crescenzi, A.Di Corcia, E.Marchiori, R.Samperi and A.Marcomini, Wat. Res., 30(1996)722. 45. E.M.Thurman, T. Willoughby, L.B.Barber and K.A.Thorn, Anal.Chem, 59(1987)1798. 46. A.Szymanski and Z.Lukaszewski, Anal.Chim.Acta, 281(1993)443. 47. K.Schulze, Tenside Surf. Det., 33(1996)94. 48. H.Palicka, ,,Trends in Detergent Formulations", 3rd International Symposium ,,Chemistry Forum", Warsaw, April, 1997. 49. P.Berth and P.Jeschke, Tenside Surf. Det., 26(1989)75. 50. S.Motomizu, T.Oshima and T.Kuroda, Analyst, 113(1988)747.

171 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63.

64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89.

E.Orthgiess and B.Dobias, Tenside Surf. Det., 27(1990)226. S.Motomizu and M.Kobayashi, Anal. Chim. Acta, 261(1992)471. S.Motomizu, Y.H.Gao, K.Uemura and S.Ishibara, Analyst, 119(1994)473. H.Liu and P.K.Dasgupta, Anal. Chim. Acta, 288(1994)237. J.Kawase, A.Nakae and M.Yamanaka, Anal.Chem., 51(1979)1640. S.Motomizu, M.Oshima and Y.Hosoi, Microchim. Acta, 106(1992)67. S.Motomizu and Y.Gao, Microchem.J., 49(1994)326. J.Liu, Anal.Chim.Acta, 343(1997)33. S.R.Epton, Trans. Faraday Soc., 44(1948)226. V.W.Reid, G.F.Longman and E.Heinerth, Tenside, 4(1967)292. V.W.Reid, G.F.Longman and E.Heinerth, Tenside, 5(1968)90. R.Wickbold, Tenside, 8(1971)130. M.Tsubouchi, N.Yamasaki and K.Matsoka, J.Am.Oil Chem.Soc., 56(1979)921. M.Tsubouchi, H.Mitsushio and N.Yamasaki, Anal.Chem., 53(1981)1957. M.Tsubouchi and J.H.Mallory, Anal.Chim.Acta, 144(1982)249. J.Marcinkiewicz, A.Gorska and W.Zwierzykowski, Tenside Surf.Det., 32(1995)240. L.Cohen, A.Moreno and J.L.Berna, Tenside Surf.Def., 34(1997) 183. N.Buschmann and H.Starp, Tenside Surf.Det., 34(1997)84. K.Vytras, Electroanalysis, 3(1991)343. W.Szczepaniak, Analyst, 115(1990) 145 I. W.Szczepaniak and M.Ren, Electroanalysis, 6(1994)341. A.G.Fogg, A.S.Pathan and D.Thorburn Burns, Anal.Chim.Acta, 69(1974)238. R.A.Lienado, Anal.Chem., 47(1975)2243. D.F.Anghel, G.Popescu and N.Ciocan, Mikrochim.Acta, (1977/II)639. N.Ciocan and D.F.Anghel, Fresenius Z.Anal.Chem., 290(1978)237. S.H.Hoke, A.G.Collins and C.A.Reynolds, Anal.Chem., 51(1979)859. W.Selig, Mikrochim.Acta, (1979/II)437. W.Selig, Mikrochim.Acta, (1979/II)437. W.Selig, Microchim.J., 25(1980)200. W.Selig, Fresenius Z.Anal.Chem., 300(1980) 183. J.Gronsveld and M.J.Faber, Tenside Surf.Det., 27(1990)231. K.Vytras, M.Dajkova and V.Mach, Anal.Chim.Acta, 127(1981)165. T.A.Neubecker, Environm. Sci. Technol., 19(1985)1232. K.Oba, K.Miura, H.Sekiguchi, R.Yagi and A.Mori, Wat.Res., 10(1976)149. J.D.Fairing and F.R.Short, Anal. Chem., 28(1956) 1827. Association of American Soap and Glycerine Producers. Analytical Subcommittee. Anal. Chem., 28(I 956) 1822. H.L.Webster and J.Halliday, Analyst, 84(1959)552. C.P.Ogden, H.L.Webster and J.Halliday, Analyst, 86(1961)22. E.A.Setzkorn and A.B.Carel, J. Am. Oil Chem. Soc., 40(1963)57.

172 90. C.D.Frazee, Q.W.Osburn and R.O.Crisler, J. Am. Oil Chem. Soc., 41(1964)808. 91. J.P.Simko, E.M.Emery and E.W.Blank, J. Am. Oil Chem. Soc., 42(1965)627. 92. R.D.Swisher, J. Am. Oil Chem. Soc., 43(1966)137. 93. W.T.Sullivan and R.D.Swisher, Envir. Sci. Technol., 3(1969)481. 94. J.Waters and J.T.Garrigan, Wat. Res., 17(1983)1549. 95. Q.W.Osburn, J. Am. Oil Chem. Soc., 63(1986)257. 96. C.J.Krueger and J.A.Field, Anal.Chem., 67(1995)3363. 97. J.A.Field, D.J.Miller, T.M. Field, S.B. Hawthorne and W.Giger, Anal.Chem., 64(1992)3161. 98. H.T. Hon-nami and T.Hanya, J.Chromatogr., 161(1978)205. 99. J.McEvoy and W.Giger, Environ.Sci.Technol., 20(1986)376. I00. M.L.Trhy, W.E.Gledhill and R.G.Orth, Anal.Chem., 62(1990)2581. I01. A.Nakae, K.Tsuji and M.Yamanaka, Anal.Chem., 52(1980)2275. 102.A. Marcomini and W.Giger, Anal. Chem., 59(1987)1709. 103. H.De Henau, E.Matthijs and W.D.Hopping, J. Environ. Anal. Chem., 26(1989)279. 104. M.S.Holt, E.Matthijs and J.Waters, Water Res., 23 (1989)749. 105. J.Waters, M.S.Holt and E.Matthijs, Tenside Surf. Det, 26(1989)109. I06. A.Marcomini, S.Busetti, S.Capri, A.Liberatori, T.Noce and A.Sfriso, 6th European Symp. on Organic Micropollutants in Aquatic Envir., Lisbon, 1990. 107. H.A.Painter and T.Zabel, Tenside Surf.Det., 26(I 989) 115. 108. H.De Henau, E.Matthijs and W.D.Hopping, J.Environ.Anal.Chem., 26(1986)279. 109. D.M.Woltering, R.J.Larson, W.D.Hopping, R.A.Jamieson and N.T. de Oude, Tenside Deterg., 24(1987)286. II0. E.Matthijs and H.De Henau, Tenside Deterg., 24(1987)193. 11 I. J.L.Berna, Ferrer J., A.Moreno, D.Prats and F.Ruiz Bevia, Tenside Surf. Det., 26(I 989) 10 I. 112. P.SchSberl, Tenside Surf.Det., 26(1989)86. 113. A. Moreno, J.Ferrer, F.Bevia, D.Prats, B.Vazquez and D.Zarzo, Wat.Res., 28(1994)2183. I14. K.Fischer, Tenside Det., 17(1980)250. 115. P.Gerike, K.Winkler, W.Schneider and W.Jakob, Tenside Surf.Det., 26(1989) 21. 116. P.Gerike, K.Winkler, W.Schneider and W.Jakob, Tenside Surf.Det., 26(1989)136. 117. P.Gerike, K.Winkler, and W.Jakob, Tenside Surf.Det., 26(1989) 271. 118. E.C.Hennes and R.A.Rapaport, Tenside Surf.Det., 26(1989)141. 119. H.Hellmann, Tenside Det., 15(1978)291.

173 120. Z.Lukaszewski, A.Szymanski and B.Wyrwas, Surfactants in the River Warta in Poznan within the Period of 1989-1994, 35th IUPAC Congress, Istanbul, 1995. 121. M.Kikuchi, A.Tokai and T.Yoshida, Wat. Res., 20(1986)643. 122. R.Wickbold, Tenside Det., 13(1976)33. 123. J.Waters and J.T.Garrigan, Water Res., 17(1983)1549. 124. W.E.Gledhill, L.R.Huddleston, L.Kravetz, A.M.Nielsen, R.I.Sedlak and R.D.Vashon, Tenside Surf.Det., 26(1989)276. 125. P.Gerike, K.Winkler, W.Schneider and W.Jakob, Tenside Surf. Det., 26(1989)136. 126. R.Wickbold, Tenside Det., 13(1976)32. 127. G.V.Scott, Anal.Chem., 40(1968)786. 128. I.Sheiham and T.A.Pinfold, Analyst, 94(1969)387. 129. T.K.Christopopoulos, E.P.Diamandis and T.P.Hudjiioannou, Anal. Chim. Acta, 143(1982) 143. 130. K.Vytras, J.Kalous and I.Symersky, Anal. Chim. Acta, 177(1989)219. 131. C.Gavach and P.Seta, Anal.Chim.Acta, 50(1970)407. 132. K.Vytras, Collect. Czech. Chem. Commun. , 42(1977)3168. 133. S.S.M.Hassan and I.H.A.Badr, Talanta, 41(1994)523. 134. M.Morak, Tenside Surf.Det., 26(1989)215. 135. I.A.Gurev, G.M.Lizunova and N.S.Bulanova, Zh.Anal.Khim., 34(1979)1809. 136. I.A.Borisova and I.A.Gurev, Zav.Lab., 45(1979)309. 137. I.A.Gurev and T.S.Vyatchanova, Zh.Anal.Khim., 34(1979)976. 138. L.Daradics, Rev.Chim. (Bucharest), 28(1977)1098. 139. E.R.Michelsen, Seifen-01e-Fette-Wachse, 104(1978)93. 140. E.R.Michelsen, Tenside, 15(1978)169. 141. B.W.Topping and J Waters., Tenside Surf. Det., 19(1982)164. 142. V.T.Wee, Wat.Res., 18(1984)223. 143. V.T.Wee and J.M.Kennedy, Anal.Chem., 54(1982)1631. 144. E.Matthijs and H.De Henau, Vom Wasser, 69(1987)73. 145. C.De Ruiter, J.C.H.F.Hefkens, U.A.T.Brinkman, R.W.Frei, M.Evers, E.Matthijs and J.A.Meijer, Intern.J.Environ.Anal.Chem., 31 (1987)325. 146. R.Weber, K.Levsen, G.J.Louter, A.J.H.Boerboom and J.Haverkamp, Anal.Chem., 54(1982) 1458. 147. D.M.Woltering, R.J.Larson, W.D.Hopping and R.A.Jamieson, Tenside Surf. Det., 25 (1988) 116. 148. D.M.Woltering, R.J.Larson, W.D.Hopping and R.A.Jamieson, Tenside Surf. Det., 24(1987)286. 149. W.Kupfer, Tenside Deterg., 19(1982)158. 150. R.D.Swisher, ,,Surfactants Biodegradation", Second Edition, Revised and Expanded, Marcel Dekker Inc., New York, Basel, 1987. 151. J.Cross, ,,Nonionic Surfactants. Chemical Analysis", Marcel Dekker Inc., New York, Basel, 1987.

174 152. T.M.Schmitt, ,~nalysis of Surfactants", Marcel Dekker Inc., New York, Basel, 1992. 153. Die Analytik der Tenside, Wickbold R. (ed.), Hills, Marl, 1976. 154. G.J.Moody and J.D.R.Thomas, ,,Potentiometry of Oxyalkylates", in ,,Nonionic Surfactants. Chemical Analysis", J.Cross (ed.), Marcel Dekker Inc., New York, Basel, 1987. 155. P.H.V.Alexander, G.J.Moody and J.D.R.Thomas, Analyst, 112(1987)113. 156. K.Vytras, V.Dvorakova and I.Zeman, Analyst, 114(1989)1435. 157. R.D.Gallegos, Analyst, 118 (1993)1137. 158. C.Baillarger, A.Mayaffre, M.Turmine, P.Letellier and H.Suquet, Electrochim. Acta, 39(1994)813. 159. Application Bulletin No. 230/1 e, Metrohm. 160. Polish Standard PN-88-C-04550/11."Determination of non-ionic surfactants in water by spectrophotometric method using Dragendorff reagent". Polish Committee for Standradization, Measures and Quality Control, 1988. 161. K.Bilrger, Fresenius Z. Anal. Chem., 196(1963)251. 162. D.Brown, H.De Henau, P.Gerike, M.Holt, E.Keck, E.Kunkel, J.Matthijs, J.Waters and R.J.Watkinson, Tenside Det. 23(1986)190. 163. B.Wyrwas, A.Szymanski and Z.Lukaszewski, Anal.Chim. Acta, 278(1993)197. 164. A.Mathias and N.Mellor, Anal.Chem., 38(1966)472. 165. C.Slagt, J.M.H.Daemen, W.Dankelman and W.A.Sipman, Z.Anal.Chem., 264(1973)401. 166. K.Tsuji and K.Konishi, J.Am.Oil Chem. Soc., 51(1974)55. 167. I.I.Kaduji and J.B.Stead, Analyst, 101(1976)728. 168. R.S.Tobin, F.I.Onuska, B.G.Brownlee, D.H.J.Anthony and M.E.Comba, Wat.Res., 10(1976)529. 169. V.T.Wee, ,,Determination of Linear Alcohol Ethoxylates in Waste- and Surface Water", in ,,Advances in the Identification and Analysis of Organic Pollutants in Water", L.H.Keith (ed.), Ann Arbor Science Publishers, Inc., Michigan, 1981,. 170. L.Favretto, B.Stancher and F.Tunis, Analyst, 105(1980)833. 171. J.Chlebicki and W.Gancarz, Chem. Analit.(Warsaw), 24(1979)675. 172. J.Chlebicki and W.Gancarz, Tenside Det., 17(1980)13. 173. W.Winkler and F.Buhl, Przem.Chem. (Warsaw), 73(1994)182. 174. W.Winkler and F.Buhl, Przem.Chem. (Warsaw), 73(1994)314. 175. Z.Kozarac, V.Zutic and B.Cosovic, Tenside Det., 13(1976)260. 176. H.Batycka and Z.Lukaszewski, Anal.Chim. Acta, 162(1984)207. 177. M.K.Pawlak and Z.Lukaszewski, Anal.Chim. Acta, 202(1987)85. 178. A.Szymanski and Z.Lukaszewski, Electroanalysis, 3(1991) 17. 179. A.Szymanski and Z.Lukaszewski, Electroanalysis, 3(1991)963. 180. A.Szymanski and Z.Lukaszewski, Electroanalysis, 7(1995) 114. 181. A.Szymanski and Z.Lukaszewski, Electroanalysis, 6(1994) 1094.

175 182. H.Batycka and Z.Lukaszewski, Anal.Chim. Acta, 162(1984)215. 183. Z.Lukaszewski, H.Batycka and W.Zembrzuski, Anal.Chim. Acta, 175(1985)55. 184. M.K.Pawlak and Z.Lukaszewski, Anal.Chim. Acta, 202(1987)97. 185. A.Szymanski and Z.Lukaszewski, Anal.Chim. Acta, 231(1990)77. 186. A.Szymanski and Z.Lukaszewski, Anal. Chim. Acta, 281(1993)443. 187. A.Szymanski, T.Jaroszynski, P.Jeszka and Z.Lukaszewski,Wat.Res., 30(1996)2465. 188. A.Szymanski and Z.Lukaszewski, Anal.Chim. Acta, 260(1992)25. 189. A.Szymanski and Z.Lukaszewski, Anal.Chim. Acta, 293(1994)77. 190. A.Szymanski and Z.Lukaszewski, Anal.Chim. Acta, 273(1993)313. 191. A.Szymanski, B.Wyrwas and Z.Lukaszewski, Anal.Chim. Acta, 305 (1995)256. 192. B.Wyrwas, A.Szymanski and Z.Lukaszewski, Talanta, 41(1994) 1529. 193. B.Wyrwas, A.Szymanski and Z.Lukaszewski, Talanta, 42 (1995)1251. 194. B.Wyrwas, A.Szymanski and Z.Lukaszewski, Anal.Chim. Acta, 331(1996)131. 195. Z.Lukaszewski, A.Szymanski and B.Wyrwas, ,,Tensammetric Determination of Non-Ionic Surfactants Adsorbed on Alive Activated Sludge", Deauville Conference: 5th Symposium on Analytical Sciences, Nice, 1997. 196. B.Wyrwas, A.Szymanski and Z.Lukaszewski, Talanta, in the press. 197. Z.Lukaszewski, A.Szymanski and B.Wyrwas, ,,Monitoring of Primary Biodegradation of Oxyethylated Alcohols in the O ECD Confirmatory Test by Tensammetry", Deauville Conference, 4th Symposium on Analytical Sciences, Brussels, 1996. 198. M. Abel and W.Giger, Anal.Chem., 57(1985)2584. !99. M. Ahel, W.Giger and M.Koch, Wat.Res., 28(1994)1131. 200. M. Ahel, W.Giger and C.Schaffner, Wat.Res., 28(1994) 1143. 201. M.Kudoh, H.Ozwa, S.Fudano and K.Tsuji, J.Chromatog., 287(1984)337. 202. M.S.Holt, E.H.McKerrell, J.Perry and R.J.Watkinson, J.Chromatog., 362(1986)419. 203. K.Yoshimura, J.Am.Oil Chem. Soc., 63(1986)1590. 204. M.C.Allen and D.E.Linder, J.Am.Oil Chem. Soc., 58(1981)950. 205. A.Szymanski, K.Tomaszewski, B.Wyrwas and Z.Lukaszewski, nonpublished data. 206. L.Nitschke and L.Huber, Fresenius J.Anal.Chem., 345(1993)585. 207. Z.Lukaszewski, A.Szymanski and B.Wyrwas, ,,Selective Determination of Oxyethylated Alcohols Having a C16.1s Hydrophobic Part in the Mixture of Non-Ionic Surfactants by Tensammetry", Euroanalysis IX, Bologna, 1996. 208. M.Ahel, C.Schaffner and W.Giger, Wat.Res., 30(1996)37. 209. A.Di Corcia, R.Samperi and A.Marcomini, Environ.Sci.Technol., 28(1994)850.

176 210. Z.Lukaszewski, A.Szymanski, B.Wyrwas and K.Tomaszewski, ,,Determination of Poly(Ethylene Glycols) by the Indirect Tensammetric Method Combined with the BiAS Separation Scheme, Deauville Conference: 4th Symposium on Analytical Sciences, Brussels, 1996. 211. Z.Lukaszewski, A.Szymanski and B.Wyrwas, ,Application of Tensammetry for the Monitoring of Metabolic Products of Biodegradation of Oxyethylated Alcohols", 6th European Conference on ElectrAnalysis, Durham, 1996. 212. Z.Lukaszewski, A.Szymanski and B.Wyrwas, ,,Poly(Ethylene Glycols) as Metabolites of Non-Ionic Surfactants Biodegradation in the Aquatic Environment" 9th International Conference on Surface and Colloid Science, Sofia ,1997. 213. U.Zoller, Wat.Res., 28(1994)1625. 214. Z.Lukaszewski, L.Burchardt and A.Szymanski, ,,Non-Ionic and Anionic Surfactants in the Lakes of the ,,Golaniecka" Stream (Poland), 5th Polish Conference of Analytical Chemistry, Gdansk, 1995.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

177

A d s o r p t i o n of s u r f a c t a n t s o n d i e s e l e n g i n e soot a n d i t s a p p l i c a t i o n i n carbody washing A. E1 Ghzaoui and S. P a r t y k a Laboratoire des Agr~gats Mol6culaires et Mat~riaux Inorganiques, ESA 5072 du CNRS, Universit6 Montpellier II, place Eugene Bataillon, C. C 015, 34095 Montpellier, France

1.

INTRODUCTION

The cleaning of a solid surface involves the removal of unwanted foreign material from its surface. In detergency, the interaction between solid surfaces and bulk phase is of fundamental interest. The amphiphilic molecules that preferentially adsorb at the interface play an important role in such phenomena. The detergent action during the cleaning is due to the interaction between the hydrophobic or polar moiety of the surfactant with the dirt and the substrate. Schwartz [1] lists three main mechanisms for removal of liquid soils from a surface by surfactant solutions. The first is the so-called rollback mechanism which depends on the wetting properties of the surfactant. Basically, the oily soil on the substrate to be washed retracts to form drops which detach from the surface. The second mechanism is emulsification of oily soil. Here the adsorption of surfactants at the oily soil-substrate interface, with lowering of the interfacial tension, may play an important role. The third mechanism is solubilization into a surfactant micelle. Another mechanism of cleaning is the formation between the soil and the surfactant of an intermediate phase which can be a lamellar liquid crystal. The extent and efficiency of solubilization of the oily soil depend on the chemical structure of the surfactant used for the detergent formulations, the temperature, the ionic strength, and the pH. Removal of particulate soils is achieved by different mechanisms and depends on the dirt type. Despite these differences, a reasonably general mechanism of deflocculation and suspension by surfactant adsorption has been successful in describing particulate soil removal for a variety of fiber surfaces [2,3]. The role of the surfactant appears to be twofold. First it aids in wetting the soil agglomerate and the surface by adsorbing to them. This causes a thin film of water to penetrate between the particles forming the agglomerate, and between the particles and the surface. As a consequence, the adhesion work required to

178 remove the particle soil from the substrate is decreased. The work of adhesion, Wa, is given by the expression [4]: Wa

= ~/SB + ~/PB -~SP

where subscripts SB, PB and SP refer respectively the interfaces between substrate and aqueous solution of surfactant, soil and aqueous solutions of surfactant and substrate and soil. Adsorption of surfactants at these interfaces can result in a decrease in 7SB and ~/PB, with a decrease in the work required for detachment of the soil particles from the solid surface. Diesel-exhaust particles (DEP) or soot particles are the main cause of urban pollution. This pollution is mainly formed by diesel engines and particularly if the fuel-air ratio is badly regulated [5]. It has an impact on h u m a n health and causes an environmental problem. These particles can be very small, and easily reach far down into lung tissue, when inhaled. They also deposit on any surfaces exposed to this pollution. So, if one wants to keep the surface clean, one has to wash it with an effective detergent formulation. The efficiency of the detergent formulation involves knowledge of the soil characteristics on the one hand, and of the surfactant adsorption on the soot particles and on the surface on the other hand. If the surfactant adsorbs on the DPE particles which strongly adhere to any surface, it alters the electrical and chemical properties of the both materials and causes an increase of the interfacial potential. As consequence, an electrostatic repulsion appears between particles within an agglomerates or between soil particles and the surface. The second effect of surfactant adsorption is a wetting phenomenon which facilates a thin film of water to penetrate between the particles forming the agglomerate. The consequence of this wetting is the defloculation of the aggregates caused by a decreasing in the attractive interaction between the particles. This effect is particularly important for hydrophobic soil. The modification of physicochemical state of the agglomerates and the solid surface is the most important conditions for the efficiency of the surfactants in the cleansing process. The aim of this chapter is a presentation of the physicochemical characteristics of diesel-exhaust particles and on experimental methodological approach to carbodies cleaning. The emphasis here is on the surfactant adsorption on soot particles and on the polyurethane surface as a model of a carbody surface. The results provided from adsorption data will help to select the best surfactants to be used for an efficiency detergent formulation.

2. MATERIALS AND EXPERIMENTAL TECHNIQUES 2.1. Experimental techniques The most widely used approach to the direct measurement of the amount adsorbed is to study the depletion of the surfactants from solution in equilibrium

179 with the adsorbent. The problem then reduces to the d e t e r m i n a t i o n of the s u r f a c t a n t molality in the presence of sufficient interface per unit volume to cause m e a s u r a b l e change in molality. The method is therefore applicable only when the specific surface area of the solid is not too small. 0.5 grams of diesel engine soot was add to 20 g of surfactant solution. These suspensions were sealed in clean glass tubes and then agitated for 12 h in a t h e r m o s t a t . The s u p e r n a t a n t was s e p a r a t e d from the solid by centrifugation for 15 min at 12000 rpm and then analysed with Total Organic Carbon (Shimatzu). The surface excess was calculated according to F= (Ct~

Ceq)ml ms A

(1)

where Ctot and Ceq are the total and the equilibrium concentrations of surfactant, ml is the initial mass of solvent, ms denotes the mass of soot and A the specific area of the adsorbent. A R a n k Brothers microelectrophoresis a p p a r a t u s with a r e c t a n g u l a r cell was applied to m e a s u r e the average velocity at which charged soot particles moved under the action of a steady and weak electric field between p l a t i n u m electrodes. From the average velocity at both stationary levels, the electrophoretic mobilities (p) of soot particles were calculated by the following relationship: V = -E

(2)

where V and E are the velocity of the particles and the electric field respectively. The electrophoretic mobilities of the soot particles were d e t e r m i n e d under the same conditions as those used to obtain the adsorption isotherms. After the a t t a i n m e n t of adsorption equilibrium and centrifugation, samples of the soot suspension from the s u p e r n a t a n t s were transferred to a t h e r m o s t a t e d microelectrophoresis cell. The surface tension of the surfactant solutions was m e a s u r e d with an electrobalance type tensiometer (Prolabo TD-2000). The area (a0) per surfactant at the air-solution interface at surface s a t u r a t i o n has been determined by applying the Gibbs equation. The comparison of the values obtained with the corresponding values at the soot-solution interface at surface s a t u r a t i o n provides informations on the packing area and eventually on the orientation of adsorbed surfactants at the soot-solution interface. The m e a s u r e m e n t s of pH were performed using a tacussel pH electrode (C 601). The d e t e r m i n a t i o n of soot particles d i a m e t e r was performed using a Mastersizer E M a v e r n apparatus. The chemical analyses of soot were performed with an X-ray emission spectrometer attached to a Stereoscan 360 Cambridge Electronic Microscope.

180

The heat of immersion data were obtained using the Calvet microcalorimeter and the surface area of the solid was measured by the BET method. For both experiments, the soot was outgassed at 150~ under a vacuum of 10 .3 torr for 5 h. The presence of organic impurities in the soot particles influences very strongly the heat of immersion and also the surface area. The turbidity was measured with a Varian spectrophotometer (Carry 3E). Turbidity is the fractional decrease in the intensity of a primary beam passing through a suspension. By analogy with the Lambert-Beer law: dI = (x + e)dx I

(3)

where T is the turbidity, e the absorbance of the suspended particles and x the path length of the light through the suspension. Considering both light absorbency and scattering phenomena in apparent turbidity yields: (4)

x'=(T+e)=--xlln(~-/

Thus, the apparent turbidity x' for a given path length is defined by the logarithm of the ratio of the light intensity I0 passing through the reference pure liquid to the light intensity I passing through the particle suspension.

2.2. A d s o r b a t e s u r f a c t a n t s

The surfactants used in this chapter belong to the four families of surfactants. Below are listed all the surfactants with their chemical formula. a. Anionic surfactant Sodium dodecyl sulfate (SDS), n-C12H25SO-4 supplied by Prolabo (France) and used as received (98 % purity). b. Cationic surfactants Dodecyltrimethylammonium bromide, (DTAB), n- C12H25N§ Tetradecyltrimethylammonium bromide, (TTAB), n- C14H29N§ supplied by Sigma (France) and used as received (98 % purity).

-,

c. Nonionic surfactants Octylbenzene polyoxyethylene, CsH17C6H4(OCH2CH2)10-OH, (TX100) Octylbenzene polyoxyethylene, CsH17C6Hn(OCH2CH2)16-OH, (TX165) Nonylbenzene polyoxyethylene, C9H19C6Hn(OCH2CH2)lo-OH, (TN111)

d. Zwitterionic surfactant n-dodecyl betaines, C12H25N§

-, (NDB).

-,

181 Among the different properties of surfactants, those resulting from their a m p h i p h a t i c structure, the property of being adsorbed at interfaces and t h a t of forming colloidal-sized clusters in aqueous solutions, are the most important. The former m a y be characterised by the effectiveness of adsorption, whereas the l a t t e r by the critical micelle concentration (cmc). The area per molecule at surface saturation, ao, is a useful m e a s u r e of the effectiveness of the surfactant adsorption at the solution-air interface, since it corresponds to the m a x i m u m value which adsorption can reach. The cmc represents the m a x i m u m solubility of the single molecules in a given aqueous medium and thus plays an i m p o r t a n t role in the s u r f a c t a n t adsorption onto solid substrates, where single ions r a t h e r t h a n micelles are involved.

Table 1 Critical micelle concentration (cmc), and area per molecule at surface saturation, (a0), for the s u r f a c t a n t molecules in deionised w a t e r Temperature cmc ao Surfactant Solvent [mmol kg -1] [nm2/molec.] [K] SDS

298

water

8.5

72

TTAB

298

water

4.0

56

DTAB

298

water

12.5

57

TX100

298

water

0.27

TX165

298

water

0.5

106

TNlll

298

water

0.1

66

NDB

298

water

2.0

48

58

The above cmc values of all studied surfactants have been established from the surface tension data r e p r e s e n t e d in Figures 1-7. Each plot shows a break which corresponds to the cmc in deionised w a t e r at 298 K. The m i n i m u m observed in the ~, value m e a n s t h a t the surfactant contains some impurities (SDS and NDB). The most interesting result which issues from the surface tension d a t a is t h a t the lowest y values are obtained for the nonionic surfactants and particularly for TX100 and TN 111.

182 70

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

65 60 'E

Z E

55 50 cmc

,,4-45 40 35 30 -8

-7

-6

-5

-4

In C Figure 1. Surface tension of aqueous solution of TTAB against the logarithm of the molality. 0 .................................................................................................................... 65 60 55 '~ 50 9

>~ 45 ,...1

C~

40

9

V 9

35

O0

9

O0

O0

30 25 -9

-8

-7

-6

-5

-4

-3

-2

lnC Figure 2. Surface tension of aqueous solution of SDS against the logarithm of the molality.

183 55 50 45 40

cn'~

>-

35 30

-12

-11

-10

-9

-8

-7

-6

lnC Figure 3. Surface tension of aqueous solution of TX100 against the logarithm of the molality.

0

.................

~

.............................................................................................................................................

55

50 Z 45

ClYIC

40

oV 00

O QO

35 30 -10

-9

-8

-7

-6

-5

-4

-3

lnC Figure 4. Surface tension of aqueous solution of NDB against the logarithm of the molality.

184

5 .........................................................................................................................................................................................................

50 C

~

'-~ 45

?.-.

40

"/ 35

30 -11

-10

-9

-8

-7

-6

-5

lnC Figure 5. Surface tension of aqueous solution of TX165 against the logarithm of the molality. 70 65 60 55 9

cmc

5O 45

V 40 35 3 0

'

-8

. . . . . . . .

~

-7

..............

~. . . . . . . . . .

-6

~

~

-5

-4

9. . . .

i

-3

-2

lnC Figure 6. Surface tension of aqueous solution of DTAB against the logarithm of the molality.

185 5

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

50 45 Z

C~

40

>~

35

% 0o O0

30 25 -13

-12

-11

-10

-9

-8

-7

lnC Figure 7. Surface tension of aqueous solution of TN 111 against the logarithm of the molality.

3.

CHARACTERISATION OF DIESEL ENGINE SOOT

The soot particles were collected in the exhaust pipe of diesel engines. The particulate emissions from a diesel engine are composed of both solid and liquid compounds and are very complex in character. Elementary chemical analysis by X-emission spectroscopy showed that it contained iron, sulphur, calcium, silicium and other elements in trace amounts: zinc, copper, aluminium. The comparison of the chemical composition of diesel engine soot and airborne dust, (Figures 8 and 9), shows that the latter contains more mineral oxide t h a n the former. The solid part of the soot powder contains primarily agglomerations of small carbonaceous subparticles. The organic quantity present in the soot was detected by the Total Organic Carbon method and surface tension measurements. The surface tension of the supernatant which is the mixing between 1 g of untreated diesel engine soot and 50 g of deionised water is 66 mN'm -1 while this value is 72.5 mN'm -1 for the mixture of washed diesel engine soot and deionised water. The decrease of the surface tension from 72.5 to 66 mN.m -1 is the consequence of the presence of organic matter. The organic component soluble in water represents about 2.20 mg for 1 g of soot. These organic molecules are adsorbed onto the surface of the soot particles. Typical chromatography analysis show the existence of more t h a n 300 peaks [6], while the quantitative chemical evaluation of the soot composition shows that there are between 10 and 15 % mineral oxides,

186

0.20 % organic polar compounds and between 84.8 and 89.8 % solid carbon. The pH of 0.5 g of soot immersed in 50 g of deionised water is 4.2.

7

l

:!; i

'~

,

'

H . !. i" ~1 ] ~ I

,,

.

"~.

";.

-~

~!!-:1

i~

tiil I

~i~..

,ii~

ii~

:: q !

,

i"

i"

-

" " . . . . . . . ...... . . . 9 ...

:.. . . . . . . .

:

"

...

9. . . . .:... . %

'

'~"~,r .1 ~.. . . . .

9

.~ : : . . . . . . . . .

, q.

.

~:!z:~

,

~. ~..~.

9:::

iii]

~!?k!:i~#..~, : ~ ,. .... .... :~'. i~!,.~.:, i., ;-

...-: ,],r -"

C:

5,

~'

9 9 ~4 ....

7

r,1 il ,-, ;! ' ~ '~1-"

i:~l

:l~ii.: . .

5r113

~ :'~ ..

.. :

9

,

r

:

-.

,, ,I ,~.: , " t , ~ ~; , { ~ p . t , . , 7 ~ . , 4 ~

,~

-, .

" -. ".

" ~t,..

,~...~'-.-;~,t~.~k.%,~,p

t-::,:-U

lr~.

6

Figure 8. Energy dispersive X-ray analysis of diesel engine soot material.

i

i

i !i' .. .:,

"J

"i

,! i.i .. :..~ ..~

i

i

::!

i~ ~i~. -:

:.. ..

.

,,-~:.~

,-.

,:

-

!#~

,

:.:~ 9

-

,

i

,,)

. ,

~

' ,:

.~

"

9

i

"

',

"

~.

",

~, _ , . .

'.t.

,

9

"

2;-;

".-, ,~ . . . . . . . . .

.

....

.,,_...

_

..,~

..

9

.

.

.

.

Figure 9. Energy dispersive X-ray analysis of airborne dust material.

.

.

:::.

187 The particles diameter was measured by the light scattering technique. Figures 10 and 11 show the average diameter of soot particles without any t r e a t m e n t and soot washed with two organic solvents (hexane and methanol).

,5

..................................................................................

2,5

1,5

0,5

diameter, ~n Figure 10. Particle size distribution of the original diesel engine soot. ,5

.......................................................................................................

3,5

2,5 2 1,5

0,5

Ii Ii!i11111 lil, i

diameter, Figure 11. Particle size distribution of the washed diesel engine soot.

188 In both cases the distribution of the particles diameter is large. The average diameter of the soot washed by two organic solvents is about 6 pm while it is about 19 pm for the u n t r e a t e d soot. This variation is certainly the result of the solubilization of organic m a t t e r which coated the particles and behaves as a cement between the particles. The solubilization of the organic m a t t e r causes a defloculation of the aggregates. This explains the decrease of the average diameter. The decrease of the particles diameter is also confirmed by scanning electron microscopy. The specific surface area measured by nitrogen adsorption at 77 K varies from 52 m 2 g-1 for the u n t r e a t e d soot to 70 m 2 g-1 for the soot washed with two solvents (Figures 12 and 13). This variation is due to desorption of organic molecules and also to the decrease of the average diameter of the soot particles.

40

S~t

35

52 m 2 g-1

9

30 E r

25

o,r~

20

~

15

o E

10

0

0,2

0,4

0,6

0,8

1

P/Po Figure 12. Nitrogen adsorption isotherm on the unwashed diesel engine soot at 77 K.

An a t t e m p t has been made to assess the polarity of the soot particles. This polarity is deduced from interactions between the solid and different probe liquids by using immersion microcalorimetry [7]. Immersion calorimetry gives the change of enthalpy resulting from the creation of the soot-liquid interface. The values of enthalpy of immersion of the treated and untreated diesel engine soot are shown in Table 2. The values obtained for the washed soot in non polar and polar liquids are exceptionally high. It is recalled that usual values of enthalpy of immersion for mineral oxides such as A1203, SiO2 in the same liquids are about 300 m J m -2 [8]. The enthalpies of immersion of the u n w a s h e d soots in the same liquids are much smaller which is understandable because the u n w a s h e d soot particles are covered by several layers of organic molecules. When

189 this coated m a t e r i a l is in contact with the pure liquids, the enthalpy of immersion is due to hydrophobic-hydrophilic interactions and consequently very low. To conclude of this section we note t h a t the enthalpy of immersion reflects a very strong surface activity of the washed soot surface. 45

0

Sbet-- 70 m2 g-1

40 -~o 35 .

~

<>0

3o

O~

O@

oq~

}

q. OO O O

.~ 25 0 r~

~

20

o 15

9 Adsomtion

10

<>Desorption

.......................

0

,

0,2

i

0,4

0,6

0,8

1

P/Po Figure 13. Nitrogen adsorption-desorption isotherms on the washed diesel engine soot at 77 K.

Table 2 Enthalpy of i m m e r s i o n of diesel engine soot in different liquids E n t h a l p y of immersion of E n t h a l p y of immersion of Liquids t r e a t e d soot [mJ m 2] u n t r e a t e d soot [mJ m 2] n-heptane

145

7O

water

791

100

formamide

917

120

190 3.1. A d s o r p t i o n o f s u r f a c t a n t s o n t o d i e s e l e x h a u s t p a r t i c l e s

We have studied the physicochemical behaviour of several surfactants at the interface between the washed soot and the aqueous solution of the surfactant. Figures 14-18 show the adsorption isotherms of anionic, cationic, nonionic and zwitterionic surfactants and the corresponding variation of the electrophoretic mobility of the soot particles with the equilibrium molality.

10

.......................................................................................................................................... 0

-0,5 00

0

9

9

-1 1,5 =

O

6

O 9

oF

z:t.

-2

og 4

-2,5

0

:s

-3

Q

-3,5

O

-4

O

O

b

0

>"

-4,5 2

4

6

8

10

12

14

16

Ceq, mmol kg-I Figure 14. Adsorption isotherm of SDS and the corresponding electrophoretic mobility curve for the washed soot at 298 K. The common characteristic for all the surfactants studied is a very clear isotherm plateau which is reached around the corresponding cmc values. The amounts adsorbed at the plateau (Fmax) are reasonable and they are similar to the data obtained for hydrophobic or hydrophilic adsorbents [9,10]. A very high Fmax is observed for SDS surfactant (8.4 gmol m-2), although the significant increase in electrophoretic mobility indicates a progressive increase of the negatively charged surface patches. This tremendous amount adsorbed means that in the interfacial region, giant aggregates and may be 3D dimensioned aggregates are present. The isotherms of adsorption of TTAB and DTAB have quite similar shapes to those for hydrophilic silica surfaces [11,12]. The Fmax in both cases is identical. The analysis of the corresponding electrophoretic mobility curve suggests an electrostatic adsorption at the beginning of the isotherm. This recalls that the soot surface is initially negatively charged and becomes positively charged even at the very low adsorption coverage.

191 4

.............................................................................................................................................................................. 3,5 lID

3,5 '~

e e o'~ o

9

O

o o

3

3

2,5 _

2

o 1,5

1,5 1

::[

1

0,5

0,5

0

0 0

3

6

9

12

15

Ceq, mmol kgl Figure 15. Adsorption isotherm of TTAB and the corresponding electrophoretic mobility curve for the washed soot at 298 K.

From an equilibrium molality of about 3 mmol'kg -1, the electrophoretic mobility is then positive and constant. The adsorption mechanism is therefore composed of two simultaneous adsorption processes: firstly the cationic surfactant molecules adsorb on all available negatively charged sites situated on the different oxides, which causes the changing of the sign from negative to weakly positive, while secondly one observes adsorption of surfactants on the hydrophobic part of the surface via dispersion hydrophobic interactions between the alkyl moiety of the surfactant and the surface. Since the length of the alkyl chain of TTAB is larger t h a n that of DTAB, the cmc appears at smaller molality which is a main reason to note that the saturation plateau of adsorption of TTAB is formed earlier. We can underline that Fmax is much higher for TTAB. The cationic surfactants form aggregates at the interface between the soot and the aqueous solution of surfactant. The larger the aliphatic tail of the surfactant, the higher is the aggregation number, and by analogy the interfacial aggregate will be bigger. Consequently the amount adsorbed is higher for TTAB than DTAB (Figure 16). The shape of the adsorption isotherm of the nonionic surfactants on the soot material ressembles the isotherms of the same surfactants adsorbed on the activated carbons [13-16]. They are of the Langmuirian type. The amount adsorbed at the plateau, Fmax, depends strongly on the length of the polar chain

192 1,8

4 -

(a) 9

1,5

"

,t

@ 3 o-

(b)

9

9

0

0

0

1,2

0,9 0

0

0

2 i

0

E

0

D

0,6 0,3

p

0

"0

0

lqP

oTXI00 oTX165

0o o

""

,g

0

oDTAB oTTAB

0

1

0

I

I

I

0,2

0,4

0,6

Ceq, mmol kg

0,8

1

0 q

I

I

I

I

0

5

10

15

20

-1

25

Ceq, mmol kg"

Figure 16. Adsorption isotherms of TX100 and TX165 on washed soot and the corresponding electrophoretic mobility curve (panel a) and those for TTAB and DTAB at 298 K (panel b). in such a way that Fmax increases when the polar chain decreases (Figure 16). Conversely, for a polar group, when the length of the alkyl tail increases, ['max increases. This phenomenon appears at a smaller molality. The general trend of nonionic surfactant adsorption is characterised by two steps: for low coverage values the surfactants are adsorbed by their hydrophilic and hydrophobic moieties. In this first stage the molecules are certainly flat on the surface. When the coverage value increases, there is a second step; the hydrophilic parts of the molecules are repelled from the surface and the surface phase becomes thicker. We can note that the negative charge of the surface is not affected by adsorption of nonionic surfactant (Figure 17).

'7, 0

1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 0,2 0

0 -0,05 -0,1 -0,15 0 O0

00

0 0

OF o~

"7 "7 r

>

-0,2 -0,25 -0,3 -0,35

0,2

0,4 0,6 Ceq, mmol kgl

-0,4 0,8

Figure 17. Adsorption isotherm of TN111 and the corresponding mobility curve for the washed soot at 298 K.

193 The adsorption of zwitterionic surfactant, NDB, on the soot surface is initially characterised by the strong adsorption at low surface coverage (Figure 18). One observes a linear increase of adsorption until the isotherm plateau is reached in the cmc region. The simultaneous measurements of the electrophoretic mobility show the changing of the sign at low coverage. This change is well correlated with the vertical part of the isotherm and suggests adsorption of surfactant via electrostatic interactions between negatively charged surface sites and the cationic group of NDB. Then, in turn, the mechanism of adsorption alreasy described above starts with the alkyl chain of surfactant on the hydrophobic patches of the surface and simultaneous formation of interfacial aggregates. The amount adsorbed at the plateau, ['max, is relatively high and comparable to the value obtained for this kind of surfactant on hydrophilic surfaces [17]. 0,6 O 9 O

0

0

0

0 9

O 9

0,5

9

0,4

0 E

0

9

0,3

0

0,2

9

o

0,1

o~t

-0,1 0

' 0

2

,

i

4

6

-0,2 8

10

12

14

Ceq, mmol kg1 Figure 18. Adsorption isotherm of NDB and the corresponding mobility curve for the washed soot at 298 K.

To summarise the results of surfactant adsorption on the soot surface, we can draw the following observations: - the soot is composed of two types of surface: hydrophilic and hydrophobic, - the surface is initially negatively charged, the negative charge of the surface can be altered by the adsorption of cationic and amphoteric surfactants while adsorption of anionic surfactant increases the negative charge of the surface, - the adsorption of nonionic surfactant does not influence the charge of the surface.

194 Furthermore the kinetic investigation of surfactants adsorption on this kind of material shows that the equilibrium of adsorption is reached after about 5 min (Figure 19). This time is relatively long compared with the time necessary for car washing which is usually from 10 to 15 min. 0,9 0,8

-

r ,

E -:. o E m.

0,70,6]0,5I

0,4 30

60

I

I

I

I

I

I

I

90 120 150 180 210 240 270 300 temps, s

Figure 19. Adsorption kinetic of TN 111 on the washed soot at 298 K.

4.

A D H E S I O N OF S O O T P A R T I C L E S TO P O L Y M E R I C S U R F A C E

At the solid-liquid interface, the strength of adhesion can be expressed quantitatively as the free energy of adhesion. In the case of solid-solid interface, adhesion is much more complicated and depends on elastic and plastic properties of the solids coming into contact, because the adherents deform each other at the region of contact. Particle adhesion is the result of forces which exist between particles and the substrate. Generally, the main cause of adhesion is believed to be the London dispersion force. Excess electric charges generated by frictional electricity can sometimes enhance the rate of soiling by catching dust particles from the air. Soil particles such as soot are not only able to adhere to the substrate by adhesive bonds but they may also be occluded in holes or crevices of the substrate. The main forces involved in the adhesion of soot particles on a carbody surface which is a polymeric surface are capillary, van der Waals, elastic and ionic type.

195

4.1. Capillary force Under humid conditions, a liquid bridge between particle and surface can be formed in two different ways: by spontaneous capillary condensation of vapours and by directly dipping the particle into a wetting film which is present on the substrate, this is the capillary force [18-25]. Due to surface tension, a liquid bridge between the particle and the surface results in a mutual attraction. At thermodynamic equilibrium the meniscus radius represented in Figure 20, is related to the relative vapour pressure by the well-known Kelvin equation: ~+~=~ln ~P r1 r2 2~,M \ Ps J

(5)

where rl and r2, R, T, p, ~,, M and ~P are the meniscus radius, the gas constant, ps the temperature, the density, the surface tension of the liquid, the molecular weight and the relative vapour pressure respectively.

/ ----r2

liquid--

f

Soot particle

rl

t_~

polymeric surface Figure 20. The model of capillary interaction between the soot particle and the polymeric surface which is covered with a liquid adsorbate. The mutual attraction between the particle and the surface results from the Laplace pressure p:

p= ~-~rk where rk is the Kelvin mean radius

(6)

196 rl r2 rk = ~ r 1 + r2

(7)

If the liquid wets perfectly, t h e n the total capillary force exerted on the particle is given by:

F(d) = nx 2(d) ~' rk

(8)

where (x) is the radius of the area. From geometrical considerations one obtains x 2 = 2Rz and for t << R, one has:

z~t-d+r

[ R]

1 1 + ~ R+r 1

(9)

For rl << r2, one can write rl ~ rk, and one has the capillary force given by:

Fc(d) = 2~R~, 1 +

~R R-rk

~t-d 1 rk

(10)

if rk << R a n d t ~ d, the capillary force becomes: : Fc = 4~yR If the surface is not an ideally wetting surface, we m u s t introduce the contact angle b e t w e e n the surface and the liquid: Fc = 4n 7 cos0 R

4.2. L o n d o n - v a n der W a a l s force From observed deviations from the ideal gas law, van der Waals concluded in 1873 t h a t molecules a t t r a c t each other. Only after development of the theory of q u a n t u m mechanics, London could quantify this s t a t e m e n t in 1930. London forces can be u n d e r s t o o d as follows: for a nonpolar atom, the time average of its dipole m o m e n t is zero, but at any i n s t a n t there exists a finite dipole m o m e n t given by the i n s t a n t a n e o u s positions of the electrons. This i n s t a n t a n e o u s dipole generates an electric field t h a t polarises any closely n e u t r a l atom, inducing a dipole m o m e n t in it. The consequence of an interaction is an a t t r a c t i o n force b e t w e e n the two atoms. If one considers an interaction between two condensed bodies, a sphere and flat plate, the interaction energy is given by the following equation:

197

1

(11)

where A is the H a m a k e r constant, R the radius of the sphere and H the separation distance. If R>>H, the interaction energy becomes: AR Va = - ~ 6H

(12)

corresponding to an adhesive force Fa, AR Fa = ~ 6H 2

(13)

The H a m a k e r constant m a y be evaluated from the equation, A = 12 ~H 2 W, where W is the energy per unit area. So the London-van der Waals force will then be given by the expression: Fa=2 r~RW

(14)

4.3. E l a s t i c f o r c e

A flat polymeric surface can be considered as an elastic solid. This m e a n s t h a t the surface has a definite shape and is deformed by external forces into a new equilibrium shape. If the force is removed, the surface reverts to its original form. The solid stores all the energy which it obtains from the external forces during the deformation, and this energy is available to restore the original shape when the forces are removed. The deformation of an elastic surface induced by a rigid sphere is described by Sneddon mechanics [24]. The elastic force is given by a more complex formula:

Fe = 2(1 - v 2)

[. R - n )

where E, v, 1"1and R are respectively Young's modulus, the Poisson coefficient, the radius of the contact area between the sphere and the flat surface and the radius of the sphere. If one a s s u m e s t h a t ~ < 1, the elastic force becomes" R E Fe = ,, fiR 1-v

(16)

198 The surface of contact between the sphere and the flat surface is defined by, = 2Rd, where d is the penetration depth. The elastic force will be given by: Fe =21/2

E R3/2dl/2 1- v2

(17)

A particle which adheres to the polymeric surface is subjected to three forces: two attractive forces, capillary and London-van der Waals forces, and a repulsive force, the elastic force. At equilibrium: Fc+ Fa- Fe = 0 4~/R + 2 ~ R W - 21/2

(18) E R3/2d 1/2 = 0 l_v 2

If we pose A - 4~R, B - 2~W and C -

(19)

21/2 1-Ev 2 ' then by r e a r r a n g i n g the above

equation we can calculate the penetration depth, d:

d = I A +1B ]C 2 R

(20)

The adhesion of a solid dirt particle such as diesel engine soot on the polymeric surface is dependent on the radius of the particle. The smaller is the radius, the larger is the depth of penetration and the larger will be the area of contact. In fact the strength of the adhesive bond is proportional to the contact area. This is the simple theoretical basis which shows that solid soils become more difficult to remove from the surface as their size decreases. This effect has been demonstrated m a n y times and in the case where the substrate is a textile fibber has been attributed to mechanical e n t r a p m e n t of the soil particles in the crevices on the fibber surface [25]. In reality the adhesion between the soot particles and the substrate is much more complex because diesel engine soot is a mixture of solid particles and organic matter. The latter can act as a hydrophobic bond between the polymeric surface and the solid particle.

5.

W E T T I N G OF C A R B O D Y S U R F A C E S BY A D S O R P T I O N OF SURFACTANTS

5.1. D e t e r m i n a t i o n o f t h e c r i t i c a l s u r f a c e t e n s i o n o f a c a r b o d y s u r f a c e The determination of the surface energy of a solid is of interest in fields such as adhesion and detergency, where forces operate across an interface. The

199

physicochemical basis for the wetting of a solid surface, S, by a liquid is as follow: a drop of liquid L, layed down on the solid in contact with air, A, will spread out until it makes a contact angle, 0, defined by the solid surface and the tangent to the liquid-air interface. 0 is measured in the liquid phase. At equilibrium, and neglecting gravity effects, this situation is described by the so-called Young equation [26]: 7LVCOS0

= ysv- 7SL

(21)

where 7LV, ysv and 7SL are respectively the surface tension of the liquid, the surface tension or surface energy of the solid in equilibrium with air and the interfacial surface tension between the solid and the liquid. Two parameters are useful in wetting relationships: the adhesion tension and the work of adhesion. The adhesion tension, i:, for a liquid on a solid is defined as the difference between the surface tension of the solid and the interfacial tension at the solidliquid interface: I: = 7SV- 7SL

(22)

The energy required to separate an unit area of solid surface from the liquid is the work of adhesion, W, [27]. This is given by the following expression: W

-~ 7LV -F 7SV" 7SL

(23)

The lower the solid-liquid interfacial tension, the stronger the adhesion and the lower will be the contact angle between the solid and the liquid. The knowledge of the surface energy of the solid is fundamental to understand the wetting. In the case of low energy solids such as polymers, Zisman [28] characterised the wetting of solids by the measuring contact angles, 0, between solid surfaces and a series of liquids, and plotting cos0 versus 7LV, the liquid surface tension. The point at which the resulting curve intercepts the line cos0 = 1, is called the critical surface tension, 7c. The 7c is the liquid surface tension required to give a contact angle of zero degrees. The more nonpolar the solid surface, the lower the value of Yc which is obtained. We have measured the critical surface tension of the polyurethane polymer which is used as the paint on carbody surfaces. The value obtained is about 45 mJ m 2. Since the liquid surface tension of water is 72 mJ m 2, it is not spontaneously spread over the polymeric surface and forms a contact angle of 62 ~ This situation is unfavourable for the removal of diesel engine soot from carbody surfaces. The effectiveness of carbody washing necessitates t h a t the washing liquid wets perfectly the surface of interest. This is why the addition of surfactants to water, to decrease the interfacial tension of the solid-liquid interface, is often necessary to enable water to wet a solid.

200 5.2. C o n t a c t a n g l e m e a s u r e m e n t s The addition of surfactants to water is a well-established means of enhancing the ability of aqueous solutions to wet and spread over solid surfaces [29,30]. The Wilhelmy plate method was used to measure the variation of contact angles. Figures (21-23), show these variations between the polymeric polyurethane model surface and a series of aqueous solutions of surfactant versus the molality of the surfactants.

1,1 1,0- ; ' 9

1,1 1,0 0,90,8 o

"9

"

0,9"" ,8 " 9

(a)

9

0,7 o

o 0,60,5 0,4 0

I

I

I

I

l

2

3

4

C, mmol kg

5

(b)

0,7-.~ 0,6"' 0,5i 0,40,3 0

I

10

-1

I

I

20

30

C, mmol kg

40

-1

Figure 21. Variation of cos 0 between the polyurethane surface and aqueous solution of SDS (a) and NDB (b) against the molality.

0 0

1,1

1,1 1,0 0,9 9 0,8 g 0,7 g 0,6

1,0"

(c)

00 9

0,90,8-

"

0,5

0,4 0,3

0,40,3

,

5

i

,

10

15

20

(dl

.o

0,7- o CD O,6 P o 0,5:

0

9

,

0

10

,

l

20

30

-l

C, mmol kg

40

-1

C, mmol kg

Figure 22. Variation of cos 0 between the polyurethane surface and aqueous solution of TTAB (c) and DTAB (d) against the molality.

201

1,2

1,2 1,0

. o ~

oo

0,8 gm O o

0,4

g, 0,6 0,4

0,2

0,2 0

0,0

I

I

0,2

0,4

9 t

9

9

O

0,8 Dt

(e)

0,6

0,0

i

1,0

0,6

0

(f)

I

I

I

1

2

3

-I

4

-l

C, mmol kg

C, mmol kg

Figure 23. Variation of cos 0 between the polyurethane surface and aqueous solution of TN 111 (e) and TN 150 (d) against the molality.

The contact angle decreases as the molality of the s u r f a c t a n t increases and becomes equal to zero at a molality below the critical micelle concentration. The molality at which the contact angle is zero depends on the n a t u r e of the surfactant. If we define the effectiveness of a surfactant to decrease the contact angle as the m i n i m u m a m o u n t needed to obtain a zero contact angle, then the nonionic s u r f a c t a n t TN 111 is the most effective, as see in Table 3.

Table 3 Molality and surface tension of aqueous solutions of s u r f a c a t n t at which the contact angle is zero Surfactant TTAB DTAB SDS NDB TNlll TN150

Molality for which the contact angle is zero [mol kg -1] 2.03 7.52 2.00 8.25 3.00 3.14

10 .3 10 .3 10 .3 10 .4 10 .5 10 .4

Surface tension for which the contact angle is zero [mJ m -2] 49.3 48.4 47.2 46.0 40.0 41.7

It has been stated t h a t aqueous solutions of surfactants wet solids by a mechanism in which the surfactant is adsorbed with the hydrocarbon chain in contact with the hydrophobic surface such as polymers. B e r n e t t and Zisman [31] a t t e m p t e d to explain the wetting by aqueous solutions on the same basis as wetting by pure organic liquids. To a first approximation the spreading is then

202 caused by the surface tension of the liquid, and the solid-liquid interfacial tension 7SL, plays only a minor role. The ability of an aqueous solution to wet a low energy surface should therefore depend upon the surface energy of the solid and the molality of the wetting substance dissolved in water to decrease the surface tension of the water below the surface tension of the solid. However, the wetting of the solid surface is more complex. On polar solids, when the solid surface bear an opposite charge to t h a t of the surfactant, aqueous solutions of surfactant induce a dewetting. At low molality the surfactant adsorbs on the surface. The alkyl moiety of the surfactant is oriented towards the solution and renders the surface hydrophobic [32]. We believe t h a t the wetting of the polyurethane surface is mainly caused by an adsorption of surfactants on the surface of the polymer. According to this approach, the solid-liquid interfacial tension, 7SL, between water and the polyurethane surface must be reduced by an addition of surfactant. Indeed, if the wetting of the surface is not caused by an adsorption, the interfacial tension can not change. The lowering of the interfacial tension is therefore related to the extent of adsorption at the solid-liquid interface. The variation of 7SL from the water-polyurethane interface to t h a t between an aqueous solution of surfactant and polyurethane can be calculated as follows: 7Wsv -7wSL : 7wLVCOSWO

(24)

ySsv - 7sSL = 7sLVCOSSO

(25)

where the exponent w refers to water and s to surfactant solution of a given molality. If we assume t h a t ~'sv does not varies, then ~,wsv= 7ss, on can writes ~/wSL + 7wLVCOSWO _-- ~sSL + 7sLVCOSSO

(26)

7wSL - 7sSL = 7sLVCOSSO -TwLVCOSWO

(27)

From this equation, we can evaluate the change in ~SL produced relative to water by an addition of surfactant. These variations are shown in Table 4. The decrease in solid-liquid interfacial tension induced by an addition of surfactants seems to be dependent of the nature of the surfactant. These variations can be related to the unequal adsorption of the different surfactants on the polyurethane surface. The plot of 7LVCOS0versus 7LV, for all the system, shows t h a t 7LvCOS0increases as 7LV decreases. According to the Young equation and assuming t h a t 7sv is unaffected by the surfactant solutions, this means t h a t the interfacial tension 7SL decreases and confirms qualitatively the calculated change of 7wSL - 7SSI seen in Table 4.

203 Table 4 Change in interfacial surface tension ~wSL- ~sSL Surface tension of aqueous solutions of surfactant at which the contact angle is zero [mN.m -1]

Surfactant TTAB DTAB SDS NDB TN111 TN150

7wSL - 7sSL ([N-m -1]

49.3 48.4 47.2 46 40 41.7

15.3 14.3 13.2 12 6 7.70

A convenient method of analysing the relationship between adsorption at the solid-liquid interface and wetting has been developed by Lucassen-Reynders [33]. When the solid surface is a low-energy surface such as polyurethane, the contact angle can be used to determine the surface excess of the surfactant at the solidliquid interface. The combination of the Gibbs adsorption equation applied to the solid-liquid interface with the Young equation, yields:

d (7 Lv c~ dlnC

) .=RTFsL

(28)

Therefore, the slope of the plot of 7LVCOS0 versus In C may provide information about the surface excess of the surfactant at the solid-liquid interface. These excesses were calculated according to the above expression. The values are shown in Table 5.

Table 5 Surface excesses of the surfactant interface Surfactant TTAB DTAB SDS NDB TNl11 TN150

surfactants

at

polyurethane-aqueous

Surface tension of aqueous solutions of surfactant at which the contact angle is zero [mN.m -1] 49.3 48.4 47.2 46 40 41.7

solution

FSL [mol m 2] 1.62 1 1.10 9.57 4.12 3.57

10 .6 10 .6 10 .6 10 .7 10 .7 10 .7

of

204

One can note t h a t the surface excess varies with the n a t u r e of the surfactant. It also varies with the length of the alkyl moiety of the surfactant. From TTAB to DTAB, the surface excess decreases from 1.62 10 .6 to 1.0 10 .6 mol m -2 related to the length of the alkyl moiety of the surfactant. This q u a n t i t y also reduces as the length of the headgroup of the nonionic surfactant increases; FSL is equal to 4.12"10 .7 mol m -2 for TN111 for which the n u m b e r of oxide groups, n = 9, to 3.57.10 .7 mol m 2 for TN150 where n = 15. The a m o u n t adsorbed at the p o l y u r e t h a n e surface seems to be correlated to the change of the solid-liquid interfacial tension, 7wSL - 7sSL. This line of reasoning would indicate t h a t a particular s u r f a c t a n t m a y be a poorer wetting agent for some s u b s t r a t e s t h a n for others, whereas for the same substrate, two different surfactants m a y show different wetting behaviour when the value of the surface tension, 7LV, of the aqueous solutions of these surfactants are the same. This is w h a t we observe in our system: the value of the surface tension 7LV at which the contact angle between the polyurethane surface and the aqueous solutions of surfactants is nil, depends on the n a t u r e of the surfactant (see Table 5).

6.

D E T E R G E N C Y OF CARBODY S U R F A C E

A simple definition of detergency is the removal of particulate soil from the carbody surface by aqueous solutions of surfactants which can alter the adhesion between the soil and the surface. The detergent process on a carbody surface is in principle takes place at an interface. It is therefore f u n d a m e n t a l l y a colloidal phenomenon. The qualitative washing of such a surface with a detersive system is complicated because of the n a t u r e of the paint surface, the soil and the detergents. The detergent composition will depend on both surface and soil. An u n d e r s t a n d i n g of carbody detergency involves several steps: artificial soiling of the paint surface, determination of the a m o u n t of soil on the surface, washing the surface with the surfactant of interest, and determination of the a m o u n t of soil retained on the surface after the washing process.

6.1. S o i l i n g m e t h o d Diesel engine soot was used as an artificial soil on the carbody surface. The comparison between the washing effectiveness of the different s u r f a c t a n t s requires a reproducible soiling method simulating realistic soiling conditions of the carbody in cities. N a t u r a l soiling occurs with the airborne soot particles come in contact with the surface. The diesel engine soot m a y be dry or wet depending on the weather. Artificial soiling with diesel engine soot can then be carried out with the soil a dry or wet state. It is difficult to soil the paint surface with a dry diesel engine soot because the method is not uniform and reproducible. The s u b s t r a t e was metallic disks painted with a polyurethane paint which is used by car manufacturer. We have applied diesel engine soot particles as a suspension in a mixed solvent of water-ethanol mixed solvent, on p o l y u r e t h a n e

205 surface and the system was heated up at 40 ~ C for 5 hours. The soiled surfaces obtained were uniform and reproducible. 6.2. D e t e r m i n a t i o n

o f t h e a m o u n t o f soil o n t h e p o l y u r e t h a n e s u r f a c e The visual cleanliness or soiling of a surface of interest is useful. It is the only method used by the consumer to appreciate the effectiveness of the washing action. Nevertheless this method is subjective and the opinion of two consumers may be different. Visual estimation can not give a quantitative amount of soil per unit area of the surface. Therefore, if we want a valid theoretical description and a comparison of the cleansing action of detergent solutions, a physical method must be used. The estimation of the amount of soil on a polyurethane surface can be evaluated by several physical techniques: chemical analyses, radioactive methods, or spectrophotometry measurements. The latter method has been used because it has the advantage of resembling the visual evaluation of the soiled surface. This appearance is essential for consumers and must be taken into account. Spectrophotometry measurements for the determination of the amount of soil deposited on a polyurethane surface are more precise and sensitive than visual estimation. The effectiveness of detergency, D, of polyurethane surface was estimated using the following expression :

D = Ro_ R___.~•I 100 R0

(29)

where Ro is the reflectance of the soiled surface and R1 the reflectance of the washed surface.

6.3. D e t e r m i n a t i o n

of the washing effectiveness of the surfactants

The washing process was as follows: the polyurethane surface was immersed in aqueous solutions of surfactants at different molalities for 5 minutes, and rinsed for 30 seconds with deionised water. The metallic disks were dried before reflectance measurements. Figures (24-26) show the dependency of the detergency effectiveness upon the nature and molality of the surfactant. For the surfactants studied, the washing effectiveness increases with the molality of the surfactant until the critical micelle concentration (cmc) is reached. Nonionic surfactants T N l l l and TX100, and anionic surfactant SDS are the most efficient, Table 6. The washing effectiveness depends on the length of the alkyl moiety of the surfactant; D decreases from 11.1 for TTAB, to 9 for DTAB. It also depends on the number of oxide groups in the nonionic surfactants; d decreases from 13.4 for TXl00 to 8 for TX165.

206

14

12-

(a)

12-

(b) 10. O

10-

O

8.

o~

oDTAB 9 TTAB

6J

420

0

I

I

I

I

I

2

4

6

8

l0

I

12

0

I

14

0

16

I

I

I

3

6

9

I

I

12

-i

I

15

18

21

-i

C, mmol kg

C, mmol kg

Figure 24. Variation of the detergency effectiveness upon the molality of SDS (a), TTAB and DTAB (b) at 298K.

10.

1412

87-

10

~6.

8

5-

6 4

(d)

9-

(c)

43 .D

9

21

2 0 0,0

9

I

I

0,1

0,2

0 0,3

0

I

I

I

I

2

4

6

8

-1

C, mmol kg

I

10

-1

C, mmol kg

Figure 25. Variation of the detergency effectiveness upon the molality of TN 111 (c) and NDB (d) at 298K.

207 4

.....................................................

12 10

9

8

o

6

o D

4

9 TX100

O

o TX165

o

2 0 0

0,5

1 C, mmol kg

1,5 -1

Figure 26. Variation of the detergency effectiveness upon the molality of TX100 and TX165 at 298K.

Table 6 Washing effectiveness of aqueous solution of surfactants Surfactant SDS TTAB DTAB NDB Tlll TX100 TX165

Washing effectiveness [%] 13.8 11.1 9 8.10 13.7 13.4 8

It follows from this that the washing process is firmly correlated with the nature and the structure of the surfactant. This relation can be explained if we consider the adsorption of these surfactants on diesel engine soot and the wetting of the polyurethane surface.

208

7. A D S O R P T I O N AS A CONDITION OF CARBODY WASHING 7.1. W e t t i n g of the c a r b o d y surface The removal of diesel engine soot from a carbody surface by aqueous solutions of surfactant is a result of two mechanisms. The first one is the wetting of the surface. The tendency of a liquid, L, to spread over a surface, S, is given by the spreading coefficient, Svs SL/S = 7SA - 7SL" 7LA = 7LA (COS0 - 1)

(30)

where the subscripts SA, SL and LA refer to the surface-air, surface-liquid and liquid-air interface. If the spreading coefficient is negative, the liquid does not spread spontaneously over the surface, and mechanical work must be done to wet the surface. In the case of a polyurethane surface, we have shown that the surface is hydrophobic. The contact angle between this surface and water was 62 ~ Consequently, SIJS = -3.84. Since the spreading coefficient is negative, the washing effectiveness of a carbody surface soiled with diesel engine soot is weak. Increasing the efficiency of the washing process involves the adsorption of surfactants on the polyurethane surface, to decrease the contact angle between the aqueous solutions of surfactants and the surface and to make S~s nil. We have shown previously that the wetting of a polyurethane surface occurs by adsorption of surfactants via London dispersion interactions between the hydrophobic moiety of the surfactants and the surface, with the polar headgroups oriented towards the solution. As a result the contact angle, 0, decreases and becomes nil. As result SIJS increases and the washing solutions spread over the surface. This causes a decrease in the adhesive forces and a diminution in the work required to remove the soot particle from the carbide surface. Adsorption of surfactants at the surface solution interface induces an interfacial pressure, ~, cause by the repulsive interactions between the surfactant molecules in the interfacial film. It is this spreading pressure which bring about a diminution of the adhesion between the soil and the surface and facilitates the removal of the soil (Figure 27).

209

Adsorption of surfactants

Soot particle

._ _. _. _. _ . _ . _ ._ _. _ ~ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ . . . . . . . . . . . . . .

/ / / / / / / / /

/

Figure 27. Removal of soot particle caused by the spreading pressure.

7.2. C o r r e l a t i o n b e t w e e n a d s o r p t i o n of s u r f a c t a n t s on d i e s e l e n g i n e soot and the washing process Adsorption of surfactants on diesel engine soot was followed by turbidity measurements. In all cases, we note a maximum dispersion of soot particles and a stabilisation of these dispersions around the cmc region, while it is more difficult to disperse soot particles in pure deionised water. Since the surfactant adsorbs on the soot via hydrophobic interaction between the alkyl chain moiety of the surfactant and the surface with the headgroup of the surfactant oriented toward the bulk phase, the soot particles becomes more hydrophilic. We have shown that electrophoretic mobilty of soot particles is not affected by adsorption of nonionic surfactant, while it is modified by adsorption of cationic, zwitterionic and anionic surfactants. Therefore the dispersion of soot particles can be explained by electrostatic forces in case of ionic surfactants and by steric interaction in the case of nonionic surfactans. The interpenetrating of two adsorbed layers of nonionic surfactants can result in a loss of transformational freedom and so to a loss of entropy and leads to repulsion between the soot particles. A theory including an entropy-repulsive energy term between two particles was developed: V r =

27t0I

(31)

In this equation, 0 is the surface coverage and I is an integral depending on the geometry of the system. From this equation, we conclude that the maximum entropic repulsion between soot particles are obtained when 0 is maximum. Spontaneous deflocculation by nonionic surfactants can be explained on the basis

210 of the spreading pressure exerted by the adsorbed nonionic layer, this leads to penetration over the contact zone between the adherents. The disjoining force Fa exerted on the adherents can be represented by: Fd = 2nrP s

(32)

where r is the radius of the particle and Ps a spreading pressure. The spreading pressure can be calculated by the Gibbs equation: d~,=- .~Fidl~ i 1

This equation is valid not only for liquid-liquid and liquid-gas interfaces, where ~, can be measured, but also for solid-liquid interface. One of consequence is t h a t the spreading pressure, i.e. the difference between the surface tension of the solid in the absence of the adsorbate, ~,o and in its presence, ),, can also be evaluated for a solid-liquid interface, then one can writes: Fi

Ps =~/0 _y = E ~Fidpi i Fi= 0

(33)

For ideal solutions of uncharged molecules, d~ is equal: dpi = RTd In c i

(34)

T h e Fi terms are surface excesses. In dilute solutions, all Fi terms can be referred to Fsolvent, which by definition is set to zero. So, in the case of adsorption of

nonionic surfactants on diesel engine soot, we can calculate the spreading pressure by the following equation: c

Ps = RT~Fdlnc o

(35)

In this equation R is the gas constant, T the absolute temperature, F the a m o u n t adsorbed of the surfactant and c the concentration of the surfactant. Therefore the adsorption of surfactants on the diesel engine soot induces an interfacial pressure caused by the repulsive interaction between the adsorbed surfactant. Then, the removal of diesel engine soot from the polyurethane surface, is facilitated by the interfacial pressure and the repulsion between the surfactant adsorbed on diesel engine soot and the polyurethane surface (Figure 28).

211

Soot particle

--~ ---_-------------.=

__. . . . . . . .

,

Adsorption of surfactants ~

~Q O Q : t : 9 .. 9 71.

"tr

.y.~. ,~.~

~'~

Figure 28. Mechanism of removal of soot particle from the polyurethane surface.

8.

SUMMARY

The quantitative evaluation of the chemical composition of the diesel engine soot material shows that there are between 10 and 15% mineral oxide, 0.20% organic compounds and between 84.8 and 89.8% solid carbon. The specific surface area varies from 52 m2"g-1 for the untreated soot to 70 m2.g1 for the soot washed with two solvents. The distribution of the average diameter also depends on chemical t r e a t m e n t and varies from 19 ~m to 6 ~m. This surface is hydrophilic and hydrophobic and negatively charged in aqueous solution. The results of the surfactant adsorption show a strong adsorption on hydrophilic and hydrophobic parts of the surface. The best wetting surfactant of a model carbody surface is the nonionic surfactant. All these experimental results have allowed to propose a mechanism of washing and to obtain an efficiency detergent formulation for carbody cleaning.

REFERENCES

1. 2.

E. Matijevic (ed.), Surface and Colloid Science, New York, 1972. E. Kissa (ed.), Detergency Theory and Technology, Marcel Dekker, Inc., New York, 1987. 3. W.G. Culter and R. C. Davis (eds.), Detergency Theory and Test Method, part I, Marcel Dekker, Inc., New York, 1972. 4. Surfactants and Interfacial Phenomena, John Wiley, New York, 1972. 5. J . B . Heywood, and McGraw-Hill (eds.), Internal Combustion Engine Fundamentals, New York, 1988.

212 6. D. Schuetzle, T. E. Jensen, D. Nagy, A. Prostak and A. Hochhauser, Anal. Chem., 63 (1993) 1149. 7. S. Partyka, F. Rouquerol and J. Rouquerol, J. Colloid Interface Sci., 68 (1979) 21. 8. H. Malandrini, F. Clauss, S. Partyka and J. M. Douillard, J. Colloid Interface Sci., 194 (1997) 183. 9. M. Lindheimer, E. Keh, S. Zaini and S. Partyka, J. Colloid Interface Sci., 138 (1990) 83. 10. J. M. Douillard, S. Pougnet, B. Faucompre and S. Partyka, J. Colloid Interface Sci., 154 (1992) 113. 11. J. L. Trompette Ph.D. Thesis, University of Montpellier, Montpellier, 1992. 12. J. L. Trompette, J. Zajac, E. Keh and S. Partyka, Langmuir, 10 (1994) 812. 13. Th. F. Tadros (ed.), Solid/Liquid Dispersion, London, 1987. 14. G. D. Parfitt and C. H. Rochester (eds.), Adsorption from Solution at the Solid/Liquid Interface, London, 1983. 15. G. H. Findenegg, B. Pasucha and H. Strunk, Colloids and Surface, 37 (1989) 223. 16. M. S. Celik, J. Colloid Interface Sci., 129 (1989) 428. 17. J. Zajac, C. Chorro, M. Lindheimer and S. Partyka, Langmuir, 13 (1997) 1486. 18. F. M. Orr and L. E. Rivas, J. Fluid Mech., 67 (1987) 723. 19. M. A. Fortes, J. Colloid Interface Sci., 88 (1982) 338. 20. E. A. Boucher, M. J. Evans and S. McGaary, J. Colloid Interface Sci., 89 (1982) 154. 21. D. N. Mazzone, G. I. Tadros and R. Pfeffer, J. Colloid Interface Sci., 113 (1986) 544. 22. H. Wiesendanger and J. Guntherodt (eds.), Scanning Tunneling Microscopy III, New York, 1993. 23. Surfactants and Interfacial Phenomena, John Wiley, Inc., New York, 1972. 24. M. Heuberger, D. Giovanni and L. Schlapbach, J. Vac. Sci. Technol. B, 14 (1996) 1250. 25. W. G. Culter and R. C. Davis (eds.), Detergency Theory and Test Method, part II, Marcel Dekker, Inc., New York, 1972. 26. T. Young, Philos. Trans. R. Soc. 95 (1805) 65. 27. A. Dupre, GautierVillars (eds.), Th~orie M~canique de la Chaleur, Paris, 1869. 28. W. A. Zisman, Advan. Chem. Ser., 43 (1964) 1. 29. R.A. Pyter, G. Zografi and P. Mukerjee, J. Colloid Interface Sci., 89 (1982) 144. 30. C. Gau and G. Zografi, J. Colloid Interface Sci., 140 (1990) 1. 31. M. K. Bernett and W. A. Zisman, J. Phys. Chem. 63 (1959) 1241. 32. T. Minassian-Saraga (ed.), Contact Angle, Wettability and Adhesion, Advances in Chemistry Series, 43, Washington D. C., 1964. 33. E. H. Lucassen, J. Phys. Chem., 67 (1963) 969.

Adsorption and its Applicationsin Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski(Editor) 9 1998ElsevierScienceB.V. All rightsreserved.

213

T h e o r y a n d a p p l i c a t i o n of p r e s s u r e s w i n g a d s o r p t i o n for t h e environment Y. Liu, D. Subramanian and J.A. Ritter Department of Chemical Engineering, Swearingen Engineering Center University of South Carolina, Columbia, SC 29208, USA Analytic, equilibrium theory-based expressions for the periodic state process performance and bed profiles of three evolving environmental pressure swing adsorption (PSA) processes are presented. The three PSA processes include single component solvent vapor recovery, single component air purification with complete clean-up, and binary solvent vapor recovery. The analytic expressions are used to carry out conceptual designs of several environmentally related and commercially relevant PSA systems, including dimethyl methylphosphonate (DMMP) and butane vapor recovery, purification of air containing ppm levels of styrene vapor using a complete clean-up cycle, and separation of n-butane vapor from a mixture of n-butane and n-heptane in an inert carrier gas. These conceptual designs show that the analytic expressions can be readily used in the preliminary design, feasibility and performance evaluation of environmental PSA processes, even by the novice.

1. I NTR ODUCT I O N Volatile organic compound (VOC) emissions such as benzene, butane, acetone, trichloroethylene, carbon tetrachloride, and styrene have been under increased scrutiny worldwide [1], with the only likely scenario being even more stringent requirements on the release of VOCs into the environment. Thus, an increasing emphasis has been placed on the development and use of air purification (AP) and solvent vapor recovery (SVR) processes. Similarly, very strict exposure limits have been placed on contaminants used in defense applications (i.e., chemical agents) because of their extreme toxicity. Moreover, experiences gained during the Gulf War have heightened the awareness of the need for better AP systems designed specifically for defense applications. The requirements for defense systems differ, however, from those needed for controlling vapor emissions. In contrast to SVR processes, which have to simultaneously purify air and recover solvent vapors, defense systems only have to purify air. Nevertheless, both applications are inherently environmentally related, where adsorption technology has been used quite extensively and successfully.

214 Carbon adsorption with steam regeneration has been one of the most frequently used methods for removing and recovering solvent vapors from air [2-8]. However, this method suffers from thermal aging of the adsorbent, equipment corrosion, and inefficiency in energy usage [9]. Clearly, carbon adsorption with steam or hot purge gas regeneration is also not very practical for defense applications because of the large bed sizes that are required as a result of the long cycle times. In this later application, once-through systems have been used with subsequent disposal of the carbon. Pressure swing adsorption (PSA) offers attractive alternatives for both SVR in industrial applications [10,11] and AP in defense applications [12]. The essential features of PSA have been delineated in the pioneering patents of Hasche and Dargan [13], Perlet [14], and Finlayson and Sharp [15], as well as in the papers of Kahle [16,17] and Skarstrom [18]; it has also been extensively discussed in the recent monographs by Ruthven [10], Yang [19], and Ruthven et al. [20]. During the 1980s, PSA gained widespread commercial acceptance [20]; and the growth in research and development, and commercialization of PSA has been rather spectacular in the last two decades. The key reason for this outstanding progress is that PSA technology can provide a very flexible and efficient means of gas separation and purification, and, for many applications, it can reduce the energy and cost of separation compared to conventional separation processes like absorption and distillation. Nowadays, PSA processes are widely used on a very large scale for a variety of gas separations [21]. In contrast, environmental and defense applications of PSA represent two relative new areas with major potential for growth. For example, the recovery of small amounts of organics from chemical processes, storage-tanks and other gaseous vents, as well as from solvent painting, purging and cleaning operations, are increasing [22,23]; and so are AP needs in defense applications [24]. Specific environmental issues related to adsorption science and technology where PSA processes have either recently been commercialized or shown some promise for commercialization include, local environmental problems, such as SVR [1,6,7,10,11,25], solvent vapor fractionation, and SOx and NOx removal from flue gas [26], and global environmental problems, such as emission control of green-house gases (CO2, CH4, N20, etc.) [27,28], recovery of CFCs in emission control of ozone depletion gases [29], and contaminant removal in defense applications [12]. The objective of this chapter is to introduce some simple analytic expressions derived from equilibrium theory that can be used for environmental PSA process heuristics, feasibility, design, development, performance, and understanding. Complete sets of expressions are given for three evolving environmental PSA processes: single component SVR from an inert carrier gas (e.g., nitrogen or air); single component AP with complete clean-up cycles, and binary SVR from an inert carrier gas, where the lighter vapor is separated from the heavier vapor. Conceptual process designs are carried out for all three cases to illustrate the use of these simple expressions. It is noted that only those

215 processes having feed streams containing one or two components in an inert carrier, with the total contaminant mole fraction being less t h a n about 15%, are treated; this is typical of environmental applications. It is also noted that the relative humidities of the gas streams are not taken into account because of the associated complexities, which supersede the use of a simple equilibrium model. Nevertheless, the general concepts and ideas conveyed here can be extended to more complex PSA processes including bulk gas separation.

2. FUNDAMENTALS OF ENVIRONMENTAL PSA PROCESSES 2.1. Principles The principle underlying PSA technology is the selective adsorption of one or more components from a feed gas mixture on a solid adsorbent, so that an adsorbed phase having a composition different from that of the feed mixture is formed when the feed is contacted with the adsorbent. The gas phase becomes richer in the less selectively adsorbed components during the adsorption step and in the more selectively adsorbed components during the desorption step. The attractive forces responsible for this adsorption phenomenon are of the van der Waals type. Thus, the adsorbed components are easily desorbed by reducing their partial pressures. The desorption process also cleans the adsorbent so that it can be reused in subsequent cycles. The partial pressures of the components can be altered by decreasing the total pressure and/or by changing the composition of the gaseous mixture. A major advantage of PSA, relative to other types of adsorption processes, such as thermal swing, is that the pressure can be changed much more rapidly t h a n the temperature. This makes it possible to operate a PSA process on a much faster cycle, thereby increasing the throughput. The adsorption capacity available for separation in PSA depends on both equilibrium and kinetic factors, but the relative importance of these factors varies greatly for different systems. The majority of PSA processes are "equilibrium driven" in the sense that the selectivity depends on differences in the equilibrium affinities. This is true also of PSA processes for the environment. In such processes mass transfer resistance generally has only a slight but nevertheless deleterious effect and reduces the performance relative to ideal (equilibrium) systems, which are the type treated here. 2.2. Basic s c h e m e s Different from PSA bulk gas separation processes, which typically utilize multiple beds and additional steps to increase the light product recovery (such as pressure equalization and co-current blowdown steps), environmental PSA processes utilize a twin bed system with a Skarstrom-type cycle. During each cycle, two beds each undergo four steps, namely, adsorption, countercurrent blowdown, countercurrent purge, and repressurization. The purge gas can come from the light product of the other bed or from ambient air, and the

216

pressurization can be realized by using the feed mixture (cocurrently) or again by using ambient air (countercurrently). While one bed is undergoing adsorption the other bed is being purged, and while one bed is undergoing repressurization the other bed is being depressurized. In this way, the beds operate 180 ~ out of phase with each other. Other cycle designs are also possible, whereby feeding is continued in one bed while the other bed undergoes blowdown, purge and pressurization. A typical cycle sequence is depicted in Fig. 1, and it may be carried out as follows. During the adsorption step, the gas mixture is fed into the bed at a constant high pressure and the less selectively adsorbed component(s) is withdrawn as the light product (or vented as clean air). During the blowdown step, the bed is depressurized from the high pressure to the low pressure by withdrawing gas through the feed end of the bed (countercurrently). The light product end of the bed is kept closed during this step. During the purge step, the depressurized bed is countercurrently purged. The gas enriched in the more selectively adsorbed component(s) is withdrawn through the feed end of the bed during both the blowdown and purge steps. During the repressurization step, the bed is pressurized from the low pressure to the high pressure with the feed gas mixture through the feed end of the bed (cocurrently) while keeping the other end of the bed closed. Once the bed reaches the high pressure, feeding commences to begin a new adsorption step and thus cycle. Eventually, these coupled beds reach a periodic state, which is also commonly referred to as the cyclic steady-state. 2.3. P e r f o r m a n c e i n d i c a t o r s

The performance of an environmental PSA-AP process is evaluated mainly by the light product purity (yp) and the process throughput (0p), which is defined as the volume of feed mixture processed per unit mass of adsorbent per unit time. For the PSA-SVR process, in addition to yp and 0p, the process performance is also judged by the solvent vapor recovery (~) and enrichment (E), and the bed capacity factor (BCF) (for single component only), which is akin to the length of utilized bed [25]. is defined as the ratio of the number of moles of the solvent vapor leaving the bed during steps III and IV to the number of moles of the solvent vapor entering the bed during step II. E is defined as the average mole fraction of the solvent vapor leaving the bed during steps III and IV divided by the mole fraction of the solvent vapor in the feed. yp is defined as the average mole fraction of the solvent vapor exiting the light product end of the bed during step II. The BCF is defined as [25] Lb

BCF = f qdz/qfL b 0

(1)

217 ~ Light Product

]3

Effluent

Feed

Feed

Blowdown

V Column A

II

Light Produc~ Pressumzation

Column B

III

Blowdown

~ Purge

IV

Effluent

Effluent

T

III

Pressurization Purge _~ Light Product I

Feed

Figure 1. Schematic of an environmental PSA process and sequence of steps in a cycle.

and represents the capacity of the bed that is used at the periodic state (measured at the end of step II) compared to the maximum capacity of the bed at

218 the feed conditions. Thus, at a fixed process throughput and when there is no solvent vapor breaking through the bed during step II, a larger BCF indicates a poorer performance. Within the realm of equilibrium theory (see the next section), Eq. 1 reduces to BCF = za Lb

(2)

where Za is simply the bed length that is covered by the concentration shock front at the end of the adsorption step. It is also convenient to define a feed (adsorption) step throughput (Of) as the volume of feed processed per unit mass of adsorbent per unit time during the adsorption step, i.e.,

Of=

Vf = v---k--f PbAbLb PbLb

(3)

The relationship between Of and 0p is given by tc

Of = ~-f 0p

0

(4)

E Q U I L I B R I U M THEORY A N D E N V I R O N M E N T A L P S A P R O C E S S DESIGN

Mathematical modelling has been widely used as a very powerful tool in the theoretical study of PSA in order to gain a clearer understanding of this rather complex process. It can also be used to predict the process performance under various operating conditions, which may save time and cost associated with pilotscale testing. A wide variety of mathematical models have been developed [20], and these can be essentially classified into two groups, namely, equilibrium and dynamic models. Dynamic simulation involves tracking the transient by repeated numerical integration of the governing equations until the periodic state is reached. It is generally quite flexible and very accurate, but difficult to implement for the novice user. For this type of modeling, interested readers are referred to Liu and Ritter [25, 30-32] for SVR processes, and LeVan, Ruthven, and Yang and their co-workers for purification processes [33-36]. In contrast, equilibrium theory is the simplest approach to modeling PSA processes, as it accounts mainly for mass conservation, and ignores transport phenomena; and in many cases, it allows the governing material balance equations to be solved analytically by the method of characteristics.

219 The usual assumptions of equilibrium theory include: isothermal operation, no axial dispersion, no axial pressure gradients, and most importantly, instantaneous local equilibrium implying negligible transport phenomena. Building on the pioneering works of Shendalman and Mithchell [37], Chan et al. [38], LeVan [39], and Pigorini and LeVan [40], Ritter and co-workers [12,41,42] developed some simple analytic expressions for PSA-SVR and PSA-AP processes that give directly the periodic state process performance in terms of the process and adsorption isotherm parameters. Additional assumptions used in the development of these expressions include neglect of the velocity changes in the column due to a non-adsorbing carrier gas and the low feed mole fraction of the adsorbing impurity, neglect of the gas phase capacity due to high partition ratios between the adsorbed and gas phases, and neglect of the pressure transient steps as they occupy only a small fraction of the total cycle time. This later assumption freezes the concentration profiles in the gas and solid phases during the pressure changing steps [43], which reduces the analysis to a two-step PSA process, composed only of feed and purge steps [39,41]. Furthermore, purging of the columns is carried out countercurrently using pure inert carrier gas emanating from an adjacent bed during the high pressure adsorption step. The analytic expressions and a conceptual process design for single component SVR from inert carrier gas are presented first; it is noteworthy that these simple expressions have been justified by comparison with a more rigorous mathematical model under limited conditions [41]. This development is followed by the formulations and a conceptual process design for complete clean-up during PSA-AP. Finally, expressions and a conceptual process design are given for binary PSA-SVR from an inert carrier gas with the lighter vapor being separated from the heavier vapor. In all cases, the designs are directly related to environmental applications of PSA. Further details of the developments of these expressions are given elsewhere [12,41,42]. It is noted at the outset that all of the analytic expressions are derived based on the general Langmuir adsorption isotherm, which is given by qi =

qs,ibiPYi 1 + Z bjPyj J

(5)

for any number of components, where i represents the component and j is summed over all components. Eq. 5 has a one-to-one correspondence with the constant separation factor isotherm,

.

qi =

ci

, R i + (1 - R i)c i through the constant separation factor,

(6)

220

1 Ri = 1 + biPHYf, i

(7)

For favorable L a n g m u i r adsorption isotherms, Ri is restricted to 0 < Ri < 1.0. The dimensionless variables in Eq. 6 are defined as * qi qi = ~ qf , ci ci = w cf

(8) (9)

3.1. S i n g l e c o m p o n e n t PSA-SVR p r o c e s s The PSA-SVR process has received increasing attention since its recent commercialization [1,10,11,25,30-32,39,41-46]. Potential markets for PSA-SVR are very large, as the use of organic solvents is ubiquitous throughout m a n y industries [5]. PSA-SVR processes also offer many advantages over conventional SVR processes by [10,11,47] (a) providing a greater portion of the total working capacity by pulling a vacuum, thereby resulting in more solvent vapor adsorbed for the same amount of adsorbent, and an improved process performance; (b) avoiding high temperatures compared to t e m p e r a t u r e swing adsorption (TSA), thus eliminating or reducing thermal aging of the adsorbent and the possibility of adsorbent combustion in the presence of flammable hydrocarbons and hot air; (c) avoiding the hot steam purge, thus eliminating the secondary waste stream in the recovered condensate (i.e., condensed steam s a t u r a t e d with minute levels of the recovered solvent); (d) avoiding extremely low t e m p e r a t u r e s compared to cryogenic condensation; and (e) providing higher separation factors and lower energy costs compared to distillation. 3.1.1. PSA-SVR p r o c e s s w i t h o u t b r e a k t h r o u g h When no heavy component b r e a k t h r o u g h occurs during the adsorption step in the PSA-SVR process, yp = 0 and ~ = 100%. Therefore, the process performance is judged only by the BCF and E. The dimensionless periodic state bed penetration is expressed as (I - R)(7 v - I) +7 V- 2~(I - R)(7 v - I)7 v

~a=

(I0)

R

where g is the dimensionless axial coordinate defined as Z -

--

q-L

(Ii)

221 L is the distance penetrated by the concentration wave in the very first feed step; it is defined as L =

vftf

(12)

The subscript on ~- in Eq. 10 indicates a periodic state variable; thus ~a = ~Za =

(BCF)Lb L

L

(13)

Introducing Eqs. 2, 3 and 12 into Eq. 13 gives the following expression

qa =

\cf 0ftf

(14)

Also note t h a t Tv and ~/M a r e related by TV =TM(z

(15)

To obtain the enrichment, first, the concentration of the heavy component exiting the column during the purge step is obtained as a function of time from !

, 7VRt-~TVtqatfR ci = 7Vt(R- 1)

(16)

Eq. 16 is then averaged over the duration of the purge step to give E as (Z

E =~ v T

(17)

It is pointed out t h a t the frozen solid phase assumption applied to the pressure changing steps in the development of the equilibrium model limits the pressure ratio (a) that can be employed for a given feed mole fraction, or vice versa [41] to y~nax <_1

(18)

222

The periodic adsorbed phase concentration profile at the end of the adsorption step is simply a shock wave that covers the bed from z = 0 to z = Za, or in dimensionless terms from ~-= 0 to ~- = ~-a. The adsorbed phase concentration profile at the end of the purge step is an expansive wave (heel). This heel extends from the bed entrance to a certain position G = Go, where ~o is the dimensionless axial position reached by the dimensionless concentration ci* = 0 (and thus qi* = 0) when it travels countercurrently. The adsorbed phase concentration profile at the end of the purge step is obtained from / *

g - ga +

qi =

~]TVR(ga - g)

( 1 - R ) ( g - g a)

(19)

and ~-0 is given by g0 = ga -

R7 v

(20)

The corresponding gas phase concentration profile is obtained from Eq. 12 through the adsorption isotherm, Eq. 5. In the design of a PSA-SVR system, the adsorbent, adsorbate, and the corresponding adsorption isotherm are usually known. Also, designers usually have information on the feed, such as the feed mole fraction (yf), feed pressure (PH) and, sometimes, the feed volumetric flow rate (Vf). There are two types of design problems. One is to determine the operating conditions for existing columns and desired process performance; and the other is to design the columns for pre-determined operating conditions and desired process performance. In the first type of design (i.e., known yf, PH and bed dimensions), if Vf and the cycle time (tc) are selected, the purge to feed ratio (7v and 7M), a and thus PL are determined for the desired process performance (E and BCF) as follows. First, qf is calculated according to the adsorption isotherm, Eq. 5, and cf is calculated from yf, Tf and PH in terms of the ideal gas law. Then the superficial feed velocity (vd is calculated from Vf and the bed diameter (db). With the calculated qf, cf and vf, and the specified feed duration (tO, L is calculated from Eq. 12. Note t h a t for this two-step PSA process, tf = 0.5tc. After L is obtained, ~-a is determined from Eq. 13 corresponding to the desired BCF. Now, 7v is obtained from Eq. 10 with R obtained from Eq. 7. a is then calculated according Eq. 17 from the specified E and calculated 7v. PL is simply obtained by dividing PH by (z, and then 7M is obtained from Eq. 15. If a or 7v (or 7M) is specified instead of Vf or tf, then the other operating conditions are obtained as follows. If a is known, 7v is calculated from Eq. 17, and then ~'a is obtained from Eq. 10. From Eq. 13 and with the calculated ~'a,

223

specified BCF and known Lb, L is determined. So, vf (and therefore Vf) is obtained for a fixed tf, or tf (and thus tc) is obtained for a fixed vf from Eq. 12. Equilibrium theory necessarily shows that it is the amount of the adsorbent (i.e., the volume of the bed), not the a r r a n g e m e n t of the adsorbent (i.e., the length to diameter ratio), that plays the role in achieving the desired process performance [41]. Therefore, to obtain the bed dimensions in the second type of design problem, the analytic expressions are used to find the adsorbent inventory (bed volume) for selected operating conditions and desired process performance. After obtaining the bed volume, complete specification of the bed dimensions requires the bed diameter or bed length to be selected a priori. For example, with known Vf and tf (tc), if (z is known, ~,v is calculated from Eq. 17 for the desired E, and then ~a is obtained from Eq. 10 with R obtained in the same m a n n e r as in the first type of design problem. Using Eq. 14, Of is solved for the desired BCF. Finally, Eq. 3 is used to determine the bed dimensions with either db or Lb specified. The first type of design is carried out for the recovery of dimethyl methylphosphonate (DMMP) vapor from an inert carrier gas using BPL activated carbon. The adsorption isotherm of the DMMP-BPL activated carbon system at 298.2 K is taken from Ritter [48]; it is plotted in Fig. 2 along with the Langmuir model correlation. The isotherm parameters and the average relative error (ARE) defined by

5.0 4.5 4.0

........

I

........

I

........

I

........

I

........

I

........

I F

3.5 __ 3.0 (33

-~ 2.5 0

E 2.0

_5

o- 1.5 1.0 0.5

0.0

L _ J -

l x 1 0 -6

,

,,,,,,I

ental data model correlation ,

l x 1 0 -5

,

,,,,,,I

,

l x 1 0 -4

, ,,,,,,I

,

l x 1 0 -3

, ,,,,,,I

,

l x 1 0 -2

, ,,,,,,I

,

l x 1 0 -1

,,,,,,,

lx10 ~

P (kPa) Figure 2. Equilibrium adsorption isotherm of DMMP vapor on BPL activated carbon at 298.2 K [48]" experimental data and Langmuir model correlation.

224 1.5

m m

I

'

I

'

I

'

I

'

I

end of purge step end of feed step

'

[

I

0.0

I

'

I

'

I

,i

edde

I I I

-

0.0

'

0.010 0.005 0.000

1.0

0.5

I

0.1

,

I

0.2

,

I

,

I

0.3 0.4

,

I

,

I ...... I

!

I

0.5 0.6 0.7 0.8

,

,,

0.9

,,,

1.0

Figure 3. Periodic adsorbed phase concentration profiles for the DMMP-BPL system at the end of the feed and purge steps.

ARE%= !00 - qi=1 ~ acb sa/ q le x'p 'qexp,i i / N

(21)

are given in Table 1. This design problem involves determining Vf and 7v (or 7 M) that are required to obtain the specified process performance of E = 10 and BCF = 0.6. The feed conditions are PH = 121.56 kPa and yf = 300 ppm. (z is chosen to be 15 ancl tc is fixed at 1,440 s ( t f - 720 s). All of the design information including the bed dimensions and adsorbent properties are tabulated in Table 1, along with the results of the design. For the desired process performance, 1124.3 SLPM of feed are processed, using yv = 1.5 (~/M _-- 0.1) and PL = 8.1 kPa. The periodic adsorbed phase concentration profiles at the end of the adsorption and purge steps are displayed in Fig. 3. The resulting adsorbed phase profiles do not change much between the end of the adsorption and purge steps during the periodic state because of the strong affinity between DMMP and BPL activated carbon. The insert in Fig. 3 shows that the simple wave at the end of the purge step starts from almost the same position as where the shock wave ended at the end of feed step. This behavior is quite typical for PSA systems with strong adsorbate-adsorbent interactions [25,43]; nevertheless, PSA is still quite effective at concentrating DMMP from 300 to 3000 ppm while utilizing 60% of the column.

225 Table 1 Conceptual design of a PSA-DMMP-SVR process without breakthrough during the feed step Design Input Adsorption isotherm parameters b (kPa-~) 8732.1 q~ (mol/kg) 3.83 ARE (%) 7.51 Bed information Lb (m) db (m) pb (kg/m 3)

0.5 0.2 431.6

Feed conditions yf (ppm) PH (kPa) Tf(K)

300 121.56 298.15

Process conditions (-) tf (s)

15 720

Design Output Operating conditions Vf (m3/s) 0.1704 (SLPM) 1124.3 7 v (-) 1.5 ~M(_) o.1 PL (kPa) 8.1

Required process performance E (-) 10 BCF (-) 0.6

3.1.2. P S A - S V R p r o c e s s w i t h b r e a k t h r o u g h

In the case of heavy component breakthrough, the periodic state bed penetration is always equal to the bed length (i.e., BCF = 1.0). Therefore, the PSA process performance is judged by yp, ~ and E. The extension of the equilibrium theory to allow for the case of breakthrough is restricted to the case of a pure carrier gas purge. Under the simplifications of the theory, breakthrough occurs when the shock t h a t forms and propagates in the feed step breaks out of the bed before the termination of the feed step. In other words, the bed capacity is not sufficient to contain the shock wave in the bed. It is mathematically simulated when the chosen process parameters determine a periodic state penetration greater than the chosen bed length. In such a case, a completely saturated bed results at the end of the feed step, with some of the heavy component vapor being lost in the breakthrough. The procedure for obtaining the heavy component enrichment is quite similar to that used in the case without breakthrough [41]; the final expression for E is

226

II

I Lb~

2

LbP bqfR _ R 7 v + - vftfcf vftfcf

E =a

(22a) 7 v ( 1 - R)

or

2

qfR 0ftfcf

E =a

R 7v +

qf 0ftfcf (22b)

7v(1-R)

To obtain 9~, the change in the adsorbed phase concentration over the purge step duration is calculated through a mass balance, where the final expression is w r i t t e n as

N = E7 v = a

2 7 V L b P b q f R _ R 7v vftfcf (l-R)

+ vftfcf

(23a)

or

2 9~ = ETv = a

qfR 0ftfcf

/

R 7v +

qf 0ftfcf

/

(23b)

(l-R)

To obtain yp, the solvent vapor concentration exiting the bed is required as a function of time over the feed step duration. The time t a k e n for the shock wave to form and propagate to the column exit is the time for which pure inert e m a n a t e s from the column exit. The rest of the feed step duration pollutes the inert light product at the feed concentration level. The average light product purity is given by

(24) -

(l-R)

The periodic adsorbed phase concentration profile is obtained in the same m a n n e r as in the case without breakthrough. However, in this case, the shock wave at the end of the adsorption step covers the whole bed from ~-= 0 to ~-= ~'b.

227 The adsorbed phase concentration profile at the end of the purge step is obtained from Eqs. 19 and 20, except t h a t ~'a in these equations is replaced by ~-b for the case with breakthrough. The design methodology is also similar to the no b r e a k t h r o u g h case. Note t h a t since the BCF is always equal to unity in the case with breakthrough, it is no longer a performance indicator, yp is the important performance indicator in environmental applications; and so are 9~ and E if the solvent vapor is recovered. In the first type of design (i.e., known yf, PH and bed dimensions), all of the operating variables (Vf, tc, a, PL and ~/v) are determinable for a specified performance (E, ~ or yp), when two of them are specified. In practice, however, 9~ and yp are not pre-determined simultaneously since they are related. In environmental applications, yp is usually specified to meet the environmental regulations, leaving ~ to be calculated. To determine these variables, Eqs. 22 to 24 are solved simultaneously for any two of the operating variables, and either or yp by noting the governing relations in Eq. 3 and Eqs. 13 to 15. For example, if Vf and tf are specified, Gb is calculated from Eq. 11 in terms of Lb. Then R and L, as well as qf and cf, are calculated in the same way as in the case without breakthrough. For the specified yp, ~,v is obtained by solving Eq. 24, and then a is obtained by solving Eq. 22 for the desired E. 9~ is calculated from Eq. 23. Specifying the bed dimensions in the second type of design also requires the simultaneous solution of the aforementioned equations to find Of. The bed dimensions are determined through Eq. 3. In this case, at least three process conditions (Vf, tf, a and ~,v) must be specified, with Vf as one of them to completely define the bed dimensions. For example, for a known feed condition (PH and Y0, to achieve the desired process performance (E and yp) for selected Vf, yv and tf, the bed dimensions are determined as follows. First, Eq. 24 is solved to obtain ~-b and then Of is calculated from Eq. 14 with Gb replacing ~'a and BCF = 1.0. The calculated Of is then used to determine the bed dimensions according to Eq. 3 with either Lb o r db specified. The calculated Of and ~v are used in Eq. 22b to obtain (z for the desired E. A second type of design is carried out for the recovery of b u t a n e vapor from an inert carrier gas using Westvaco's BAX activated carbon. The adsorption isotherm of the butane-BAX system at 298.2 K [49] is plotted in Fig. 4 along with the L a n g m u i r model correlation; the isotherm p a r a m e t e r s and ARE are given in Table 2. The design is carried out in exactly the same m a n n e r as in the case without breakthrough. The specific value of each p a r a m e t e r is given in Table 2 along with the design outputs. In this case, two beds with different dimensions are obtained for the same operating conditions (both specified and calculated conditions) and the same process performance, as a result of solving Eq. 24 for yv. With respect to the overall mass balance constraints and viable process and operating conditions, both bed designs are feasible. However, the periodic

228

7.0

........

6.0

E

.......

|

9 ----

5.0 m 0

I

.......

I

.......

I

.......

I

.......

I

.......

I

experimental data modelcorrelation

/

m..,..,,~,.J 9 -

, .... ,,I

'

' ' ' "';

4.0 3.0

(3"

2.0 1.0 0.0

........ I

....... J

,,,,,I

, ,,,,,.I

, ,,,,,,

'o

'o

'o

b

o

o

o

o

o

X

X

X

X

X

X

X

X

X

P (kPa) Figure 4. Equilibrium adsorption isotherm of n-butane vapor on BAX activated carbon at 298.2 K [49]" experimental data and Langmuir model correlation. Table 2 Conceptual design of a PSA-n-butane-SVR process with breakthrough during the feed step Design Input

Design Output

Adsorption isotherm parameters b (kPa -1) 0.1514 qs (mol/kg) 6.5194 ARE (%) 14.62

Operating conditions (z (-) 1.5 PL (kPa) 16.21 TM (-) 0.2

Bed information pb (kg/m ~)

Process performance 517.97

Feed conditions yf(-) PH (kPa) Tf (K)

0.15 121.56 298.15

Process conditions Vf (SLPM) ~v (.) tf (s)

4000.0 1.5 720

Required process performance E (-) 5.0 yp (ppm) 10.0

(%)

Bed dimensions Lb (m) (specified) db (m)

99.993

1.5 0.398 or 1.177

229 adsorbed phase concentration profiles shown in Fig. 5 show that only the smaller bed is physically realistic. 1.5

'

I

'

I

'

I

'

I

'

I

'

I

'

I

'

I

'

I

'

(a) 1.0 realistic behavior 0.5

0.0 1.5

l

,

I

,

'

I

'

I

I

,

I

'

I

,

'

I

I

,

'

I

,

I

,

"Ix

,

I

,

I

,

I

'

I

'

I

'

I

'

I

'

unrealistic beh_~_avior 1.0

-'-

-"-

"-"-

(b)

"-

~

....

\ 0.5

-

~

-

\ 0.0

, 0.0

I 0.1

,

I 0.2

,

I 0.3

,

I 0.4

,

I 0.5

,

I 0.6

,

I 0.7

,

I 0.8

,

I 0.9

,/ 1.0

z/L b Figure 5. Periodic adsorbed phase concentration profiles for the n-butane-BAX system at the end of the feed (solid lines) and purge (dashes) steps for the (a) smaller and (b) larger bed designs.

The periodic adsorbed phase concentration profiles at the end of the purge step are quite different in each case. The profile for the larger bed (Fig. 5b) shows that the adsorbed phase concentration in a certain mass transfer region exceeds that corresponding to the feed condition, which is physically impossible but occurs mathematically to satisfy the mass balance condition. The profile for the smaller bed (Fig. 5a) does not exhibit this unusual behavior; in fact, its behavior is characteristic of that obtained with rigorous numerical modeling [25]. Thus, the smaller bed corresponds to the only realistic design. It is also interesting to compare the adsorbed phase concentration profile in Fig. 5a for butane with that

230 in Fig. 3 for DMMP. The profiles are quite different, with butane exhibiting significant movement of the simple wave during the purge step. This is indicative of a large amount of butane adsorbing and desorbing during each periodic cycle, and it is a direct reflection of the adsorbent/adsorbate affinity.

3.2. Single component PSA-AP process with complete clean-up PSA-AP is a well established process. More than thirty industrial companies in the United States design and manufacture PSA-AP systems [51]. In fact, there has been a great deal of interest recently in the design and development of PSAAP systems [34,35,44,45,51,52], especially for use in chemical defense systems where specific concerns have been raised pertaining to the residual contaminants that remain in PSA-AP beds [25]. This so called "heel" within the beds is characteristic of PSA processes, since PSA beds are never meant to be completely regenerated. What this means is that PSA-AP systems used in defense applications continuously desorb contaminant vapors from the beds even after the contaminant is no longer being exposed to them. This makes it rather difficult for a military vehicle to return to the base after being exposed to the contaminant vapor in the field, because toxic vapors are continuously desorbed from the AP system as long as it is operating; and this desorption can persist for some time. Thus, the objective here is to introduce new vacuum swing adsorption (VSA) cycles for AP with complete regeneration of the beds during every cycle [12]. The conceptual design of a PSA-AP system involves the same procedure as that used for the PSA-SVR processes, both with and without breakthrough. However, the expressions derived by Ritter et al. [12] for complete clean-up only apply to PSA cycles without breakthrough during the feed step; so 9{ = 100%, yp = 0 and the periodic state process performance expressions depict in Eqs. 7 to 17 are valid. Based on these expressions, the critical volumetric purge to feed ratio required for complete clean-up is given by v

--I

(25)

R By noting Eq. 15, Eq. 25 is readily adapted to yield the critical pressure ratio (ac) in terms of yM as

ac =

1

~,MR

(26)

Also, according to the definition of R (Eq. 7), Eq. 26 becomes

(Zc =

(1 + bPHyf) M Y

(27)

231 It is again pointed out t h a t the frozen solid phase assumption applied to the pressure t r a n s i e n t steps in the equilibrium model constrains the pressure ratio t h a t can be employed for a given feed mole fraction or vice versa [41]. For the case of complete clean-up, this feature limits the applicability of the analytic expressions to feed mole fractions constrained by y~nax<~/l_ + 4bPH~/M - 1 2bP H

(28)

The case of incomplete clean-up is governed by Eq. 18. As in the case when there is no b r e a k t h r o u g h during the adsorption step, the periodic adsorbed phase concentration profile at the end of the adsorption step is a shock wave which covers the bed from ~-= 0 to G= ~'a. Due to the nature of complete clean-up, however, the contaminant is completely removed from the feed end of the bed at the end of the purge step. Therefore, no contaminant heel is left in the bed at the start of a new cycle. So, the adsorbed phase concentration profiles at the end of the feed and purge steps are obvious. To design a PSA-AP process with complete clean-up, once the feed condition is determined (PH and Y0 for a known isotherm system, 7cv required for complete clean-up is fixed, i.e., according to Eq. 25, 7cv is only a function of PH, yf and b (the Henry's law constant) at the operating temperature. Therefore, in the design of the complete clean-up PSA cycle for known feed conditions, the first step is to determine 7cv. As mentioned above, Eqs. 7 to 17 are valid for this complete cleanup case, and the design methodologies for both types of designs are essentially the same as in the case of a PSA-SVR process without breakthrough. The only notable differences between theses two cases are t h a t 7cv is determined for a known feed mixture (it is not chosen arbitrarily), and ac is related to 7M (it is not an independent design variable). Also, PSA-AP process designs usually do not consider ~ as a performance indicator. To illustrate, a conceptual PSA-AP process design with complete clean-up is carried out using the styrene-BAX activated carbon system. The adsorption isotherm at 298.2 K for this system [50] is plotted in Fig. 6 along with the model correlation. The isotherm model p a r a m e t e r s and ARE are given in Table 3. This design involves the determination of the bed dimensions, 7cv and ac for known feed conditions (PH = 101.3 kPa and yf = 1000 ppm ), and selected 7M = 0.5, Vf = 1 m 3 STP/min (1000 SLPM) and tf = 600 s. The choice of 7M = 0.5 is a compromise between the need for producing a clean light product and the need for reducing ac. The desired bed utilization at the periodic state is also specified (BCF = 0.6) with 40% of the bed being used as a guard against breakthrough. The selected 7 M has to satisfy Eq. 28, which gives 7 M ~_ 0.043; and the applicability of the equilibrium correlations constrained yf < 3430 ppm under the specified conditions (see Eq. 28).

232

5.0

i

i

i

i

iiii

I

........

I

........

I

........

I

........

I

I

.......

I

4.5 4.0

e x p e r i m e n t a l data

3.5 3.0 -~

2.5

0

E 2.0 1.5 1.0 0.5 0.0

. . . . . . . .

lx10 -6

[

~

lx10 -5

,

1

lx10 -4

,

lx10 -3

,

......

lx10 -2

I

lx10 -1

I

I I

IIII

lx10 ~

P (kPa) Figure 6. Equilibrium adsorption isotherm of styrene vapor on BAX activated carbon at 298.2 K [49]: experimental data and Langmuir model correlation.

First, the ideal gas law is used to calculate cf from the known yf, PH and Tf. Then Eq. 5 is used to calculate qf, Eq. 7 is used to calculate R, and 7cv and ac are obtained from Eqs. 25 and 26, respectively. Now, Eq. 10 is used to calculate ~- in terms of the calculated R and 7cV; and then the calculated ~-, is used to calculate Of through Eq. 14. Finally, Eq. 3 is used to determine the bed dimension, i.e., LD for a selected db, in this case. All of the inputs and outputs of this design are given in Table 3. Only two relatively small beds are required to achieve the specified performance, even though styrene is a relatively strong adsorbate. Moreover, the chosen process feed conditions are indicative of actual PSA-AP systems used in defense applications [33]. What allows these beds to be so small in this case is the complete clean-up of the beds during every cycle. This gives rise to a significant adsorbent capacity compared to PSA systems that leave a significant heel in the beds (e.g., see Figs. 3 and 5a). Also, PL = 1.17 kPa is still quite reasonable for conventional vacuum systems.

3.3. B i n a r y PSA-SVR p r o c e s s from inert carrier gas Many ternary systems exist throughout various industries, consisting of two dilute vapors in a relatively inert carrier gas (i.e., nitrogen or air), where it is

233 Table 3 Conceptual design of a PSA-styrene-AP process with complete clean-up during every cycle Design Input

Design Output

Adsorption isotherm p a r a m e t e r s b (kPa 1) 416.62 qs (mol/kg) 4.0498 ARE (%) 11.28

Operating conditions (Zc (-) 86.4 PL (kPa) 1.17 Vcv (-) 43.2

Bed information pu (kg/m a)

517.97

Bed dimensions db (m) (specified) Lb (m)

1000.0 101.3 298.15

7cM (-)

0.06 0.128

Feed conditions yf (ppm)

PH (kPa) Tf (K)

Model constraints yfmax (ppm)

> 0.043 3430.2

Process conditions Vf (SLPM)

1000.0

7 M (-)

0.5

tf (s)

600.0

Required process performance BCF (-) 0.6

advantageous to separate and recover the lighter vapor (typically a solvent or reaction medium like methylene chloride) from the heavier vapor (typically a monomer like acrylic acid) for recycle, resale, etc. In many cases, streams of this nature are still being vented to the atmosphere, or condensed as a binary mixture and then discarded as a waste [52]. Another example is the methylethylketone (MEK)-toluene system (common in the textiles industry), which is being removed from air using very large steam regenerated activated carbon beds [53]. But these systems necessarily generate a secondary waste stream consisting of the organiccontaminated condensate from such processes. Moreover, because the MEK and toluene are recovered as an organic liquid mixture, they cannot be recycled and are discarded as a waste [53]. With the development of these binary PSA-SVR processes, secondary waste streams can be avoided and the potential for recycle of at least one of the solvents becomes feasible. In the context of a two-component mixture, the goal is to separate the two impurities into two enriched products. One of the products is enriched in the lighter component, and the other in the heavier component, with both products

234 being comprised mostly of inert carrier. As proposed recently by Pigorini and LeVan [40] and extended by Subramanian and Ritter [42], the PSA cycle is operated as a typical bulk gas separation process, producing one of the products during the feed step as a mixture of the lighter impurity and inert carrier gas, and the other product during the purge step as a mixture of all three components, but relatively enriched in the heavier impurity. In this case, the purge step duration is terminated just when the lighter component is purged out of the bed. They [40,42] both used equilibrium theory to analyze this situation. The resulting analytic expressions are employed here to carry out the second kind of process design, where the column dimensions are determined based on a chosen set of operating conditions. In this design, the two beds undergoing feed and purge are coupled, and the feed and purge step durations are equal. Moreover, the formulation assumes a pure inert carrier gas purge throughout the purge step. But, this is true only for a determinable fraction of the purge (feed) step duration, after which the lighter impurity contaminates the effluent of the bed undergoing feed. So, an actual PSA separation process of the type being designed necessarily requires an intermediate vessel to store the pure inert carrier gas effluent for use as purge throughout the purge (feed) step duration. The analysis of the wave phenomena is done in the context of fully convex mixed Langmuir isotherms (see Eq. 5), or equivalently by the constant separation factor isotherm, which is given by Cl

,

ql = Zl

+ (1- R1)o~. + (RI/R2X1- R2)o~

(29)

for component 1 and similarly for component 2 with the subscripts interchanged. The separation factor, R/, given by Eq. 7, has been invoked in reaching Eq. 29 from the basic mixed Langmuir isotherm model. The inputs to the design are the feed and isotherm parameters, or the so called triad, {R1,R2,~}, where G is defined as

(30)

t c 1,f )t q~,f ) o is the solid phase ci, f is the concentration of component i in the feed, and qi,f concentration that exists in equilibrium with the feed concentration,

ci, f

corresponding to a pure-component feed. Note that R1 < ~IR2 ensures that component 1 is always the stronger adsorbing component. 7v is obtained from

235 v

. a2(bl -ao)

Y=ao(2albl-aobl-a2)

(31)

where ao, bo, al and bl are the Rieman invariants [42] that are directly related to the feed conditions and the adsorption isotherm parameters through the constant separation factors; they are obtained from Eqs. 32, 33, 36 and 37, respectively.

a~

n - 1 - 2h +2hR 1 + ~ 1 2R2 -~IR2 - 2hR1 +2hR 2 +fiR 2 ~ 1

(32)

rl-1 - 2h + 2hR 1 - ~ 1 2 - 2hR 1 + 2 h R 12-qR 2 ~ 1

(33)

b o=

q 2 R 2 -qR

where NI=(1 +2n +rl 2 -4nR2 -4nR1 +4nRIR2)

h=q

1 R1

(34) (35)

and al=

1 vlR2

(36)

1

bl=R1

(37)

For a given Vf, the feed step (and purge step) duration of the process is obtained from qRiR2boao ( 2 a l b l - a 2 -aobl ) -of= (b 1 _ao)

(38)

in which the dimensionless feed step duration is defined as vftf

~f =

~

(ql'f ]L b

(39)

236

or

0ftf

(40)

xf =(ql,f~c~,f/

Furthermore, fixing the fraction, f, of the pure inert carrier molar flux emanating from the column undergoing feed, that must be diverted as purge into the other coupled column undergoing purge, determines the pressure ratio of the process. If f is the fraction chosen for the process, then a is given by cz=

al b~ -a~ ao(albl + bobl - a l b o -aobl)f

(41)

which ensures that the molar amount of pure inert carrier gas is the correct amount required to force periodicity of the characteristic conditions. The recovery of the lighter component is defined as the fraction of moles of the lighter component fed into the process that is recovered as light product in the feed step; it is given by - a l X a 1 - ao)~blqR 2 -1) ~L=(2(~~al- bla o - a2 ~boqR2 - 1)

(42)

The enrichment of the lighter component in the light product is defined as the ratio of its mole fraction (or dimensionless concentration) in the product (at the rolled-up level) to that in the feed. This enrichment is expressed as (43)

EL'II= bl~boqR2

The recovery of the heavier component during the purge step, ~H is always 100%, since in this case the heavier component concentration wave is just contained inside the bed at the end of the feed step. The enrichment of the heavier impurity in the purge effluent relative to its mole fraction in the feed is expressed as EH = ( b l - a ~ l a 2 b

o §

(al - ao)(2 l - (Rl(ao - Ri)(1 q + al))(1 R - R1)2- boR1)

aoal (al bl + bo bl - albo -ao bl)f

(44)

The enrichment of the lighter impurity in the purge effluent relative to its mole fraction in the feed is expressed as

237

(bl_aolao2bo (al-ao)(2-~lR2(ao+al))(botlR2-1) 1~_(~22_~1)(1_R2 ) EL'IV=

aoal (al bl + bo bl - albo -ao bl )f

(45)

Note that the frozen solid phase assumption restricts the applicable pressure ratio to 1 o~<.(yI + Y2 )

(46)

This also restricts f in Eq. 41 to f>

al bo(bl -ao )/Yl + Y2,) ao(albl+ bob1 - albo - a o b l )

(47)

The specific characteristics of this binary SVR process results in the heavier component concentration wave front just reaching the light product end of the bed at the end of the adsorption step, and all of the lighter component just being purged from the feed end of the bed at the end of the purge step. Therefore, the bed is saturated with feed at the end of the adsorption step and only the heavier component (with the inert carrier gas) is left in the bed at the end of the purge step. The heavier component gas phase concentration profile at the end of the purge step is represented by a constant concentration spanning from the feed end of the bed to a determinable position, zp, beyond which a heavier component expansive simple wave profile exists. The constant heavier component concentration, C*l,const, is determined by , klk + 1 - h + ~(k + 1 - h ) 2 + 4 h 1 Cl,const = k+l+h+x/(k+i h) 2 + 4 h

(48)

where h is given by Eq. 35, 1"1is given by Eq. 30, and k = R1 - qR2 1-R 1

(49)

The simple wave starts from zp, or in dimensionless form, ~p, and is determined by the following relation:

al

% = 1- bo

(50)

238 where the dimensionless bed length, ~, is defined as =

Z -

(5~)

-

Lb and a l and b0 are given by Eqs. 36 and 33, respectively. Within the simple wave region, the heavier component concentration varies from Cl,const to zero, and the relationship between these concentrations and their axial position is given by = 1 - alb~

-RlCl -Rlk)2 k2

(52)

Therefore, the simple wave profile is readily plotted using Eq. 52 in terms of the heavy component concentrations between Cl,const and zero. The conceptual design of the binary PSA-SVR process is somewhat different from those discussed in previous sections because of the nature of the specific case considered, i.e., the purge step terminates just when all of the lighter component is purged out of the bed. In this case, if f is chosen, the process performance and a are uniquely determined based solely on the feed conditions (PH, yf,I and Tf) and the Langmuir isotherm parameters of the two solvent vapors. However, the selection of f must satisfy Eq. 47, and the calculated (z must satisfy Eq. 46. To start the design, the triad {R1, R2, q} is calculated corresponding to the known feed conditions (yf, i, PH and TO and the adsorption isotherm of each component using Eqs. 7 and 30 with ci,f calculated from the ideal gas law and the corresponding q~ calculated from the pure-component Langmuir isotherm model. Then Eqs. 32 to 37 are used to calculate ao, bo, al and bl. The process performance is then determined using Eq. 42 for 9{L, Eq. 43 for EL, II, Eq. 44 for EH and Eq. 45 for EL,~ for a selected f. Note that 9{H is always equal to 100% in this case. The required ~v is calculated from Eq. 31 and a is calculated from Eq. 41. If tc is specified, Of (0p) is calculated from Eq. 40 using the value of xf obtained from Eq. 38; or conversely, the corresponding tc is calculated using Eq. 40 if Of (0p) is specified. Of (0p) is then used to determine Lb or db from Eq. 3 if one of them is known; or to determine Vf from Eq. 3 for the specified bed dimensions. The binary n-heptane-(1)-n-butane (2)-BAX activated carbon system is selected to illustrate the conceptual process design of a PSA-SVR separation process. The adsorption isotherm data and Langmuir correlations for these systems at 298.2 K are displayed in Fig. 7, and the isotherm parameters and AREs are given in Table 4. The Langmuir representation of this n-butane-n-heptane - BAX system corresponds to R1 = 0.071, R2 = 0.352 and ~ = 5.413. Remember that the purge step of this process has a duration which just corresponds to complete removal of the lighter impurity (butane in this case). The equilibrium model limits a < 8.33 and f _> 0.346; so, an f = 0.5 is chosen. In this example, U, a and bed dimensions,

239 Table 4 Conceptual design of a PSA-n-heptane-n-butane solvent vapor separation process from inert carrier gas with n-butane separated from n-heptane Design Input

Design Output

Adsorption isotherm p a r a m e t e r s n-heptane (1): b (kPa ol) 5.37 qs (mol/kg) 4.9235 ARE (%) 6.176 n-butane (2): b (kPa -1) 0.1514 qs (mol/kg) 6.5194 ARE (%) 14.62

Operating conditions a (-) 5.76 PL (kPa) 21.1 7v (-) 1.906

Bed information pb (kg/m a)

480.0

Feed conditions yf,1 yf,2 PH (kPa) Tf (K)

0.02 0.10 121.56 298.15

Process conditions Vf (SLPM) tf (s) f (-)

4000.0 720.0 0.5

Bed dimensions Lb (m) (specified) db (m)

1.5 0.548

Process performance EL, II (-) ~i)~L (-) EL, IV (-) EH (-) Of (m3/kg/s)

1.184 0.40 1.812 3.022 3.565• 10 .4

Model constraints f (-) a (-)

_>0.346 _<8.33

as well as the process performance are sought for the given feed conditions (PH = 121.56 kPa, y f , 1 - 0 . 0 2 , y f , 2 - 0.1 and Tf = 298.2 K), a specified Vf = 4000 SLPM, and tc = 1440 s (or tf = 720 s). All of these conditions are listed in Table 4, along with the design outputs. For the complete removal of the lighter impurity (butane vapor) during the purge step, 7v = 1.906 and a = 5.76 are needed. This pressure ratio is within the range given by the model limitation ((z _< 8.33) and corresponds to a moderate PL = 21.1 kPa. A light product stream consisting of 11.84% n-butane in nitrogen (EL, II = 1.184) and a heavy product stream consisting of 6.044% n-heptane (EH = 3.022) and 18.12% n-butane (EL, rV = 1.812) in nitrogen are produced with the corresponding recovery of the n-butane of 40.0% during the adsorption step. Note that the recovery of n-heptane is 100% in this process. For the specified to, a Of = 3.565• .4 m3/kg s (at actual feed conditions) is obtained, which is used to determine the bed dimensions. With Lb specified at 1.5 m, the required db is 0.548 m.

240 7.0 6.0

n-heptane data n-butane data model correlations

9

A 5.0 --~

4.0

E cr

3.0

0

v

2.0 1.0 0.0 ! 0

0

X

| 0

X

0

X

0

X

0

X

0

X

0

X

X

0

X

P (kPa)

Figure 7. Equilibrium adsorption isotherms of n-heptane and n-butane vapors on BAX activated carbon at 298.2 K: experimental data and Langmuir model correlations.

1.5

'

I

'

I

'

I

'

I

'

I

'

I

'

I

'

I

'

I

'

C1 ,C 2 1.0 0.10 ..

o

i i I xl

0.00 0.5

-

t~ r

,

I',,. OO O~ C~

~o~o~o od6dd~

c1

" -

-

0.0

,

I,

0.0 0.1

I,

I,

0.2 0.3

I,

I,

a,

0.4 0.5 0.6

i,

I

,

0.7 0.8 0.9 1.0

z/t Figure 8. Periodic gas phase concentration profiles for the n-butane and n-heptane-BAX system at the end of the adsorption (solid lines) and purge (dash) steps; note that c2" - 0 everywhere inside the column at the end of the purge step.

241 The periodic gas phase concentration profiles at the end of the adsorption and purge steps are plotted in Fig. 8. At the end of the purge step, a constant n-heptane concentration of Cl,const - 0.306 exists in the column from ~ = 0 to - 0.813, i.e. a p l a t e a u region exists in the column. The simple wave starts at - 0.813 and it reaches zero at ~ - 0.993 The almost negligible region from - 0.993 to ~ = 1.0 is free of adsorbate and only consists of inert carrier gas. The insert in Fig. 8 shows this adsorbate-free region very clearly.

~P ~p

4.

SUMMARY

Simple analytic expressions derived from equilibrium theory are introduced for three evolving e n v i r o n m e n t a l PSA processes: single component solvent vapor recovery from inert carrier gas; single component air purification with complete clean-up during every cycle, and binary solvent vapor recovery from an inert carrier gas, where the lighter vapor is s e p a r a t e d from the heavier vapor. These expressions can be readily used for process heuristics, feasibility, design and development, as well as performance and understanding. Conceptual process designs are carried out for all three cases to illustrate the use of these simple expressions. It is anticipated t h a t because of their simplicity and ease of use, the theoretical developments and case studies presented here will be useful to both the novice and the expert for rapidly carrying out p r e l i m i n a r y studies of new PSA processes with e n v i r o n m e n t a l applications.

ACKNOWLEDGEMENTS The authors gratefully acknowledge financial support from the US National Science Foundation u n d e r G r a n t CTS-9410630, and from the Westvaco Charleston Research Center.

NOMENCLATURE Ab a ao al b bo bl ci c*i ci,f

bed cross sectional area, m 2 negative R i e m a n n i n v a r i a n t negative R i e m a n n i n v a r i a n t corresponding to the feed state negative R i e m a n n i n v a r i a n t corresponding to the pure inert carrier state positive R i e m a n n i n v a r i a n t , or isotherm p a r a m e t e r , k P a -1 positive R i e m a n n i n v a r i a n t corresponding to the feed state positive R i e m a n n i n v a r i a n t corresponding to the pure inert carrier state fluid phase concentration of component i , mol/m 3 dimensionless fluid phase concentration concentration of component i in the feed, mol/m ~

242

db EL,II EL.IV EH

f k L Lb N N1 PH PL

qs qi q*i q~ qexp,i qcal,i

R Ri

Tf 9~ t tc tf Vf

Vf yi yf, i z Za

bed diameter, m enrichment of the lighter impurity in the feed effluent enrichment of the lighter impurity in the purge effluent enrichment of the heavier impurity in the purge effluent fraction of the pure inert carrier gas light product used as purge intermediate parameter penetration of the square profile in the very first feed step, m bed length, m number of experimental data intermediate parameter feed pressure, kPa purge pressure, kPa adsorption isotherm parameters, mol/kg adsorbed phase concentration of component i, mol/kg dimensionless adsorbed phase concentration adsorbed phase concentration in equilibrium with pure component i at ci,f, mol/kg experimental adsorbed phase concentration of component i, mol/kg calculated adsorbed phase concentration of component i, mol/kg isotherm separation factor constant separation factor for compone~:t i feed temperature, K solvent vapor recovery time, s cycle time, s feed step time, s superficial feed velocity, m/min feed volumetric flow rate, m3/s gas phase mole fraction of component i gas phase feed mole fraction of component i axial coordinate, m bed length covered by the concentration shock at periodic state, m

Greek letters o~ (~0

P 7v ~/M

Ph T T

P

pressure ratio negative invariant positive invariant volumetric purge-to-feed ratio molar purge-to-feed ratio bulk density of the packing, kg/m 3 dimensionless time dimensionless purge step duration

243 G

Op Of

dimensionless axial coordinate, z/L dimensionless periodic state bed penetration dimensionless axial coordinate, z/Lb dimensionless ratio process throughput, m3/kg s adsorption step throughput, m3/kg s

REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

D.J. Pezolt, S. J. Collick, H. A. Johnson and L. A. Robbins, Environmental Progress, 16 (1997) 16. M.M. Mattia, Chem. Eng. Prog., 66 (1970) 74. J.W. Drew, Chem. Eng. Prog., 71 (1975) 92. P.C. Wankat and L. R. Partin, Ind. Eng. Chem. Process Des. Dev., 19 (1980) 446. P.N. Cheremisnoff, Pollution Engineering, June (1986) 26. D.R. Gerard, J. Testing and Evaluation, 17 (1989) 106. J.R. Graham and M. Ramaratram, Chem. Eng., February (1993) 6. Y. Takeuci, A. Shigeta and H. Iwamoto, Sep. Technol., 3 (1993) 46. D.M. Ruthven, Principles of Adsorption and Adsorption Processes, John Whiley & Son, New York, 1984. R.J. Holman and J. H. Hill, Paper presented at AIChE 1992 Annual Meeting, Miami Beach, FL, November, 1992. T. Hall and L. Larrinaga, Paper presented at the I&EC Special Symposium, American Chemical Society, Atlanta, GA, September, 1993. J.A. Ritter, Y. Liu and D. Subramanian, Ind. Eng. Chem. Res., 5 (1998) in press. R.L. Hasche and W. N. Dargan, US Patent No. 1794377 (1933). G.A. Perley, US Patent No. 1896916 (1933). D. Finlayson and A. J. Sharp, UK Patent No. 365092 (1930). H. Kahle, Chemie Ing. Technik, 23 (1953) 144. H. Kahle, Chemie Ing. Technik, 26 (1954) 75. C.W. Skarstrom, US Patent No. 2944627 (1958). R.T. Yang, Gas Separation by Adsorption Process, Butterworth, London, 1987. D.M. Ruthven, S. Farooq and K. S. Knaebel, Pressure Swing Adsorption, VCH Publishers, New York, 1994. S. Sircar, Adsorption: Science and Technology, A. E. Rodrigues, M. D. LeVan and D. Tondeur (eds.), Kluwer Academic Publishers, (1989) 285. G.E. Keller, Am. Chem. Soc. Symp. Ser., 223 (1983) 145. G.V. Baron, Separation Technology, E. F. Vansant (ed.), Elsevier Science B. V., (1994) 201.

244

24. D.E. Tevault, Personal Communication, US Army ERDEC, Aberdeen Proving Ground, MD, 1995. 25. Y. Liu and J. A. Ritter, Ind. Eng. Chem. Res., 35 (1996) 2299. 26. E.S. Kikkinides and R. T. Yang, Ind. Eng. Chem. Res., 30 (1991) 1981. 27. K.T. Chue, J. N. Kim, Y. J. Yoo, S. H. Cho and R. T. Yang, Ind. Eng. Chem. Res., 34 (1995) 591. 28. S. Sasaki, S. Matsumoto, M. Fujitsuka, T. Tanaka and J. Ohtsuki, IEEE Transactions on Energy Conversion, 8 (1993) 26. 29. M. Suzuki, Paper presented in Fifth International Conference on Fundamentals of Adsorption, PL-1, California, May, 1995. 30. Y. Liu and J. A. Ritter, Ind. Eng. Chem. Res., 36 (1997) 1767. 31. Y. Liu and J. A. Ritter, Adsorption, 3 (1997) 151. 32. Y. Liu and J. A. Ritter, Adsorption, 4 (1998) 159. 33. D.K. Friday, M. D. LeVan, J. J. Mahle and L. C. Buettner, Fundamentals of Adsorption, M. Suzuki (ed.), Elsevier, Amsterdam, (1993) 169. 34. J . J . Mahle, D. K. Friday and M. D. LeVan, Ind. Eng. Chem. Res., 35 (1996) 2342. 35. D.M. Ruthven and S. Farooq, Chem. Eng. Sci., 49 (1994) 51. 36. E.S. Kikkinides, J. A. Ritter and R. T. Yang, J. Chin. Chem. Eng., 22 (1991) 401. 37. L.H. Shendalman and J. E. Mitchell, Chem. Eng. Sci., 27 (1972) 1449. 38. Y . N . I . Chan, F. B. Hill and Y. W. Wong, Chem. Eng. Sci., 36 (1981) 243. 39. M.D. LeVan, Ind. Eng. Chem. Res., 34 (1995) 2655. 40. G. Pigorini and M. D. LeVan, Ind. Eng. Chem. Res., 36 (1997) 2306. 41. D. Subramanian and J. A. Ritter, Chem. Eng. Sci., 52 (1997) 3161. 42. D. Subramanian and J. A. Ritter, Chem. Eng. Sci., 53 (1998) 1295. 43. J.A. Ritter and R. T. Yang, Ind. Eng. Chem. Res., 30 (1991) 1023. 44. J.A. Ritter and R. T. Yang, Chem. Eng. Comm., 108 (1991) 289. 45. Y. Liu, C. E. Holland and J. A. Ritter, Sep. Sci. Technol., (1998) in press. 46. Y. Liu, C. E. Holland and J. A. Ritter, Sep. Sci. Technol., (1998) in press. 47. R. Gautam and C. M. Yon, AIChE Annual Meeting Preprints, Miami Beach, FL, Nov. 2-6, (1992) 459. 48. J.A. Ritter, Ph.D. Dissertation, State University of New York at Buffalo, Buffalo, 1989. 49. E.D. Tolles, Personal Communication, Westvaco Charleston Research Center, Charleston, SC, 1996. 50. D.H. White, Adsorption and Ion Exchange: Fundamentals and Applications, M. D. LeVan (ed.), 84 (1988) 129. 51. D.H. White and P. G. Barkley, Chem. Eng. Prog., January (1989) 25. 52. J.D. Schroer, Personal Communication, 3V, Inc., Georgetown, SC, 1994. 53. A.H. Carter, Personal Communication, Reeves International, Spartanburg, SC, 1994.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

245

M o l e c u l a r m o d e l l i n g of a d s o r p t i o n a n d d i f f u s i o n p r o c e s s e s i n z e o l i t e s i n r e l e v a n c e to e n v i r o n m e n t p r o t e c t i o n R. Vetrivel a, R.C. Deka a, S.B. Waghmode a, S. Sivasanker a, K. Mizukami b, H. Takaba b, M. Kubo b and A. Miyamoto b aCatalysis Division, National Chemical Laboratory, Pune - 411 008, India bDepartment of Materials Chemistry, Graduate School of Engineering, Tohoku University, Sendai 980-77, J a p a n

Unlike several other practical heterogeneous catalysts, zeolites are highly crystalline and well characterized materials. The surfaces amenable for the approach of molecules, the catalytic active sites for adsorption and the space available for their reaction are well-defined. The above facts have led to the development of zeolites as the breeding ground for molecular modelling methods. In turn, the molecular modelling methods have played a crucial role in improving our understanding of several facets of zeolite catalysis, thus establishing a symbiotic relation. In this work, we bring out the application of molecular modelling methods to understand, interpret and to some extent predict the properties of zeolite based environment-friendly catalysts. The studies applied to design zeolite catalysts which are prospective candidates to replace environmentally hostile catalyst materials such as mineral acids, chlorides of aluminum, zirconium, iron etc. are presented. Two typical cases, where modelling has been carried out on zeolite catalysts in order to understand the mechanism of complex organic transformations, are described. In case I, the yields in the conversion of several spirolactones to enones were rationalized and the role of shape-selectivity in controlling the yield are brought out. In case II, the distribution of Na and RE in zeolite-Y and their consequence on the yield of S-N acetals are understood. Further, the adsorption and diffusion characteristics of alkylaromatics in various zeolites are studied by force-field based energy minimization calculations. These studies have brought out the power of molecular modelling methods for i) the initial screening of zeolite catalysts in shape selective reactions, ii) to identify the energetically favorable and unfavorable locations for the molecules insides the pores of zeolites and iii) to deduce the energy barriers for the diffusion of the molecules. The energetically favorable locations of 1,2-dichlorobenzene, its electronic interaction with C12 and promoter inside zeolite K-L are studied. The results are useful to understand the mechanism of selective formation of 1,2,4-trichlorobenzene. Additionally, the

246 attempts based on molecular modelling studies towards the design of zeolite catalysts for pollution control such as deNOx and removal of chloroflurohydrocarbons by adsorption over zeolites are described. The mode of activation of c g 4 and H20 over Ga-ZSM-5 are brought out. The influence of the extra framework cations on the adsorption of CF2C12 on CsNaY is revealed. Overall molecular modelling studies throw light on the underlying chemical forces - their nature and magnitude, which control the behavior of the reacting molecules inside the zeolite pores.

1. I N T R O D U C T I O N Molecular modelling methods basically comprise of i) molecular graphics and molecular fitting techniques, ii) force-field based calculations and iii) q u a n t u m chemical calculations. The force-field based calculations include energy minimization, Monte Carlo and molecular dynamics calculations. We have used all the above methods at appropriate places to study zeolite catalyzed reactions. In the following sections, we describe the application of molecular modelling methods to understand the adsorption and diffusion of molecules inside zeolites. These processes are particularly relevant to the protection of the environment, because either these molecules are conventionally synthesized using environmentally hostile catalysts or these molecules themselves are environmentally hostile. We describe typical case studies where complex steps of an organic transformation occurs in a single step over zeolite catalysts (sections 2 and 3), production of the required isomer of alkylaromatics (section 4) and chloroaromatics (section 5) occurs with high selectivity, selective catalytic reduction of NO (section 6) and the adsorption of chloroflurocarbons (CFCs) inside zeolites (section 7).

0

THE ROLE OF S O L I D ACID CATALYSTS IN THE S Y N T H E S I S OF ACYCLIC E N O N E S

2.1. B a c k g r o u n d a n d m e t h o d of a p p r o a c h Acyclic enones are important intermediates in the total synthesis of natural products and complex organic molecules [1-3]. The acid catalyzed intramolecular transformation of spirolactones over methanesulphonic acid - P205 mixture is the conventional procedure used in the synthesis of these enones[4,5]. The difficulties in working up and the requirement of a large excess of the acid reagent are the drawbacks of this procedure. Recently, there has been several reports wherein hazardous catalysts were replaced by environment-friendly zeolite catalysts for organic reactions [6-8]. Pillai and coworkers have demonstrated the successful use of zeolite catalysts for the synthesis of acyclic enones [9,10] and in the present study, we apply molecular modelling methods to

247 model the spirolactone to enone conversion. The structural relations between the spirolactone and the zeolite voids have been analyzed by molecular graphics and fitting techniques, which otherwise need to be determined by tedious experimental studies. The success of conversion of specific spirolactones (1-5) shown in Figure 1 over zeolites with different pore architecture is analyzed. The reasons for the r e q u i r e m e n t of 'stoichiometric' quantity, r a t h e r t h a n 'catalytic' quantity of zeolites are probed. The electronic interactions between the organic molecules [spirolactones (1-5) and enones (6-10)] shown in Figure 1 and the zeolite H-Y host lattice are studied to u n d e r s t a n d the mechanism of this conversion reaction.

1

6

2

8

5

g 4

5

tO

Figure 1. The 3-D structural view of reactant spirolactones and the product enone molecules. Their actual dimensions are given in Table 1. The conversions shown above were carried out over various zeolites and the yields obtained [9] are given in Table 2. The terminal hydrogen atoms are not labeled.

248

2.2. D i m e n s i o n o f t h e r e a c t a n t a n d p r o d u c t m o l e c u l e s The size and shape of the molecules are crucial p a r a m e t e r s t h a t decide their diffusivities inside the constrained space of zeolite voids. The molecular size of a guest molecule is usually characterized by a critical diameter, de [11], a LennardJones length constant, Sm [12], or a m i n i m u m kinetic diameter of molecule, dm [13]. The molecular sizes defined by above methods pose problems such as, either they are too difficult to estimate or the estimated values are not reliable. We use more realistic values to describe the size and shape of the molecules. For the energetically favorable conformation of the molecules, the largest dimensions (a x b x c) in m u t u a l l y perpendicular directions of the molecules are measured. Assuming t h a t the molecules are exactly fitting inside the smallest possible rectangular box as shown in Figure 2, the dimensions of the molecules are the dimensions of the box. The dimensions of the various reactants and product molecules are given in Table 1.

Figure 2. The molecule is fitted inside the smallest possible rectangular box and the dimensions of the box (a x b x c) are the dimensions of the molecule. The fitting of spirolactone (1) inside a rectangular box is shown as a typical case.

The product enone molecules (6-10) are of similar or smaller dimension t h a n the r e a c t a n t spirolactone molecules (1-5). It is convention to neglect the largest dimension (a) of the molecule [14], since the molecules diffuse into the cb-annels and cages with their largest dimension lying parallel to the channel axis. In general, the product molecules have small dimensions with larger strain energy. In Table 1, we report the strain energy due to bonded and non-bonded interactions. The analysis of the individual contributions of various bonded and non-bonded terms to the strain energy of these molecules are discussed elsewhere [15]. We note t h a t the bonded energy terms are unfavorable, while the nonbonded terms are favorable as indicated by almost zero or negative values.

249 Compared to the reactants, products have unfavorable strain energy, particularly non-bonded energy. These results indicate t h a t the products may be more polar, which could lead to their stronger adsorption inside the zeolite void volumes. This is in correspondence with the experimental findings [9,10] t h a t the final product h a d to be solvent extracted after the reaction. A probable mechanism and the characteristics of the intermediates are reported elsewhere [16].

Table 1 The dimensions of various molecules at their m i n i m u m energy configuration. The geometry of the molecule is optimized to obtain m i n i m u m strain energy using the force field calculations described in CHARMM software package distributed by Molecular Simulations Inc., USA. The n u m b e r s of the molecules are as assigned in Figure 1 Molecule No. of Dimension (A) Strain energy (kcal/mol) atoms Total Bonded Nonbonded 1 22 7.00 x 4.25 x 4.00 15.7889 21.5913 -5.8035 2 25 7.00 x 4.25 x 4.25 6.4875 10.2886 -3.8035 3 29 7.50 x 5.50 x 5.50 20.8836 21.8529 -0.9693 4 32 7.50 • 6.00 • 6.00 19.8921 24.8102 -4.9182 5 35 7.75 x 5.65 x 5.65 8.5793 12.3631 -3.7837 6 7 8 9 10

19 22 26 29 32

6.50 6.00 6.75 7.10 8.00

• • • • x

4.50 4.65 5.10 5.50 5.25

x • • • x

4.50 4.65 5.10 5.50 5.25

17.8080 15.6723 27.7583 21.7622 18.5582

19.9514 15.6289 30.1872 24.8070 17.8135

-2.1435 0.0434 -2.4270 -3.0447 0.7447

2.3. I m p o r t a n c e o f p o r e d i m e n s i o n s a n d p o r e a r c h i t e c t u r e o f z e o l i t e s The pore diameters of various zeolites are known from reported crystal structures [17]. It is possible to study the fitting of these molecules in several zeolites. The product yields of enones over different zeolites obtained in the experimental study by Pillai and coworkers [9,10] are given in Table 2. The yields of the enones (6-10) over various zeolites listed in Table 2 could be rationalized from the .pore diameters of the zeolites. H-ZSM-5 has 2-d channels with diameters of 5.1A and 5.4A, while H-ZSM-12 has l-d elliptical channel of diameters 5.7A and 5.9A. H-Beta has 2-d channels with diameters of 6.5A and 7.5A. H-EMT and H-Y have larger supercages with elliptical (7.1A • 7.4.&) and circular (7.4A) openings, respectively, into the supercage. The bicyclic lactones (6 and 7) are small enough to enter the medium (with 10-m rings) pores of

250 H-ZSM-5 as well as the large (with 12-m rings) pores of H-ZSM-12, H-Beta, H-EMT and H-Y. However, the tricyclic lactones are having an a n g u l a r conformation. They can not diffuse into the channels of H-ZSM-5 and H-ZSM-12. Thus the product yields of various enones obtained over different zeolite catalysts are in correspondence with their structural fitting. It can be generalized t h a t large pore zeolites with structures are efficient catalysts for the conversion. The range of dimensions of organic molecules are just in between t h a t of 'medium' (H-ZSM-5) and 'large' (H-ZSM-12, H-Beta, H-EMT and H-Y) pore zeolites. The roles of shape selectivity in controlling the product yield are clearly brought out from these results.

Table 2 The yield of various enones over different zeolites (Temperature = 150~ Time = 6 h, Amount of catalyst = 5.0 g., Amount of substance = 0.5 g., Solvent = Hexane) [9] Catalyst Yield of enones (mol. %)

H-ZSM-5 H-ZSM-12 H-Beta H-EMT H-Y

(6)

(7)

(8)

(9)

(I0)

69 79 77 84 92

61 68 70 78 95

0 0 52 64 90

0 0 58 66 89

0 0 56 68 86

2.4. I m p o r t a n c e o f e l e c t r o n i c i n t e r a c t i o n s Although the fitting of the molecules inside the voids of the zeolite catalysts is a necessary condition, the electronic interactions are also expected to play a vital role in controlling the yield of the reaction. Typical results of experimental findings by Pillai and coworkers [9,10] for the formation of enone (8) from spirolactone (3) at 150oc for 6 h, over H-Y are given in Table 3. The results indicated t h a t the yield obtained is dependent on the solvent media as well as on the a m o u n t of catalyst used. When the a m o u n t of catalyst is reduced, the n u m b e r of supercages as well as the catalyst surface available for the r e a c t a n t s decrease. This leads to a drastic decrease in the product yield (please compare the experiments 1 and 5 in Table 3) for a given reaction. In experiment 5, the catalyst a m o u n t is reduced to 1/100th of the amount in experiment 1, while keeping all the experimental conditions same. The product yield decreased from 90% to less t h a n 1%. If the reactions are occurring on the external surface, such a drastic decrease in product yield is not expected. The adsorption of molecules on external surface is expected to be weak compared to t h a t inside the cage. Hence, the active sites on the surface will be easily replenished, compared to those sites inside the cages. Thus the results in Table 3 indicate t h a t the reactions m a y be occurring only inside the cage and not on the externa! surfaces.

251 Table 3 Comparison of the yield of enone (8) over zeolite H-Y at various reactant/catalyst weight ratios. The effect of different solvents on the final product yield is also shown (Temperature - 150oc, T i m e - 6 h) Reactant/ Weight of No. of Weight of No. of No. Catalyst Super reactant Reactant Catalyst Solvent* Yield (g) cages (g) molecules Weight (%) ratios 1 10.0 1.0xl021 1.0 3.6x1021 0.10 H 90 2 10.0 1.0xl021 1.0 3.6x1021 0.10 M 19 3 5.0 5.0x102~ 0.5 1.8x1021 0.10 M: H (1:4) 58 4 0.2 2.0x1019 1.0 3.6x1021 5.00 M: H (1:4) 8 5 0.1 1.0x1019 1.0 3.6x1021 10.00 H <1 * H" Hexane; M: Methanol.

We studied the electronic properties of the molecules as well as the mechanism of electron transfer between organic molecules and the framework of zeolite H-Y by the semi-empirical q u a n t u m chemical calculations. The calculations were carried out using the s t a n d a r d MOPAC code [18] using the AM1 Hamiltonian [19]. The total energy of the reactant, and the product molecules are given in Table 4. The total energy values of P r o d u c t + Water are grater t h a n t h a t of reactant. These results are in correlation with the strain energy obtained from force field calculations (Table 1). The net charges on various atoms calculated from Mulliken population analyses are also given in Table 4 and they indicate t h a t the ketonic oxygens are more negative t h a n the

Table 4 Electronic properties of reactants and products Total Molecule Change on oxygens energy (eV) Ring Ketonic 1 -1829.25 -0.28 -0.32 2 -1986.88 -0.24 -0.30 3 -2270.42 -0.23 -0.30 4 -2426.12 -0.23 -0.30 5 -2582.01 -0.24 -0.30 6 7 8 9 10

-1481.28 -1637.71 -1920.73 -2076.03 -2232.75

-0.28 -0.29 -0.28 -0.28 -0.29

Bond order between C and O Ring 0.934 0.923 0.924 0.923 0.923

Ketonic 1.857 1.831 1.825 1.826 1.832 1.868 1.883 1.866 1.866 1.884

252 ring oxygens. Hence the ketonic oxygen atoms are expected to have interaction with the Bronsted acid site of zeolite. The orders of the ~ bond in the ring as well as the ketonic bond are also given in Table 4. Since H-Y was found to give the best yield (Table 2) among all the zeolites used, we studied in detail the interaction of molecules with the H-Y framework. The 12-member pore opening in zeolite H-Y has C6v s y m m e t r y along the 'b' axis, as shown in Figure 3 The symmetric molecules such as benzene [20] and 1,3,5-trimethylbenzene [21] are known to take symmetric locations at the center of this pore opening. Hence, t e t r a m e r i c cluster models centered at 01 and 04 are considered to r e p r e s e n t the active site.

Figure 3. The molecular graphics view of the 12-member pore opening into supercage of zeolite H-Y as viewed along the b-axis. O1 and 04 oxygens are alternatively repeating. Thus there is C6v symmetry along b-axis.

The adsorption complex formed between the molecule and the zeolite is simulated by a cluster model containing a t e t r a m e r of zeolite and the organic molecule. The cluster model is derived by fitting the molecule inside the supercage of H-Y. The two criteria considered in creating the cluster models were t h a t there should be interaction between the Bronsted acid site and ketonic oxygen, as well as the created orientation should not lead to any bad contacts between the van der Waals radii of atoms in the molecule and zeolite framework.

253 The organic molecules are found to have multi-site interaction with the framework [16]. In the 'zeolite cluster:organic molecule' complex, it was found that the major interaction was between the methylene hydrogens of the organic molecule and the oxygens of the framework, which support the mechanism proposed by Pillai and coworkers [9,10].

2.5. S a l i e n t o u t c o m e The yield obtained in the conversion of several spirolactones to enones are in correspondence with the structural fitting of these molecules inside zeolite channels or cages. Thus the shape selective catalytic behavior of various zeolites has been rationalized and some generalizations on the mechanistic aspects have been derived from the molecular fitting and quantum chemical calculations. The interaction energy of the products with the framework are more favorable than the interaction energy of reactants with the framework. Thus the final products formed have to be solvent extracted and the cages are not available for other reactants. The specific orientations and interactions of the organic molecules at their minimum energy conformations are useful to understand the mechanism of conversion.

3. S-N A C E T A L S F O R M A T I O N BY C-S BOND CLEAVAGE 3.1. B a c k g r o u n d a n d m e t h o d of a p p r o a c h Rajappa and coworkers [22,23] proposed a novel route for the synthesis of functionalised ketene S,N-acetals. 1-Methylamino-l-methylthio-2-nitroethylene is an important inter-mediate in the synthesis of a popular ulcer drug ranitidine [24]. Ranitidine, a H2-receptor antagonist, is a powerful inhibitor of gastric acid secretion and is extensively used in peptic ulcer therapy [25]. In general, nitroketene S,N-acetals have been shown to be useful intermediates for several nitroheterocycles by Rajappa and coworkers [26,27]. Due to the disadvantages in the conventional routes, a novel route shown in Figure 4 was reported [22,23] for the first time using rare earth exchanged NaY zeolite catalysts. Here, we report the results of force field calculations adopted for the molecules involved to understand their geometry and conformational flexibility. The visualization and the calculations of energy were performed using QUANTA/CHARMM software packages distributed by Molecular Simulations Inc. The force field equations are as described by Gelin and Karplus [28] and the parameters are listed elsewhere [29]. The zeolite-Y lattice was modeled from the crystal structure reported from X-ray crystallographic studies [30]. 3.2. M o d e l o f c a g e s a n d w i n d o w s in zeolite-Y There are many distinct types of cationic sites, three of which are significant for cation occupancy in faujasite-type zeolites described as I, II and III (Figure 5) [31]. The electrostatic field strength felt by the cations in each type of extra framework site is different and hence the catalytic activity induced by them will

254

/SMe R--N---C\sMe + CH3NO2

/SMe R--NH--C--SMe \CH2NO2

~

la-i

MeS\ y

-MESH

2a-i R

3a-i

R

a= ~ ~ ' , ~

OMe

b= ~ ~ ' - , ~

R

d= @

O

/C---CHNO 2

RHN

OMe

g= ~

Z,~ ~

O OMe

e= ~ ~ ' , , , ~

/

U= ~ " ~ p h

O Ph

O Figure 4. Condensation of nitromethane with the carbonimido dithoates (la-i) leading to variable yields of nitro ketene S, N-acetals (3a-i) with the elimination of methane thiol. The possible intermediates (2a-i), which on cleavage of C--S bond leads to the product are also shown. The efficient use of zeolite-Y for these novel condensation reactions are reported by Rajappa & coworkers [22,23].

t Figure 5. The schematic representation of the supercage in zeolite-Y framework. T atoms (where T=Si or AI) are at the vertices of the lines. The oxygen atoms are approximately midway between the T atoms and they are not shown in the Figure for clarity. The different possible extra framework cationic sites are shown.

255 also be different. F u r t h e r , the presence of cations in different types of sites will alter the d i m e n s i o n of windows opening to the cages. Therefore, s h a p e selectivity can be visualized as a function of the n a t u r e a n d location of cations. T h e r e are 16 type I, 32 type II a n d 48 type III cationic positions in the faujasitic s t r u c t u r e . For N a Y with an Si/A1 ratio of 4.2, all type I sites a n d n e a r l y two t h i r d s of type II sites will be occupied by Na. It is k n o w n t h a t Na in type I sites is very stable a n d it is difficult to pull it out from the small cage. Hence in RE (70%) NaY, the large tripositive l a n t h a n i d e ions will occupy type II sites [32]. It should be noted t h a t a m o n g the 32 type II sites available in the unit cell, only one fifth will be occupied by r a r e e a r t h cations. However, it is possible t h a t the l a n t h a n i d e s m a y also occasionally e n t e r the sodalite cage a n d hence m a y not have any influence on the shape selectivity. As shown in Figure 5, each supercage h a s four 12-T ring windows. A m o n g these four windows, one or none of t h e m m a y be blocked by r a r e e a r t h cations. However, the other 12-T windows are fully open for molecular traffic. The dimensions of the 12-T window and the supercages with different cationic contents are given in Table 5.

Table 5 Reduction in d i m e n s i o n s of the 12-m window a n d the supercage in zeolite-Y due to different cation occupancy Case

P r e s e n c e of e x t r a f r a m e w o r k cationsa

1 2 3 4

None 16 cations/u.c, in type I site 8 cations/u.c, in type II site 32 cations/u.c, in type III site

Dimensions of 12-m window (/k) 7.4 7.4 7.4 7.4

x x x x

7.4 7.4 6.4 5.4

Dimensions of supercage (/k) 12.4 12.4 12.4 10.4

x x x x

12.4 12.4 12.4 10.4

x x x x

12.4 12.4 10.4 10.4

a Cation radii are assumed to be 1.ON; u.c. = unit cell.

3.3. M o d e l of r e a c t a n t a n d p r o d u c t m o l e c u l e s The force-field e n e r g y m i n i m i z a t i o n calculations were p e r f o r m e d for all the r e a c t a n t s (la-i) a n d products (3a-i) shown in Figure 4. The total energy values for the molecules in t h e i r favorable conformations are given in Table 6. From Figure 4, it can be seen t h a t in the r e a c t a n t molecules, the s u b s t i t u e n t 'R' is far from the other a t o m s a n d the conformation of the molecule is not d e p e n d e n t on 'R'. In contrast, the s u b s t i t u e n t 'R' is closer to -nitro a n d -SMe groups in the product molecules, respectively, a n d hence the conformation is found to depend on the n a t u r e of 'R'.

256 Table 6 Total s t r a i n e n e r g y a n d the d i m e n s i o n s of various molecules c a l c u l a t e d for t h e equilibrium structures No. of a t o m s

Molecule No.

22 25 35 31 22 25 29 26 29

la lb lc ld le If lg lh li

Total e n e r g y (kcal mol 1) -2.96 -7.12 9.74 -0.93 -3.26 -5.82 -2.85 0.59 2.55

D i m e n s i o n s (/k) a • b • c 9.75 • 6.40 • 1.75 9.75 • 6.25 x 3.75 10.10 • 8.40 • 5.60 10.60 • 5.60 • 4.75 8.50 • 6.40 • 1.75 7.75 • 6.50 x 3.80 8.10 • 5.90 x 4.10 10.75 • 6.25 • 1.80 10.25 • 6.40 x 4.00

Yield %

23 3a -3.97 10.60 x 6.40 x 3.50 25.00 a 26 3b -10.65 9.80 • 7.00 x 3.60 65.00 36 3c -1.28 8.75 • 7.75 x 5.60 45.00 32 3d -10.10 10.00 • 7.50 x 4.00 0.00 23 3e -0.88 8.30 x 7.40 x 3.25 76.00 26 3f -4.74 8.90 • 6.20 x 3.00 82.00 30 3g -2.86 8.75 x 6.70 x 4.25 75.00 27 3h 0.70 9.50 x 7.25 x 3.50 80.00 30 3i -0.50 10.30 x 8.30 x 3.75 50.00 a A cyclic product from self-condensation of la was also obtained in 53% yield along with this compound.

3.4. C o r r e l a t i o n o f m o l e c u l a r d i m e n s i o n s to z e o l i t e - Y v o i d d i m e n s i o n s The t h r e e l a r g e s t d i m e n s i o n s (a • b • c) for the m i n i m u m e n e r g y c o n f o r m a t i o n of the molecules were calculated, as described in section 2.2 a n d c o r r e l a t e d to the d i m e n s i o n s of pores of zeolite-Y. It was found t h a t the efficiency of NaY was very low for t h e s e conversions. NaY will c o r r e s p o n d to case 4 as shown in Table 5, b e c a u s e only 50% occupancy in type I sites are possible owing to overcrowding. Accordingly, the d i a m e t e r of the void space in a s u p e r c a g e will be r e d u c e d to 10.4/k from 12.4/k. More crucially, the d i a m e t e r of 12-m windows which are the e n t r y points to s u p e r c a g e s is also reduced. C o m p a r i n g the d i m e n s i o n s of the molecules given in Table 6, it is clearly e v i d e n t t h a t the e n t r y a n d exit of r e a c t a n t a n d p r o d u c t molecules to the s u p e r c a g e t h r o u g h 12-T window is severely restricted. However, w h e n t h e r e is 70% e x c h a n g e of Na + by r a r e e a r t h cations to form RE (70%) NaY catalyst, the s i t u a t i o n will c o r r e s p o n d to case 3 s h o w n in Table 5, since t h r e e Na + are r e p l a c e d by one RE cation. W h e n c o r r e l a t i n g t h e m o l e c u l a r

257 dimensions with the zeolite void dimensions, it is customary to neglect the largest dimensions (length) of the molecule. This is because the molecule can and does energetically prefer to enter the cage through its smaller dimensions. There are more detailed analyses on the effect of shape of the molecules in the literature. The model proposed by Choudhary and Akolekar [33] is a typical example which support the fact t h a t the molecules prefer to enter the pores through their smaller dimensions. 3.5. M o l e c u l a r f i t t i n g Several criteria could be assigned to the observed efficient ~ bond cleavage reaction over RE (70%) NaY. The r e a c t a n t and product molecules have to be smaller t h a n the 12-m windows, because they have to enter or leave supercages through these windows. The active sites are inside the supercages since the proposed i n t e r m e d i a t e s can be formed only inside the supercages owing to their large dimensions. The molecules can enter the cage in certain preferred orientations, which are graphically visualized by m a t c h i n g the r e a c t a n t s with the windows. As discussed in the previous section, the largest dimension of the molecule does not matter. On this basis, when the r e a c t a n t molecule is too small, say as in the case of l d, then products are formed in very low yield, probably because of their short residence time inside the cage due to free diffusion of the molecules. When the r e a c t a n t molecule is too large, say as in the case of lc, then also the yield decreases. Actually the size of l c (8.4 x 5.6/k) is slightly larger t h a n the 12-m window and the size of l d (5.6 x 4.75 A) is slightly smaller t h a n the 12-m window, but for all other r e a c t a n t s the second largest dimension is in the range 6.0-6.5/k. However in the case of 3a, although the molecule is small, the product yield is low. The low yield may be due to the observed formation of a cyclic product [23] by self-condensation of two r e a c t a n t molecules (la). The selfcondensation reaction occurs only with r e a c t a n t l a because of the presence of an active methylene group, with a highly acidic proton. The dimensions of the product molecules are close to the d i a m e t e r of 12-m windows. The yields obtained are found to be a function of their sizes and those with comparatively larger dimensions such as 3c and 3i are obtained in lower yields. Hence, there is a kind of molecular recognition m e c h a n i s m by the port entries to the cage which decide the efficacy of the reaction. The formation of the i n t e r m e d i a t e and cleavage of the ~ bond in this i n t e r m e d i a t e occur ideally inside the supercage. The dimensions of the proposed i n t e r m e d i a t e s are such t h a t they easily fit in the supercage of RE (70%) NaY. In Table 7, the contributions of individual energy t e r m s to the total strain energy are given. It was found t h a t the dimensions of the molecules are not dependent on just the n u m b e r of atoms but more on their conformation. The conformations themselves are controlled by the attractive electrostatic interaction (Eelectrostatic) and can non-bonded van der Waals interactions (EvDw), as can be seen from the values given in Table 7. The i n t e r m e d i a t e s were always found to have greater

258 s t r a i n energies t h a n the r e a c t a n t s or products. There are also m a n y degrees of conformational freedom for these molecules. The exact conformation adopted by the molecule m a y also be influenced by external factors such as electrostatic interactions between the molecule and the zeolite, the electrostatic field inside the cages, solvents used for the reaction medium and the t e m p e r a t u r e of the reaction. In some conformations, there is a possibility of the reacting group being sterically shielded, resulting in non-reactivity of the r e a c t a n t molecules. This m a y be the reason for the absence of a 1:1 correlation of molecular dimensions to the yield obtained.

Table 7 Contributions of individual energy term s (kcal mo1-1) to the total s t r a i n energy for the final energy minimized molecules

Molecule

Ebond length

Ebond angle Edihedralangle Eelectrostatic

EVDW Eimproper

Etotal

torsion

la lb lc ld le If lg lh li

0.05 0.07 0.35 0.18 0.02 0.02 0.07 0.18 0.19

0.21 0.29 2.07 1.26 0.12 0.22 0.36 0.34 1.09

0.00 0.30 5.35 4.63 0.00 0.21 0.20 0.00 4.39

-3.56 -8.15 -0.44 -7.00 -3.30 -6.07 -3.28 -2.78 -5.77

0.33 0.35 2.39 -0.02 -0.09 -0.21 -0.21 2.86 2.64

0.00 0.00 0.02 0.03 0.00 0.00 0.00 0.00 0.01

-2.96 -7.12 9.74 -0.93 -3.26 5.82 -2.85 0.59 2.55

3a 3b 3c 3d 3e 3f 3g 3h 3i

0.02 0.01 0.27 0.12 0.08 0.02 0.05 0.15 0.15

0.15 0.29 1.41 0.61 0.39 0.31 0.42 0.04 0.49

1.08 0.98 1.04 0.60 2.06 0.15 0.13 0.58 2.30

-3.70 -10.17 -1.10 -9.82 -2.90 -3.27

-1.83 -2.03 -3.07 -1.64 -1.11 -1.97 -3.52 -0.53 1.38

0.32 0.25 0.15 0.03 0.60 0.01 0.24 0.34 0.23

-3.97 -10.65 -1.28 -10.10 -0.88 -4.74 -2.86 0.70 0.50

-0.93 -5.04

3.6. S a l i e n t o u t c o m e From the present study, the following conclusions can be drawn: the supercages in zeolite-Y have the requisite characteristics for catalyzing the condensation of n i t r o m e t h a n e with N - s u b s t i t u t e d carbonimidodithioates. The effect of the Si/A1 ratio and the n a t u r e of extra framework cations in modifying the cage and window dimensions[34] can be elucidated using molecular modelling

259 studies. The force field energy minimization calculations predict the equilibrium configuration of molecules. The shape and size of these molecules provide the rationalization of the yields obtained for the reactions involving these molecules.

4.

S H A P E SELECTIVE P R O D U C T I O N OF ALKYLAROMATICS

4.1. B a c k g r o u n d Alkylation of benzene and substituted benzene leads to alkylaromatics and conventionally this process is carried out over Friedel-Crafts catalysts [35]. Alkylaromatics have found several applications as intermediates or as end products in the manufacture of synthetic fibres, rubber, plastics, detergent, highoctane gasoline and fine chemicals [36,37]. Most of the Friedel-Crafts catalysts are liquids and there is always working-up problem in addition to corrosivity, toxicity and polluting effluents. The recent awareness and the universal aversion towards environmentally hostile materials are the reasons, which led to the search for environmentally friendly solid acid catalysts. Considering the above two requirements, zeolites emerge as consensus candidate material with desired properties. A fascinating structure-related aspect of the zeolite catalysis is molecular shape selectivity [38-40]. The subtle interplay of "configurational" diffusion and intrinsic kinetics of reactions in the intracrystalline pore system enable zeolite catalysts to differentiate between molecules or transition states involved in a reaction on the basis of their size and shape, and thus direct the reaction along specific paths. The diffusion of molecules in zeolite pores plays the major role in this shape selective process. The understanding of its mechanism can greatly facilitate the design of zeolite catalysts. In this work we attempt to find out a suitable zeolite catalyst for selective synthesis of p-IBEB which is the key intermediate in the production of (z-(4-isobutylphenyl) propionic acid, a popular analgesic drug ibuprofen [41]. The conventional synthesis route for the production of p-IBEB involves the alkylation of isobutylbenzene with ethene over Lewis-acid catalysts [42] in which the selectivity towards p-IBEB is typically 17.6%. Recently, the possibility of using zeolite catalysts for the production of p-IBEB by disproportionation of isobutylbenzene and a polyalkylbenzene over HY zeolite [43] has been shown to lead to better selectivity (46.3%). Although HY zeolite is better t h a n conventional catalyst, selection of this specific zeolite is a random choice r a t h e r than a logical selection and no experimental results are available for any other zeolites. Hence, we analyzed the efficacy of different zeolites. We have chosen large pore zeolites with l-d pores, where only single file diffusion is possible, zeolites with 2-d pores, where different molecular reorientation is possible at channel intersections and zeolites with 3-d pores where cage to cage translations are possible, with total dynamic freedom inside the cages. The influence of change in pore architecture on adsorption and diffusion behavior of the molecules is studied in detail.

260

4.2. A p p r o a c h Force field energy minimization technique is an efficient method for studying the location and conformation of large guest molecules within the micropores of zeolites. We demonstrate here that this approach can be extended to study diffusional behavior of alkylbenzenes in zeolites. We use this technique to study the adsorption and location of ethylbenzene (EB), isobutylbenzene (IBB) and o-, m- and p-isomers of isobutylethylbenzene (IBEB) in several large pore fully siliceous form of zeolites. The atomic positions of the fully siliceous zeolite framework are held fixed and the adsorption and diffusion behavior of the molecules are studied at zero surface coverage. The methodology that we have used for calculating diffusion energy barrier is based on energy minimization procedure. This technique is widely used in the recent past, for example, to study alkylation of naphthalene [44], to study the skeletal isomerisation of l-butene to isobutene [45] and to study amination of methanol [46]. 4.3. D i f f u s i o n a l b e h a v i o r of a l k y l a r o m a t i c s in FAU, LTL, MAZ and MOR Conformational analysis of the alkylaromatics is performed by allowing cooperative motion of the alkyl groups and the strain energy values for different conformers of the molecules are determined. For the energetically favorable conformation, the three largest dimensions (a x b x c) in mutually perpendicular directions of the molecules are given in Table 8. When correlating the dimensions of the molecules with the pore diameter of zeolites for molecular fitting purposes, it is customary to neglect the largest dimension (a) of the molecules [14]. The molecules prefer to enter the cages through their smallest dimensions on the basis of interaction energy criteria. Hence, only the other two dimensions (b and c) of the molecules have to be compared with the size of the pore openings. Comparing the dimensions (b and c) of the molecules in Table 8, it is observed that 'small' (with 8-member rings) and 'medium' (with 10-member rings) pore zeolites are too small to accommodate isomers of IBEB, while 'large' pore zeolites are suitable for the diffusion of IBEB isomers. A list of 20 large pore

Table 8 Dimensions of different organic molecules as derived from the force-field calculations Dimensions//k

Molecules EB p-DEB IBB m-IBEB o-IBEB p-IBEB

6.32 8.47 7.93 8.45 7.50 10.47

b 4.97 5.28 5.07 5.95 6.85 5.22

c 2.74 4.23 3.88 5.55 4.64 4.46

261 Table 9 The channel s t r u c t u r e and f r a m e w o r k density of large pore molecular sieves containing 12-member rings Pore Framework No Code Zeolite Channel diameter density dimensionally (/k) (T/lO00/~a) 1 AFS MAPSO-46 3 6.3 • 6.3 13.7 2 FAU FAUJASITE 3 7.4 x 7.4 12.7 3 AFY COAPO-50 2 6.1 x 6.1 12.5 4 BEA BETA 2 6.4 • 7.6 15.0 5 BPH BERYLLOPHOSPHATE 2 6.2 • 6.7 16.4 6 EMT EMC-2 2 6.5 x 7.4 12.9 7 GME G M E L I N I T E 2 7.0 x 7.0 14.6 8 MEI ZSM- 18 2 6.9 x 6.9 14.3 9 OFF OFFRETITE 2 6.7 x 6.7 15.5 10 AFI ALPO-5 1 7.3 • 7.3 17.5 11 AFR SAPO-40 1 6.7 x 6.9 14.9 12 ATO ALPO-31 1 5.4 x 5.4 19.2 13 ATS MAPO-36 1 6.5 x 7.5 16.4 14 BOG BOGGSITE 1 7.0 x 7.0 15.6 15 CAN CANCRINITE 1 5.9 • 5.9 16.7 16 LTL LINDE-L 1 7.1 x 7.1 16.4 17 MAZ MAZZITE 1 7.4 x 7.4 16.1 18 MOR M O R D E N I T E 1 6.5 x 7.0 17.2 19 MTW ZSM-12 1 5.5 x 5.9 19.4 20 ROG ROGGIANITE 1 4.2 • 4.2 15.6

zeolites k n o wn in the l i t e r a t u r e is given in Table 9. Among these, excluding the n a t u r a l zeolites, a l u m i n o p h o s p h a t e structures, zeolites with several polymorphs, 7 zeolites are considered for detailed analysis. We compare here, the diffusion characteristics of the alkylbenzenes in 12-m channels of a variety of different siliceous zeolites. The molecular graphics picture in Figure 6 shows the three supercages in FAU and diffusion p a t h of the molecules from A to C, via B. Figure 6 includes the diffusion energy profile for p-IBEB. The m i n i m u m energy configuration of pIBEB in faujasite is also shown in Figure 6. It is clear from Figure 6 t h a t when the molecule moves from one supercage to a nothe r the molecule energetically prefers to be n e a r the wall r a t h e r t h a n at the center of the supercage. Its interaction energy decreases and becomes m i n i m u m w h e n the benzene ring of pIBEB is at the center of the 12-member ring. As the molecule diffuses towards the center of the second supercage from 12-m ring its interaction energy increases and again becomes m a x i m u m at the center of the second supercage. Due to this high energy conformer, a diffusional energy b a r r i e r exists for cage to

262

"8 E

-60 rr

uJ - 7 0 z LIJ

B C

Ic.) 9 r r~ - 9 0 ILl z -I00 -I10 11.4

22.8

DISTANCE TRAVELLED BYTHE MOLECULE(~.)

Figure 6. Variation of interaction energy ofp-IBEB with faujasite lattice during cage to cage diffusion through 12-m windows. The molecular graphics picture depicts the three supercages in different planes of the faujasite lattice. A typical minimum energy configuration of p-IBEB during the diffusion calculation is shown. A, B and C in the molecular graphics picture show the center points of the three supercages. In the graph, A, B, and C show the interaction energy values at the center points of three supercages.

cage diffusion of the molecule in faujasite. Similar diffusional profiles are calculated for EB, p-DEB, IBB, m-IBEB and o-IBEB and the overall behavior of the molecules are found to be the same. The diffusional energy barriers of all the alkylbenzenes are given in Table 10. In case of zeolite-L, the diffusion of the molecules along the 12-m channel is studied. The diffusion p a t h for the molecules is defined by a pair of points (A and B) on the channel axis at opposite ends of the section of the channel u n d e r investigation, as shown in Figure 7 and the diffusion calculations are carried out in the same way as in faujasite. The energy profiles for p-IBEB in zeolite L are shown in Figure 7 and the diffusional energy barriers for all the molecules are given in Table 10. A series of sharp peaks with high intensity is observed at regular intervals for m- and p-IBEB in zeolite L. These sharp peaks in zeolite L are due to the barrel shaped cage between two 12-m rings.

263 Table 10 Diffusional energy barriers in kJ/mol for different molecules in large-pore zeolites Alkylbenzenes Zeolites Faujasite Zeolite L Mazzite Mordenite EB 26.92 38.69 14.09 6.74 p-DEB 35.36 38.09 7.65 7.82 IBB 31.65 35.87 11.21 10.13 m-IBEB 28.38 43.69 9.78 17.95 o-IBEB 32.74 40.87 50.78 95.69 p-IBEB 31.65 28.43 10.69 6.44

12-m RING

E ..'< - B O

,,....

~L-~. c

B -'L BARREL

A

n~ IJJ

z tu - g o z o_ o a: - I 0 0 bJ p_z i

-I10

i

0

7.48

..

14.96

22.44

DISTANCE T R A V E L L E D BY THE MOLECULE (,~)

Figure 7. Variation of interaction energy of p-IBEB with zeolite L lattice during its diffusion along the c-axis through the barrel shaped cages. The molecular graphics picture shows the diffusion path of the molecule alongwith a typical minimum energy configuration ofp-IBEB during the diffusion.

Zeolite mazzite is another hexagonal but one dimensional zeolite with pore diameter of 7.4/k. Two types of smaller channels are present: the first consists of stacked gmelinite cages surrounded by six-member rings, the second is between two cross-linked rows of cages and is surrounded by eight-member rings. The calculated diffusional energy barriers for the molecules in mazzite are given in Table 10. The detailed studies on the influence of molecular dimensions of alkylbenzenes in diffusing the 12-m channel of mordenite has been the subject of

264 our previous theoretical studies [47]. In case of mordenite, an elliptical 12-m channel (6.5 /k x 7.0 /k) runs parallel to [001] and has small side pockets (2.6/k x 5.7 /k) parallel to the [010] direction which connect to the next 12-m channel. The calculated diffusional energy barriers of the molecules in mordenite are given in Table 10. It can be seen that the diffusional energy barriers for the isomers of IBEB are significantly different even though there are only small variation in the dimensions of the molecules. It is evident t h a t the diffusion of the r e a c t a n t molecules, namely EB, p-DEB and IBB have energy barriers of 6.74 kJ mol 1, 7.82 kJ mol 1 and 10.13 kJ mo1-1 respectively. The energy barrier for the diffusion of p-IBEB is also of the same order (6.44 k J mol~). However, there exists an energy barrier of 17.95 k J mo1-1 for m-IBEB and a significantly large energy barrier of 95.69 kJ mo1-1 for o-IBEB. 4.4. D i f f u s i o n a l b e h a v i o r of a l k y l a r o m a t i c s in O F F , M T W and CAN In case of offretite it can be observed t h a t there is an 8-m channel r u n n i n g perpendicular to the 12-m channel. The variation of the interaction energy between the molecules and the framework as well as the location and orientation of molecules were calculated as for the other zeolites. The diffusion energy barriers calculated from the energy profiles are given in Table 11. It is observed t h a t when the ethyl and isobutyl groups are nearer to windows, the interaction is unfavorable and when they are nearer to the surface the interaction is favorable. ZSM-12 has slightly elliptical pore of diameters 6.2 x 5.7/k. The molecules were allowed to diffuse through 5 unit cells in the b-direction. The variation of interaction energy between the molecules and the framework was studied. The diffusion energy profile showed a single m a x i m u m and a single m i n i m u m in each unit cell. The diffusion energy barriers calculated from the energy profiles for all the molecules are s u m m a r i z e d in Table 11. Next, the diffusion of the molecules in cancrinite was studied. The s y m m e t r y and a r r a n g e m e n t of the 12, 6 and 4-m channels in CAN are exactly same as in OFF. However, the pore diameter is considerably smaller (5.9/k) t h a n t h a t of OFF (6.7/k). The molecules were diffused through 5 unit cells along the c-direction. The variation of the interaction energy between the molecules and CAN as the molecules diffuse in c-axis was calculated. The diffusion energy profiles show two peak maxima within an unit cell. The diffusion energy barriers Table 11 Diffusional energy barriers in kJ/mol for different molecules in large-pore zeolites Alkylbenzenes Zeolites OFF MTW CAN EB 11.11 4.25 8.59 IBB 8.61 16.10 5.50 m-IBEB 30.00 38.21 81.51 o-IBEB 149.68 243.60 68.45 p-IBEB 15.43 19.13 14.65

265 calculated from the energy profiles are again summarized in Table 11. It is also observed from the absolute values that all the molecules are tightly fitting in CAN. 4.5.

Salient outcome

The molecule-zeolite interaction is studied elaborately using force field based energy minimization calculations. It is shown that the macroscopic diffusion behaviors could be understood from microscopic molecular level interactions. The major factors that influence the diffusion of the molecules inside the 12-m channel are i) pore diameter, ii) pore architecture iii) dimension of the molecules and iv) flexibility of the molecules, assuming the chemical composition of the zeolite frameworks are same. In addition to the above factors i) the nature of transition state complex ii) number and nature of tetra or trivalent cations isomorphously substituted in place of Si, iii) the number and nature of charge compensating extra framework cations, iv) temperature and v) pressure also influence the diffusion of the molecules in a minor way. Here, we concentrated on the major factors assuming that the dimension and flexibility of transition state complexes are in proportion to the product molecules. These results also provide the information on the nature of sites inside zeolites where the molecules have favorable and unfavorable interactions. From the results presented above the features of the diffusion of IBEB can be summarized as given below: i) Force field energy minimization calculation of the diffusion energy profile is a simple technique relative to experimental studies of diffusion and a reliable technique for logical catalyst screening. ii)The molecules prefer the configurations, where there is maximum interaction between the surface of the zeolite and the alkyl groups of the molecules. ill)The significant energy barriers exist for m- and o-IBEB in mordenite, o-IBEB in mazzite, and for none of them in faujasite and zeolite L. OFF and MTW are predicted to have good efficiency in the separation of o-IBEB from other isomers, whereas CAN can have good efficiency in the separation of all the three isomers. The actual values of interaction energy of different molecules with the zeolite framework depends on how good these molecules fit inside the channels.

5.

S E L E C T I V E C H L O R I N A T I O N O V E R Z E O L I T E K-L

5.1. B a c k g r o u n d 1,2,4-Trichlorobenzene (1,2,4-TCB) is a well-known termite exterminator [48], an additive in insulating and cooling fluids used for electrical engineering applications [49]. Conventionally, FeC13 is used as a catalyst for chlorinating 1,2-dichlorobenzene [1,2-DCB] to 1,2,4-TCB. But the process produces undesired 1,2,3-TCB (1,2,4-TCB/1,2,3-TCB = 2.2). Higher selectivity for 1,2,4-TCB (1,2,4-TCB/1,2,3-TCB-6.0) is achieved using solid acid catalyst, namely zeolite

266 K-L and it improves to -~ 14.0 in the presence of the promoter monochloroacetic acid[50].

5.2. A p p r o a c h In this study, we describe the application of the combination of molecular graphics, force field calculations and q u a n t u m chemical calculations to u n d e r s t a n d the m e c h a n i s m of selective chlorination of DCB to TCB over K-L zeolite and the role played by monochloroacetic acid as a promoter. The force field calculations were carried out as described in the previous section (4.2) to locate the favorable adsorption sites. Further, the electronic structure of the adsorbed complex are studied by q u a n t u m chemical cluster calculations. Although accurate ab initio calculations are desired, we opted for semi-empirical calculations since our interest was to derive electronic properties of a cluster model as large as SilsO48H24. In this study we used the PM3 Hamiltonian. The calculations were carried out using MOPAC version 6.0 program [18]. 5.3. Force field c a l c u l a t i o n s to locate favorable a d s o r p t i o n sites The molecular graphics picture of zeolite L lattice as viewed along the C-axis is shown in Figure 8. There are two crystallogrphically distinct t e t r a h e d r a l sites, namely, T1 and T2. The T1 sites are on the periphery of the 12-member window,

b

Figure 8. The molecular graphics picture of zeolite-L lattice. The 12-member channels run along the c-axis. There are two crystallographically distinct tetrahedral sites where Si or AI are located. They are shown as squares (Yl) and circles (T2).

267

b

t_c Figure 9. The initial (I) and final (F) positions of the diffusion path as well as minimum energy location for I,2-DCB in zeolite -L. The active site of zeolite -L which surrounds 1,2DCB is marked by a circle.

while Te sites are on the periphery of the barrel shaped cage. The location of DCB in zeolite L framework is not reported in the literature, although its location inside zeolite Y has been identified from synchrotron study [51]. We carried out force field calculations to identify the diffusion characteristics of 1,2-DCB in the fully siliceous zeolite L. The initial (I) and the final (F) points studied for the diffusion p a t h inside the channel along c-axis as well as a typical minimum energy location for the DCB are shown in Figure 9. The results indicate that the molecule is in a energetically favorable location when the phenyl ring is at the center of 8-member window which is perpendicular to the main channel. Similar calculations were carried out for the promoter-monochloroacetic acid and the reactant-chlorine to identify their minimum energy locations. It was found that monochloroacetic acid had a smaller diffusion energy barrier t h a n 1,2-DCB and chlorine showed several minima along the main channel indicating that there are many possible adsorption sites. It should be noted t h a t the minimum energy locations identified for these molecules correspond to single phase adsorption of these molecules. In order to study the influence of the intermolecular interaction between the molecules on the final locations of the promoter and reactants, a geometry optimization calculation was performed. These calculations were carried out by considering only van der Waals forces between the molecules and thus the final adsorption sites for 1,2-DCB, chlorine and monochloroacetic acid were derived. It was observed t h a t all these molecules could be accommodated inside a single "barrelshaped" cage. In Figure 9, the active site in zeolite which has maximum interaction with the molecule is highlighted as a circle.

268 5.4. Q u a n t u m c h e m i c a l c l u s t e r c a l c u l a t i o n s to u n d e r s t a n d t h e mechanism A cluster model containing all these atoms inside the circle is chosen for further electronic structure calculations. The valency of oxygens on the periphery of this cluster model is saturated by adding hydrogen atoms. The position of these hydrogen atoms are along the original O-Si vector with an O-H distance of 1.03/k. The cluster model chosen for quantum chemical calculations is shown in Figure 10. PM3 calculations were performed for the cluster model of the complex shown i n Figure 10 where all the three molecules are present inside zeolite L cage. Calculations were also carried out for cluster models where the individual molecules are adsorbed. These results are given in Table 12.

Cl

Cl

"

b

to Figure 10. The quantum chemical cluster model chosen to study the interaction between the promoter-monochloroacetic acid and the reactant-chlorine & 1,2-DCB.

The presence of monochloroacetic acid is found to cause a decrease of 3.0 to 3.5 A in the pore diameter of the zeolite L. Thus, the geometric restriction imposed by the presence of monochloroacetic acid is also a reason for improving the selectivity of 1,2,4-TCB. The total energy of the individual molecules are also given in Table 12. The adsorption energy values calculated for C12 and monochloroacetic acid are reasonable chemical values, whereas the adsorption energy of 1,2-DCB is an overemphasized value, which may be due to the inadequate representation of VDW forces between phenyl ring and the zeolite

269 Table 12 PM3 energy values calculated for the cluster model of LTL, relevant molecules, and the adsorption complexes Bare cluster or molecule Complex between zeolite cluster or molecule Zeolite cluster -15873.873 C12 -627.048 - 16503.602 C1CH2 COO H - 1187.877 - 17060.916 DCB -1405.441 -17255.000 C12 + DCB + C1CH2COOH -19072.532

lattice. I m p r o v e m e n t s in identifying the location and orientation of the molecules in terms of geometry optimization calculations are in progress. We analyzed the net charges on various atoms of the free molecule as well as in their adsorbed state. It is observed t h a t the molecules are more polarized in their adsorbed state, particularly the atoms closer to the zeolite framework. The net charge on various atoms of the zeolite cluster and their changes when the molecules are adsorbed were analyzed. It is observed t h a t the net charge on Si and 0 varies in the range of 1.43 to 1.49, a n d - 0 . 6 0 to -0.69, respectively. The s y m m e t r y related atoms have uniform charges and the Si atoms on the periphery of the cluster are less positive compared to those in bulk of the cluster. It is found t h a t the silicon atoms, r a t h e r t h a n oxygen atoms have electronic interactions with chlorine and monochloroacetic acid.

5.5. S a l i e n t o u t c o m e The efficacy of the combination of molecular modelling techniques in studying various aspects of the m e c h a n i s m of chlorination of 1,2-DCB have been brought out. Although both 3 and 4 positions of 1,2-DCB are a m e n a b l e for electrophilic chlorination, geometrical restrictions favor chlorination at 4-position. The 'barrel-shaped' cage in zeolite L is ideally dimensioned to accommodate the promoter and r e a c t a n t molecules. The promoter, monochloroacetic acid decreases the pore d i a m e t e r by 3.0 to 3.5 /k, thus creating the required geometric environment to achieve selectivity. The actual n a t u r e of electronic interaction between the zeolite framework and the molecules are brought out from the analysis of electron population on various atoms.

6. R E M O V A L O F NO BY S E L E C T I V E C A T A L Y T I C R E D U C T I O N 6.1. B a c k g r o u n d and m e t h o d of a p p r o a c h Pt-based composite catalysts supported on monoliths are conventionally used for the removal of NO from automobile exhausts. However, these three-way NO removal catalysts are weak to large excess of oxygen in s t a t i o n a r y exhausts such as diesel engines and lean-burn oil engines of power plants [52,53]. Hence, the

270 removal of NO by selective reduction with hydrocarbons over metal exchanged zeolite catalysts are attempted [54]. Zeolites are extremely suited since it stabilizes the metal under oxidative atmosphere by providing a suitable redox matrix. NO by itself and its oxidation products are potential pollutants which deteriorate our environment in several ways [52,53]. Recently, the direct decomposition of NO into N2 and 02 over Cu -ZSM-5 catalyst was reported by Iwamoto et al. [55]. The synergistic roles played by Cu and ZSM-5 framework for the reaction were emphasized in a detailed study [56]. The selective catalytic reduction of NO with different hydrocarbons such as methane, methanol, ethene, propene and propane has been tried using various metal-exchanged zeolites such as Ga-ZSM-5 [57-59], Ce-ZSM-5 [60], Co-ZSM-5 [61,62], Cu-ZSM-5 [63], Mn-ZSM5 [64], Ni-ZSM-5 [64], Pd-ZSM-5 [65], In-ZSM-5 [66], Cu-MOR [67], and CuSAPO-34 [68] as well as H-ZSM-5 [69]. However, final consensus on a successful catalyst formulation has yet to be evolved. In particular, much attention has been given to Ga ion-exchanged ZSM-5 introduced by Yogo and co-workers, [57,58] because of their high activity and selectivity with methane as the reductant, which otherwise was regarded as an inactive agent [70].

6.2. Approach Quantum chemical(QC) calculations based on density functional theory (DFT) were performed using the DMOL package of BIOSYM Technologies, Inc. The geometry optimization calculations were carried out using a minimal numerical basis set [71]. The total energy for the final optimized geometry was then evaluated using a double numerical polarization basis set [71]. A JMW local type functional [72] was used for the exchange-correlation energy terms in the total energy expression. The cluster model for QC calculations has been based upon the crystal structure of ZSM-5 reported by X-ray diffraction study [73] modified according to the equilibrium geometry given by MD simulation [54], where the T12 site was considered for aluminum substitution. Earlier quantum chemical studies [74] have also reported T12 as the energetically favorable site for the incorporation of aluminum. Figure 11 shows the conformation of the oxidized form of the Gaexchanged site, namely [GaO] § species within the zeolitic lattice as obtained from large model MD calculation [54]. In our QC calculations, a single A104 tetrahedron was considered as the representative of the active site, with adjacent silicons replaced by hydrogens and charge deficiency compensated by [GaO] § cationic species. We believe that the cluster model specified above is adequate to reproduce basic features of the zeolitic active site and local electronic structure. At the same time, the modest size of the cluster allows for meeting more rigorous demands of QC methodology and testing various computational parameters.

271 A1 Si

V

"

0 2.

o

Figure 11. The structure of [GaO]-ZSM-5 as derived from molecular dynamic calculations at 300 K after 5000 time steps of 2.5 • 1015 seconds. A1 is substituted in place of Si at T12 site.

6.3. Location of extra framework gallium The way in which the gallium ions are dispersed in Ga-ZSM-5 is expected to be the cause of catalytic activity. These materials have been simulated in order to u n d e r s t a n d the salient features of this catalyst which give it its high activity. We also simulated the Ga203 lattice. The striking difference is the high coordination, namely six coordination for Ga in the bulk Ga203. low and coordination namely three coordination for gallium ions in Ga-ZSM-5. Various gallium species such as Ga 3+, [GaO] +, [Ga(OH)] 2+ and [Ga(OH)2] + were simulated inside ZSM-5. [Ga(OH)2] + is the likely cationic species in the h y d r a t e d form. The dehydration of gallium hydroxide at 600 K to form gallium oxide occurs as follows: [Ga(OH)2] §

-

-

[GaO] § + H20

It was observed t h a t the [GaO] § cation was more dynamic t h a n the hyroxylated gallium ions and both of them had more mobility t h a n the Ga § ions on the surface or bulk of the Ga203. The computer graphics illustrations shown in Figure 12 show the structure of [Ga(OH)2] § ion in Ga-ZSM-5. P r e l i m i n a r y studies on the behavior of water molecules showed t h a t water undergoes preferential adsorption on [GaO] § ion leading to [Ga(OH)2] § causing the reversible reaction in the above equation. The adsorption behavior of water is described in detail in the following section. It was found t h a t the oxygen attached to Ga § h a d the most mobility indicating the easier approach ability and less steric hindrance of gallium to NO and hydrocarbons.

272

A1

0 2-

H+

Si

O

Figure 12. The structure of [Ga(OH)2]-ZSM-5 as derived from molecular dynamic calculations at 300 K after 5000 time steps of 2.5 x 1015 seconds. A1 is substituted in place of Si at T12 site.

H

o

,* O

O

Figure 13. Cluster models considered to represent the active site [GaO]-ZSM-5 with the molecular formula of GaO-[AI(OH)4]. A single (a), double (b) and triple (c) coordination of Ga to the oxygens in the [AI(OH)4]- unit are considered.

As molecular dynamics can not provide a quantitative description of chemical bonding, the final coordination of [GaOl unit to the zeolite framework has been verified by q u a n t u m chemical DFT calculations. [GaO] can coordinate to the [AI(OH)4] -1 t e t r a h e d r o n in three different ways. They are single coordination, double coordination in a bridging position between two oxygen atoms, and triple coordination in a three-fold site between three framework oxygens as shown in Figures 13 a, b and c, respectively. The DFT total energy minimization for the

273 three configurations showed [75] that the doubly co-ordinated bridging site was the most stable structure and that this should be the most favorable coordination mode for the ion-exchanged cation.

6.4. A d s o r p t i o n o f CH4 over [GaO]-ZSM-5 In general, selective C-H activation of saturated alkanes by various catalytic systems is an extensively studied subject due to its importance in the transformation of the relatively inert alkanes to more useful products. Among other catalytic systems cation-exchanged zeolites have received specific attention due to their valuable catalytic properties in an environmentally important process of the removal of nitrogen oxides from exhaust gases, as mentioned above. The mechanism by which the reduction of nitrogen oxides by hydrocarbons proceeds is very complicated as it involves many intermediate steps and not their nature but even the sequence of elementary steps is still far from understood. Here, we describe our QC calculations on the adsorption of various molecules over [GaO] § exchanged ZSM-5. In the first step, the geometry optimization calculations on the methane molecule in the vicinity of the Ga site in the cluster model shown in Figure 14 were carried out. We located the position of weakly adsorbed methane.

t3a

H

Figure 14. The optimized geometry of the GaO-[AI(OH)4] in its double coordinated mode, which is the energetically favored one. This cluster model was used for studying the interaction of methane and water molecules.

274 Figure 15 shows the optimized configuration of a methane molecule adsorbed on the Ga site and displays the dissociation of one of the C-H bonds in m e t h a n e compared to t h a t of the gas phase molecule. It can be seen from Figure 15 t h a t the [GaOl + active site had enough polarizing ability to dissociate an approaching m e t h a n e molecule. We have studied two possible routes of the m e t h a n e dissociation. The first - Figure 15a leads to a methyl attached to the Ga a+ ion and hydrogen forming a hydroxyl group with the extraframework oxygen. The second - Figure 15b assumes the reverse attachments, namely H attached to the Ga 3+ ion and CH3 connected to the extraframework oxygen to make a methoxy group. The calculated adsorption energy values are given in Table 13. (b)

(a) O H

9-..J QQ%

A1C3~Ga

O CH3 ~~..

"

H

'~

~~CH3

Figure 15. Two possible dissociative modes of adsorption of methane on GaO-[AI(OH)4]. Quantum chemical geometry optimization calculations led to these two configurations. The configuration with Ga-CH3 and O-H (a) is energetically more favorable than the configuration with Ga-H and O-CH3.

Table 13 Adsorption energy values of guest molecules on [GaO] site. The adsorption energy has been estimated according to the equation: Eadsorption = Ehost 4- Eguest - (Ehost: guest complex), where Ehost, Eguest and (Ehost: guest complex) denote the total energy of the zeolite cluster system, the incorporated guest molecule and the complex formed by the zeolite cluster:guest molecule, respectively Adsorption Host cluster Guest molecule energy (kcaYmol) GaO-ZSM-5 GaO-ZSM-5 GaO-ZSM-5

CH4 dissociatively adsorbed into Ga-CH3 a n d - O H CH4 dissociated into Ga-H and O-CH3 H20 dissociative adsorbed

-63.0 -31.4 -77.4

275 The optimized geometry for the transition state was also analyzed. The saddle point character of the transition state was confirmed by the analysis of the Hessian restricted to optimized coordinates. The only negative frequency corresponded almost exclusively to the in-plane movement of the dissociated hydrogen atom between its neighboring carbon and oxygen atoms, supplemented by the Ga-O stretch. Thus, the reaction coordinate was chosen based on these results. 6.5. A d s o r p t i o n o f H20 a n d its i n f l u e n c e on t h e m e t h a n e a c t i v a t i o n process We studied the approach of water molecule to the [GaOl § site, by geometry optimization calculations. Water hydrates the [GaO] § site to [Ga(OH)2] + reversibly as shown in the equation in section 6.3. It was suggested that the nucleophilic and electrophilic region around the [GaO] § site might be significantly changed and the physical hindrance might be imposed due to hydration. The water molecule effectively adsorbs on Ga via an oxygen atom and consequently, one of its O-H bond is broken leading to increased coordination to gallium. The adsorption energy of water molecule is more favorable than that of methane molecule as shown in Table 13. This result reveals that the physical hindrance may be imposed by water molecules for the reactant methane molecule and the nucleophilicity around Ga may be also decreased. Thus, the structural and electronic factors explain the reasons for high sensitivity of Ga-ZSM-5 towards water. 6.6. S a l i e n t o u t c o m e Selective reduction of NO using hydrocarbons over Ga exchanged ZSM-5 is studied using molecular modelling tools. Specifically, the results of mechanistic aspects over Ga-ZSM-5 are reported here. The nature, co-ordination and the location of Ga are analyzed. It is noted that partially u n s a t u r a t e d coordination of oxygen in [GaO] § species increases it dynamic nature inside the pores of ZSM-5. Adsorption studies of CH4 revealed the various modes of adsorption of CH4. The dissociative mode where -CH3 a n d - H are attached to Ga and O of [GaO] § species, respectively was found to be the favorable mode. Dissociative adsorption of water is even more favorable t h a n that of CH4 and leads to the conversion of [GaO] § into [Ga(OH)2] § thus saturating the coordination of Ga and O as well as reducing their dynamic nature.

7. A D S O R P T I O N OF C F C s IN F A U J A S I T E Z E O L I T E 7.1. B a c k g r o u n d a n d m e t h o d of a p p r o a c h Chlorofluorocarbons (CFCs) are widely used as solvents, refrigerants, foam rubber blowing agent, and propellants. It is well known that the emission of CFCs lead to global environmental problems such as the stratospheric ozone layer depletion and causing green house effects. Although Molina and Rowland

276 [76] had pointed out the destruction of the ozone layer in 1974 itself, after nearly decade, F a r m a n and coworkers [77] discovered the CFCs as the cause for the destruction of the ozone layer. Since then there are attempts to catalytically decompose or to trap them inside suitable adsorbents [78-89]. There are previous experimental work on CFC-zeolite interactions but concentrating on catalyst regeneration [90,91] on the dehydroflurination reaction[92] and on the dynamics of the CFCs [93]. Traditionally the active carbon was as employed adsorbent for CFCs from refrigerators or gas stations. However, there are difficulties in handling, desorbing and recycling, due to unstable nature. Recently there are reports indicating zeolites as novel candidate materials for adsorbing CFCs. Alkali ion exchanged Y zeolites are extensively studied by Kobayashi et al. [88,89]. In this background, we studied the adsorption behavior of dichlorodifluromethane (CF2C12) in CsNaY using Grand Canonical Monte Carlo (GCMC) method [94]. 7.2.

Detailed distribution

o f A13+, Cs § a n d N a § in C s N a Y z e o l i t e

Although the SiO2 framework structure of Y type zeolite is well known, the fine details of the structure of CsNaY (Si/A1 = 2.43), such as the distribution and location A13§ Cs § Na § have not been established. We determined the A1 distribution in the NaY by using molecular dynamics simulations [95], to reproduce the occupancies and sites of Na § cations obtained experimentally by Fitch and coworkers [96]. These findings were further confirmed by NMR simulations shown in Figure 16. 29Si MAS NMR spectrum was simulated for the

c~

8g

90

I gg

I Ig

12g

-Chem Shl F~c(ppm)

Figure 16. The simulated 29Si MAS NMR spectra of our estimated aluminum distribution model in the zeolite NaY with Si/A1 = 2.43. This simulated spectrum reproduce the one observed by experimental report [97]. This distribution of aluminum was derived from several trial and error MD calculations.

277 'AI' distribution predicted by MD, using 'Catalysis' module in the BIOSYM molecular modelling package. Since the calculated chemical shifts were in good a g r e e m e n t with the experimental spectrum by Klinowski et al. [97], we determined the structure of the CsNaY (Si/A1 = 2.43) on the basis of the NaY structure. There are several sites for the e x t r a f r a m e w o r k cations in Y type zeolites. The locations of these cationic sites are shown in Figure 5. In our Na56Si13c~k15603s4 model 32 Na § cations are located in type II sites, 16 Na § cations in type I' sites, and 8 Na § cations in type I sites. As discussed earlier (section 3.2) Na in the type II sites can be easily exchanged by Cs § cations as compared to the Na § cations in the site I and I', because only the cations in the site II are exposed to the supercage of the CsNaY and experimentally CsNaY is p r e p a r e d by the Cs ionexchange of the NaY. The structure of Cs32Na24Si136A1560384 around a super cage is shown in Figure 17.

Cs

INa

Figure 17. The structure of CsNaY with Si/A1 = 2.43. The exchanged cations are shown as solid spheres, whereas the zeolite framework is shown as stick model. The large spheres are Cs and the smaller spheres are Na.

7.3. Adsorption of CF2C12 in CsNaY by GCMC simulation We performed GCMC calculations at various pressures of the CF2C12, in order to clarify the atomistic mechanism of the CF2C12 adsorption in the CsNaY (Si/A1 = 2.43). It was shown [90] t h a t the simulated adsorption isotherm of the CF2C12 in the CsNaY is in a g r e e m e n t with the earlier experimental results [88,89]. In the present GCMC calculations, the interatomic potential p a r a m e t e r s derived from

278 DFT calculations were used. Thus a calculation procedure to reproduce and predict the adsorption isotherm of CFCs in zeolites, was established. In order to discuss the atomistic adsorption mechanism of the CF2C12 in the CsNaY, the adsorption sites and distributions of the CF2C12 in the CsNaY at various pressures were investigated. The results of simulation were visualized using the computer graphics (CG). Figure 18 represents a CG image of the adsorbed CF2C12 in the CsNaY at low pressure of 0.001 kPa. It was noted t h a t the locations of CF2C12 were very regular and systematic. The molecules occupied sodalite cages. This phenomenon can be related to the position of the Cs § cations; the CF2C12 interacts with Cs § cations of the CsNaY and not with the SiO2 of the zeolite framework. Thus the strong interaction between the Cs § cation and the F & C1 atoms, in zeolite and the molecules, respectively are the reasons of this selective adsorption of the CF2C12. The reason of the selective adsorption can also be explained by the fact t h a t the location of the Cs § cations in type II sites allow a direct interaction with the CF2C12 molecules entering through the supercage of the CsNaY.

Figure 18. The results of GCMC calculations at a pressure of 0.001kPa. The CF2C12 molecules, exchanged cations and the zeolite framework are shown as solid spheres, dots and stick models, respectively. CF2C12prefers the ordered packing inside sodalite cages.

279

Figure 19. The results of GCMC calculations at a pressure of 0.007 kPa. The CF2C12 molecules, exchanged cations and the zeolite framework are shown as solid spheres, dots and stick models, respectively. CF2C12molecules prefer to agglomerate with each other.

CG image of the CF2C12 in the CsNaY at higher pressure of 0.007 k P a is shown in Figure 19. It can be observed t h a t the space for the adsorption of the CF2C12 around the Cs cations are already filled and the CF2C12 molecules start to aggregate with each other in the supercage of the CsNaY. It is revealed t h a t the adsorption of the CF2C12 at higher pressure follows a different mechanism compared to t h a t at low pressure. George et al. [98] reported the sorption energy of CF2C12 in zeolite-Y as a part of their more general study to explore the possibility of zeolites as scrubbers for CFC. They also reported a significant difference in the binding energy values for the guest molecules when they are located inside the sodalite cage and the supercage; the difference can be up to 10 kcal/mol. 7.4. S a l i e n t o u t c o m e The detailed structure of Cs32Na24Sil~15603s4 has been worked out. The results of GCMC calculations suggest t h a t the location and the a m o u n t of the exchanged cation in zeolite-Y have a profound influence on the adsorption site and a m o u n t of CFCs. This is an information which is helpful to design novel adsorbents for CFCs with high efficiency and ability. Thus the cage effect in

280 addition to the cation effect is also brought out. The most exciting outcome of our study is the aggregate of molecules formed at higher pressures and the potential for CsNaY as a sorbent for CF2C12.

8.

CONCLUSIONS

The value of zeolite catalysts for the synthesis of fine chemicals are brought out. Novel cyclization and rearrangement reactions occur in a single step with very high selectivity inside zeolite cages. Computational studies by molecular graphics and molecular fitting procedure allow one to derive a set of desirable criteria expected in the catalyst for a specific organic reaction and the nature of substituent required in the reactants. This is a computationally efficient procedure to be adopted as pre-experimental activity. The diffusion calculations are described, which are again beneficial for the screening of zeolites in order to achieve the desired selectivity. The studies on alkylaromatics are detailed and the predictions have been made for zeolite systems, where there are no experimental reports. Additionally, they also provide an understanding of the adsorption sites and adsorption processes. The studies on selective chlorination confirmed the experimentally observed behavior and threw light on the role of promoters. The studies on deNOx reaction mechanism and CFC adsorption mechanism over zeolites have brought out the power of highly accurate computations to provide information on novel catalyst systems for pollution abatement.

ACKNOWLEDGEMENTS We thank Prof. C. N. Pillai, Dr. S. Rajappa and Dr. A.P. Singh for fruitful discussions. One of us (RCD) thank CSIR, New Delhi for financial support in the form of senior research fellowship.

REFERENCES 1. 2. 3. 4. 5. 6. 7.

M. Karpf and A.S. Dreiding, Helv. Chim. Acta., 64 (1981) 1123. P.E. Eaton, Tetrahedron, 37 (1981) 4479. Y.V.S.N. Murthy and C.N. Pillai, Tetrahedron, 48 (1992) 5331. P.E. Eaton, G.R. Carlson and J.T. Lee, J. Org. Chem., 38 (1973) 4071. Y.V.S.N. Murthy and C.N. Pillai, Tetrahedron Lett., 31 (1990) 6067. C.R. Mordi, R. Fields and J. Dwyer, J. Mol. Catal., 92 (1994) 51. U. Kameswari, C.S. Swamy and C.N. Pillai, Stud. Surf. Sci. Catal., 84 (1994) 1959. 8. M. Ziolek and P. Decyk, Stud. Surface Sci. Catal., 84 (1994) 1579. 9. R. Sreekumar, Y.V.S.N. Murthy and C.N. Pillai, J. Chem. Soc., Chem. Commun., 1624 (1992).

281 10. R. Sreekumar, Y.V.S.N. Murthy and C.N. Pillai, in: Catalysis-Present and Future, P. Kanta Rao, R.S. Beniwal (eds.), Wiley Eastern Ltd., New Delhi, 1995, 99. 11. D.M. Ruthven, Am. Chem. Soc. Symp. Ser., 40 (1977) 320. 12. E.M. Flanigen, J.M. Bennett, R.W. Grose, J.P. Cohen, R.L. Patton and R.M. Kirchner, Nature, 271 (1978) 512. 13. D.W. Breck, Zeolite Molecular Sieves; Structure, Chemistry and Use, Wiley, New York, (1974) 634. 14. A. Chatterjee and R. Vetrivel, J. Chem. Soc., Faraday Trans., 91 (1995) 4313. 15. A. Chatterjee, R. Vetrivel, R. Sreekumar, Y.V.S.N. Murthy and C.N. Pillai, in: Catalysis-Modern Trends, N.M. Gupta, D.K. Chakrabarty (eds.), Narosa Publishing House, New Delhi, 1996, 68. 16. A. Chatterjee, R. Vetrivel, R. Sreekumar, Y.V.S.N. Murthy and C.N. Pillai, J. Mol. Cat. A: Chemical (in press). 17. W.M. Meier, D.H. Olson and Ch. Baerlocher, Atlas of Zeolite Sructure Types, Zeolites, 17 (1996) 1. 18. J.J.P. Stewart, MOPAC Manual, (Version 6.0) Quantum Chemistry Program Exchange - Program No. 539, Indiana University, Bloomington, 1990. 19. M.J.S. Dewar, E.G. Zoebisch, E.F. Healy and J.J.P. Stewart, J. Am. Chem. Soc., 107 (1985) 3902. 20. A.N. Fitch, H. Jobic and A. Renouprez, J. Phys. Chem., 90 (1986) 1311. 21. H. Himei, E. Maruya, M. Yamadaya, M. Kubo, R. Vetrivel and A. Miyamoto, J. Clay Sci. Soc. Jpn., 34 (1994) 125. 22. A.R.A.S. Deshmukh, T.I. Reddy, B.M. Bhawal, V.P. Shiralkar and S. Rajappa, J. Chem. Soc., Perkin Trans. I, (1990) 1217. 23. T.I. Reddy, B.M. Bhawal and S. Rajappa, Tetrahedron, 49 (1993) 2101. 24. G.R. White, Ger. Often., 2 621 092 (1976). 25. M.J. Daly, J.M. Humphray and R. Stables, J. Pharmacol., 72 (1981) 49. 26. S. Rajappa and M.D. Nair, Advances in Heterocyclic Chemistry, A.R. Katritzky and A.J. Boulton (eds.) 25 (1979) 113. 27. S. Rajappa, Tetrahedron, 37 (1981) 1453. 28. B.R. Gelin and M. Karplus, Biochemistry, 18 (1968) 1256. 29. B.M. Bhawal, R. Vetrivel, T.I. Reddy, A.R.A.S. Deshmukh and S. Rajappa, J. Phys. Org. Chem., 7 (1994) 377. 30. W.J. Mortier, H.J. Bosmans and J.B. Uytterhoeven, J. Phys. Chem., 76(1972) 678. 31. D.W. Breck, J. Chem. Educ., 48 (1978) 678. 32. P.B. Venuto and P.S. Landis, Adv. Catal., 18 (1968) 259. 33. V.R. Choudhary and D.B. Akoleker, J. Catal., 117 (1989) 542. 34. R. Szostak, Handbook of Molecular Sieves, Van Nostrand Reinhold, New York, 1992. 35. F.C. Friedel and J.M. Crafts, Bull. Soc. Chim. Fr., 27 (1877) 530. 36. F.A. Drahowzal, Friedel-Crafts and related reactions, Vol. 2, G.A. Olah (ed.), Wiley, New York, (1964) 446.

282 37. R.M. Roberts and A.A. Khalaf, Friedel-Crafts alkylation chemistry, A century of discovery, Marcel Dekker, New York, 1982. 38. P.B. Weisz, New Horizons in Catalysis, T. Seiyama and K. Tanabe (eds.), Elsevier, Amsterdam, 7 (1981) 3. 39. P. B. Weisz, Chemtech., 3 (1973) 498. 40. E. G. Derouane, Catalysis by zeolites, B. Imelik, C. Naccache, Y. B. Taarit, J. C. Vedrine, G. Coudurier and H. Praliaud (eds.), Elsevier, Amsterdam, 5 (~980) 5. 41. J. -P. Rieu, A. Boucherle, H. Cousse and G. Mouzin, Tetrahedron, 42 (1986) 4095. 42. I. Shimizu, Y. Matsumura, Y. Tokumoto and K. Uchida, Eur. Pat. Appl., EP 373 362, June 20, 1990. 43. Y. Tokumoto, I. Shimizu and S. Inoue, Eur. Pat. Appl., EP 414 207, Feb. 27, 1991. 44. J.A. Horsely, J.D. Fellmann, E.G. Derouane and C.M. Freeman, J. Catal., 147 (1994) 231. 45. R. Millini and S. Rossini, Stud. Surf. Sci. Catal., 105 (1996) 1389. 46. T. Kiyoura, J. Ishikawa and H. Phala, J. Catal., 170 (1997) 204. 47. R. C. Deka and R. Vetrivel, Chem. Commun., 2397 (1996). 48. J.E. Milan and W.E. Wimer, US Patent No. 4835327 (1989). 49. U. Beck, in: Ullmann's Encyclopaedia of Industrial Chemistry, W. Gerhartz, Y.Y. Stephen, F.T. Campbell, R. Pfeffercorn and J.F. Rounsaville (eds), VCH, Weinheim, A6 (1986) 340. 50. S. Sharma, Ph. D. Thesis, University of Poona, Pune, India, 1997. 5 I. Z.A. Kaszkur, R.H. Jones, J.W. Couves, D. Waller, C.R.A. Catlow and J.M. Thomas, J. Phys. Chem. Solids, 52 (1991) 1219. 52. J.N. Armor, Appl. Catal., B 1 (1992) 221. 53. F. Jansen and R. Meijer, Catal. Today, 16 (1993) 157. 54. A. Miyamoto, H. Himei, Y. Oka, E. Maruya, M. Katagiri, R. Vetrivel and M. Kubo, Catal. Today, 22 (1994) 87. 55. M. Iwamoto, H. Furukawa, Y. Mine, F. Uemura, S. Mikuriya and S. Kagawa, J. Chem. Soc., Chem. Commun., (1986) 1272. 56. M. Iwamoto, H. Yahiro and K. Tanda, Successful Design of Catalysts, T. Inui (ed), Elsevier, Amsterdam, 1989, 219. 57. K. Yogo, S. Tanaka, M. Ihara, T. Hishiki and E. Kikuchi, Chem. Lett., (1992) 1025. 58. K. Yogo, M. Ihara, I. Terasaki and E. Kikuchi, Chem. Lett., (1993) 229. 59. Y. Li and J.N. Armor, J. Catal., 145 (1994) I. 60. M. Misono and K. Kondo, Chem. Lett., (1991) I001. 61. M. Kishida, T. Tachi, H. Yamashita and H. Miyadera, Shokubai, 34 (I 992) 148. 62. Y. Li and J.N. Armor, Appl. Catal., B 1 (1992) L31. 63. S. Sato, Y. Yu-u, H. Yahiro, N. Mizuno and M. Iwamoto, Appl. Catal., 70 (1991) LI.

283 64. 65. 66. 67. 68.

Y. Li and J.N. Armor, Appl. Catal., B 2 (1992) 239. V. Nishizawa and M. Misono, Chem. Lett., (1993) 1295. E. Kikuchi and K. Yogo, Catal. Today, 22 (1994) 73. G. Mabilon and D. Durand, Catal. Today, 17 (1993) 285. T. Ishihara, M. Kagawa, Y. Mizuhara and Y. Takita, Chem. Lett., (1992) 2119. 69. Y. Kintaichi, H. Hamada, M. Tabata, M. Sasaki and T. Ito, Catal. Lett., 6 (1990) 239. 70. X. Zhang, A.B. Walters and M.A. Vannice, J. Catal., 146 (1994) 568. 71. B. Delley, J. Chem. Phys., 90 (1990) 508. 72. L. Hedin and B.I. Lundqvist, J. Phys. C, 4 (1971) 2064. 73. E.M. Flanigen, J.M. Bennett, R.W. Gorse, J.P. Cohen, R.L. Patton, R.M. Kirchner and J.V. Smith, Nature, 271 (1978) 512. 74. E.G. Derouane and J.G. Fripiat, Zeolites, 5 (1985) 165. 75. R. Vetrivel, M. Kubo, H. Himei, E. Maruya, M. Katagiri, E. Broclawik and A. Miyamoto, Stud. Surf. Sci. Catal., 92 (1995) 233. 76. M.J. Molina and F.S. Rowland, Nature, 249 (1974) 810. 77. J.C. Farman, B.G. Gardiner and J.D. Shanklin, Nature, 315 (1985) 207. 78. J.I. Darragh, GB Patent No. 1 578 933 (1980). 79. A.E. Feiring, US Patent No. 4 258 225 (1981). 80. A.E.Feiring, J. Fluorine Chem., 25 (1981). 81. S. Imamura, T. Shiomi, S. Ishida, K. Utani and H. Jindai, Ind. Eng. Chem. Res., 29 (1989) 810. 82. J.L. Graham, D.L. Hall and B. Dellinger, Env. Sci. Technol., 20 (1986) 703. 83. J. Skalny, V. Sobek, P. Lukac and A. Varga, Proc. 9th Int. Symp. Plasma Chem., (I 989) 810. 84. A.P. Hangen and E.A. Elphingstone, J. Inorg. Nucl. Chem., 36 (1974) 509. 85. R. Ohnishi, W.L. Wang and M. Ichikawa, Appl. Catal. A, 113 (1994) 29. 86. Y. Takita, T.Imamura, Y. Mizuhara, Y. Abe and T. Ishihara, Appl. Catal. B. 1 (1992) 79. 87. S. Chandrasekar, K.S. RamaRao, P.S. SaiPrasad, R. Srinivas and P. Kanta Rao, Proc. 13 th National Symposium on Catalysis, 2-4, April 1997, Indian Institute of Petroleum, Dehradun, India, Stud. Surface Sci. Catal., (in press). 88. S. Kobayashi, K. Mizuni, R. Aizawa, S. Kuahiyama, Y. Koinuma and H. Ohuchi, Ind. Eng. Chem. Res., 30 (1991) 2340. 89. S. Kobayashi, S. Kuahiyama, R. Aizawa, K. Mizuni, H. Ohuchi and M. Shinbo, Nippon Kagaku Kasishi, (1994) 828. 90. M. Sugimoto, T. Murakawa, T. Hirano and H. Ohashi, Appl. Catal. A, 95 (1993) 257. 91. M. Sugimoto, H. Katsuno, T. Hyasaka, K.I. Hirasawa and N. Ishikawa, Appl. Catal. A, 106 (1993) 19. 92. C.P. Gry and D.R. Corbin, J. Phys. Chem., 99 (1995) 16821.

284 93. T.J. Udovic, J.M. Nicol, P.R. Cavanagh, J.J. Rush, M. Crawford, C. Grey and D.R. Corbin, Abstracts of Papers of the American Chemical Society Meeting, 208 (1994) 362. 94. K. Mizukami, H. Takaba, Y. Oumi, M. Katagiri, M. Kubo, A. Stiriling, E. Broclawik and A. Miyamoto, Stud. Surf. Sci. Catal., 105 (1997) 1811. 95. H. Himei, M. Yamadaya, Y. Oumi, M. Kubo, A. Stirling, R. Vetrivel, E. Broclawik and A. Miyamoto, Microporous Materials, 7 (1996) 235. 96. A.N. Fitch, H. Jobic and A. Renouprez, J. Phys. Chem., 90 (1986) 1311. 97. J. Klinowski, S. Ramdas, J.M. Thomas, C.A. Fyfe and J.S. Haetman, J. Chem. Soc., Faraday Trans. 2, 78 (1982) 1025. 98. A.R. George, C.M. Freeman and C.R.A. Catlow, Zeolites, 17 (1996) 466.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

285

M o d e l i n g of m e t a l i o n s o r p t i o n p h e n o m e n a i n e n v i r o n m e n t a l s y s t e m s S. Yiacoumi and J. Chen School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0512, USA

A systematic approach for modeling metal ion sorption phenomena occurring in aqueous solutions is presented here, with a discussion on environmental applications. Three different models are introduced: the surface complex formation model (SCFM) to describe sorption equilibrium; the KINEQL model to describe sorption kinetics; and the HYDROGEOCHEM model to describe transport of metal ions in treatment systems, as well as in natural subsurface systems. 1. I NTR ODUCT I O N Sorption or adsorption is a gas-solid or liquid-solid phenomenon defined as the accumulation of particular component(s) at the surface between the two phases. Unbalanced forces of attraction between the gas or liquid and solid phases result in an increase of concentration of the particular component(s) on the solid phase. Sorption can be categorized into physical sorption and chemisorption based on the strength of these forces. Physical sorption involves only relatively weak forces, while in chemisorption a chemical bond is formed between the sorbate component(s) and the sorbent components on the solid surface. In the past several decades, a dramatic increase of metal contaminant volume has posed many serious environmental problems. The most common treatment processes like precipitation and ion-exchange are usually effective in reducing the extent of contamination, but are not economical. Removal by various sorbents, such as activated carbon, has emerged as one of the most effective technologies for removing organic and inorganic pollutants from water and wastewater [1]. The preparation, disposal and recycle of sorbents and ion-exchange resins, however, need a substantial amount of energy. Searching for cost-effective sorbents, as well as modeling sorption processes, has become the focus of attention of many researchers. One promising technique to accumulate metals is by using biopolymers and non-living organisms as sorbents [2,3]. Biopolymers are extracted from and have common chemical properties with non-living organisms. It has been well documented that biopolymers possess a high potential to sequester and accumulate inorganic ions present in aqueous

286 solutions. Studies on biopolymers such as calcium alginate by Chen et al. [3] showed that after usage, the volume of calcium alginate may decrease dramatically by drying the beads. Furthermore, used beads, after being contacted with a weakly acidic solution, can be reused and their effectiveness for metal ion removal is comparable to that of fresh beads. Metal ion waste streams from various sources also contaminate natural surface and subsurface systems. As a result, contamination problems of surface water and groundwater have increased significantly in the past two decades [46]. Forecasting water pollution by mathematical models, as well as searching for remediation techniques, is necessary for making strategies to meet challenges in the future. Among several mechanisms, sorption plays an important role in the transport of contaminants in natural systems. In addition, sorption of metal ions by naturally occurring particles can change the surface properties of particles, which in turn affects the stability and transport of these particles in natural systems. Recent studies show that the change of surface properties of particles also results in changes in particle flocculation kinetics [7]. Metal ion sorption experiments in batch and fixed-bed columns reactors and modeling of metal ion sorption equilibrium, kinetics and transport in natural and treatment systems have been the focus of many researchers in the last several decades. This chapter summarizes the findings from experimental studies and presents a systematic modeling procedure for describing metal ion sorption and transport in one-dimensional fixed-bed and subsurface systems. In all the cases examined in this work, sorption and transport are studied in aqueous solutions.

2.

METAL ION SORPTION EQUILIBRIUM

In this section, factors that affect metal ion sorption equilibrium are first summarized, followed by a brief introduction of equilibrium models. The surface complex formation model, as well as examples to illustrate the application of this model, is presented.

2.1. Factors i n f l u e n c i n g metal ion sorption equilibrium Several studies have shown that sorption of metal ions from aqueous solutions is strongly pH dependent [8,9]. An increase of the solution pH results in a decrease of surface charge and an increase of negatively charged sites and eventually an increase of metal ion binding [10,11]. Many studies demonstrated that metal ion sorption by activated carbon, hydrous oxides and biopolymers increases with increasing pH [8,11-16]. Chen et al. [11] reported that sorption of copper ions strongly depends on solution pH and increases from 10 to 95 % in pH ranging from 2.3 to 8. A dramatic change in pH and emission of small gas bubbles were observed during the experiments, which may result from sorption of hydrogen ions and/or redox reactions. Reed and Matsumoto [17] also showed that the sorption edge for cadmium ions ranges from

287 pH 3 to 9. Chen et al. [3] demonstrated that sorption of copper ions by calcium alginate strongly depends on solution pH; the metal ion binding increases from 0 to 90 % in pH ranging from 1.5 to 5.0. Rao et al. [18] studied the sorption of Cu 2§ by G. lucidum and A. niger at an initial concentration of 0.5 mM. They found that the binding had an increasing trend in pH 2 to 6, with the maximum between pH 5 and 6. For some types of sorbents, pH may play a different role in sorption, depending on the initial metal ion concentration. Hao et al. [19] reported that when the initial concentrations of Cu 2§ Pb 2§ Zn 2§ Cd 2§ and Cr 2§ were 10 ppm and U. lactuca was used as sorbent, the binding of metal ions was independent of pH. When the concentrations of these metal ions were 100 ppm, however, the sorption increased with increasing pH. Ke et al. [20] also reported similar results for the Ag § sorption by Datura cells. These studies suggest that at least two binding sites are involved: one site is pH-independent and displays a greater affinity with a lower availability than the other site, which is pH-dependent. Ionic strength also plays an important role in sorption of metal ions. In the sorption by activated carbon and hydrous oxides, this effect is quite different as compared to that by biopolymers. The metal ion sorption increases with an increase in ionic strength when activated carbon and hydrous oxides are used [8,10,14]. Corapcioglu [21] demonstrated that the surface charge of activated carbon decreases with increasing ionic strength. These two observations result from the compression of the electrostatic double layer (EDL), while the competition for the functional groups between metal ions and other ions in solution is less important. Different trends are observed when biopolymers are used as sorbents. Chen and coworkers [3,12] conducted experiments of copper ion sorption by calcium alginate and found that a decrease in ionic strength results in an increase of copper ion removal. Chang and Hong [22] reported t h a t mercury uptake by Pseudomonas aeruginosa decreases with increasing ionic strength. Sorption of copper and cobalt investigated by J a n g et al. [23] indicated t h a t ionic strength does not affect strongly the sorption at concentration of copper above 60 ppm. At concentration of copper below 60 ppm, the copper binding increases with decreasing ionic strength. Cho et al. [24] showed that there was no significant decrease in the binding of Cd 2+ and Zn 2§ up to ionic strength of 10 .3 M, but the extent of sorption decreased when ionic strength exceeded 10 .2 M. Kuyucak and Volesky [25] found that in seawater where ionic strength and pH values are high, it is possible to have shifting of equilibrium toward sorption of light metal ions (e.g., Ca, Na, K, Mg). In distilled water, on the other hand, it was found that the equilibrium uptake of these sequestered metal ions, which are replaced by protons, is decreased. All these experimental observations are contrary to what is observed in the sorption of metal ions by hydrous oxides and activated carbon. In the case of biopolymers, the competition for the functional groups between metal ions and other ions in solution plays a more important role t h a n the compression of EDL. Since the number of functional groups available is fixed at

288 a certain pH, the sites available for metal ion uptake decreases with increasing ionic strength.

2.2. Models for m e t a l ion s o r p t i o n e q u i l i b r i u m The most common models used for sorption processes are the Langmuir and Freundlich isotherms. The major advantage of these models is their simplicity; however, both models fail to predict the effects of several important factors, such as pH and ionic strength. If the model parameters are obtained based on experiments under one set of conditions, the models cannot give accurate predictions for another set of conditions. For example, it has been shown that metal ion sorption increases dramatically in a short pH range. If the parameters of the empirical equations are based on experiments at a certain pH in this short range, these parameters cannot be used to calculate sorption equilibrium at a different pH. In addition to the pH effect, it is experimentally shown that the type and concentration of electrolytes and complexing agents play important roles in sorption as discussed in Section 2.1. Empirical equations fail to predict accurately sorption equilibrium under varying ionic strengths. Additionally, the empirical models cannot give a fundamental understanding of ion sorption. Numerous investigations have been carried out in the past several decades to overcome these problems. Several models have been developed to describe the mechanisms of metal ion sorption at solid-liquid interfaces. They include the Gouy-Chapman-Stern-Grahame model, the ion-exchange model, the ion-solvent interaction model and the surface complex formation model (SCFM) [9]. Among these models, SCFM is able to take into account the effects of pH, ionic strength, concentration of metal ions and other factors. It has been found that SCFM is very successful in predicting metal ion sorption by hydrous oxides [8,9,13-15,26], activated carbon [9-11] and biopolymers [3,12]. 2.2.1. I n t r o d u c t i o n of surface c o m p l e x f o r m a t i o n m o d e l The surface complex formation model (SCFM) considers that sorption depends on three interrelated processes: surface ionization, complex formation and the formation and presence of an EDL adjacent to sorbent surfaces [9]. There are several models proposed based on the considerations of the EDL formation and the types of functional groups. Different considerations of the electrostatic layers adjacent to solid surfaces lead to the formulations of the diffuse layer model, the Basic Stern model, the constant capacitance model and the triple-layer model [9,27]. The schematic representation of the triple layer model is shown in Figure 1. Hydrogen ions are usually considered to be sorbed in the inner layer and the other ions are regarded as being sorbed in the outer layer. By comparing acid-base titration data with modeling results, Westall and Hohl [27] concluded that there are no differences among various models. Hayes and coworkers [15,16], however, reported that the location of sorbed ions is strongly dependent on the relative bonding affinity of ions for the functional groups on the sorbents. Studying the effect of the ionic strength, one can distinguish the location of ions

289 sorbed. According to the experimental and modeling results of Hayes and coworkers, ions of higher affinity with surfaces are sorbed in the inner layer, while those of lower affinity are sorbed in the outer layer. The triple-layer model is considered valuable as it can predict sorption w h e n ions have lower or higher affinity with surfaces. In addition, modeling studies have revealed t h a t in addition to free m e t a l ions, metal ion hydroxides m a y also be sorbed [13,14,26].

Mm+OHO - . . . . . d,- M m + ~~.JY13 rJ:

M(0H)m O . . . . . . -:'- M m+

H

+

OH OHy

O ...... ~- X + O

.

X

_

+

Mm+

OH 2 OH~---~-YE

0 . . . . . . -L

H+ Mm+

OH/

o

[3

d

13"13

(Yd

C1

~o

(~o

Figure 1. Schematic illustration of surface complex formation model.

290 At the surface, functional groups for ion sorption are regarded as one average site or functional group, which is treated as an analog to complexing ligands in solution. Under different pH conditions, the functional group has different surface charge. Two models are commonly used, the one-pK model or the two-pK model. At a given pH, the one-pK model considers the functional group in the form of SOH ~/2+ and SOH 1/2- [28], while the two-pK model treats the functional group in the form of SO, SOH and SOH~ [13,14,26]. There are no significant differences between the one-pK and two-pK models in terms of describing metal ion sorption equilibrium [9]. On the other hand, Reed and Matsumoto [17,29] proposed a multiple-site model. They used three types of sites (or functional groups) instead of one average functional group to represent the surface sites of activated carbon. Their modeling results gave a good representation of their experimental data. Basically, the model is the same as the diffuse layer model with the exception that multiple sites are involved in the calculations. The diffuse layer model considers that all the ions are sorbed in the inner layer [9,27]. If the affinity of ions for the sorbent has to be considered, however, the triplelayer model must be employed. Better representations may result if one combines the multiple-site and triple-layer models in the calculations. In this case, the calculations will become extremely complicated and the number of the parameters may be too large to be determined.

2.2.2. Mathematical description of surface complex formation model Metal ion sorption by the functional groups on the surface of the sorbents results in the removal of metal ions from solution. The functional groups are treated as analogs to complexing ligands in solution. Electrolyte, represented as XY, is used to adjust the ionic strength in order to obtain information on the complexation of background electrolytes and the EDL. The general surface protolysis reactions, the electrolyte surface reactions and the metal ion sorption reactions for the triple-layer two-pK SCFM are listed in Table 1. The triple layer refers to the location of the ions in three different layers, i.e., o, 13 and d-layers as shown in Figure 1, while the two-pK refers to surface ionization corresponding to three types of surface species (SO-, SOH and SOH~) and associated with two equilibrium constants, KH1 and KH2. There are several possibilities for the removal of metal ions: removal may result from the formation of the surfacemetal complexes SO-M m+ or SOM(OH)I m-l)+ (a hydrolysis product of SO'M m+), or a combination of both. In addition, complexation of metal ions by multidentate surface sites, i.e., formation o f -(SO')nM(OH)}m-/)+, " may contribute to the binding of metal ions. All the ions except the hydrogen ion, normally, are assumed to be sorbed in the outer layer, or 13-layer (see Figure 1). If the functional groups have strong affinities for the ions, these ions could be sorbed in the inner layer as suggested by Hayes and coworkers [15,16].

291 Two key t e r m s h a v e to be defined before the f o r m u l a t i o n of the m a t h e m a t i c a l model. T h e y are the components a n d species which are used in the equilibrium model, M I N E Q L , by Westall et al. [30]. According to the definition by Westall et al. [30], the components in a m a t h e m a t i c a l sense form the i n d e p e n d e n t basis set from which every species can be defined a n d upon which the m a s s balance equations are based. In a chemical sense, the components are a set of chemical entities such t h a t every species can be r e p r e s e n t e d as the product of a reaction involving only these components and no component can be r e p r e s e n t e d as the product of a reaction involving only the other components. M a t h e m a t i c a l l y , species r e p r e s e n t log linear combinations of the components; chemically, a species is the product of a chemical reaction involving the components as r e a c t a n t s . A s s u m i n g a s y s t e m in which there are Mx aqueous species with concentrations xi, My sorbed (absorbed, adsorbed, complexated a n d ion-exchanged) species with concentrations yi a n d Mp precipitated species with activities of unity, one can express the c o n c e n t r a t i o n s or activities of the three types of species in t e r m s of Na aqueous c o m p o n e n t s with concentrations cj, Ns sorbent components with concentrations sj a n d two electrostatic components co a n d %. The specific equations, k n o w n as equilibrium relationships, which give the concentrations or activities of the t h r e e types of species are [13,26,30]:

Na a x xi = KX 1-] Ckik ' k=l (5

yi-

aYk

=

/

a

bYk caYocl3Yl3

lick

Sk

i = 1,2 ..... M x,

(la)

i = 1,2 .... , M y ,

(lb)

i = 1,2 ..... M p ,

(lc)

J~k=l

1= KP Na 1-[ CkPk , k=l where,

K x - modified equilibrium c o n s t a n t of the i-th aqueous species, x _ stoichiometric coefficient of the k-th aqueous component in the i-th aqueous aik species, K y - modified equilibrium c o n s t a n t of the i-th sorbed species, Y - stoichiometric coefficient of the k-th aqueous or electrostatic component in aik the i-th sorbed species, b ik y - stoichiometric coefficient of the k-th sorbent component in the i-th sorbed species, K p - modified equilibrium c o n s t a n t of the i-th precipitated species, aPk- stoichiometric coefficient of the k-th aqueous component in the i-th p r e c i p i t a t e d species.

292 Table 1 Mathematical description of two-pK triple-layer surface complex formation model Reactions

Equilibrium expressions**

Surface protolysis reaction + SOH + H + ~ SOH2

[ OH l

ISOHl{H+} - K.1

ex.(_ e~ o/

[so ]{H+}

SOH r SO- + H +

~()-I-I-]

= K H 2 e x p ( ~ -]

Electrolyte, XY, surface reactions SOH + X + <=>S O X + + H +

[SOH]{X + } =Kxexp

SOH + H + + Y- r SOH~Y

[SOH~Y-]

Outer-sphere metal ion (M m+) sorption*

nSOH + M m+ +/H20 r (SO-)n M(OH)I m-l)*

kT

-

_ Kyexp[_ e(r

E(O-)nM(OH)Im-"+1{"+}'n+" [SOH]n {M m+ }

- ~,fl

)/

n = KM(OH)/

(n +/)H + exp

kT

I(SO)nM(OH)Im-l-n)+l{H+}(n+l) Inner-sphere metal ion (M m+) sorption*

nSOH + M m+ +/H20 r (SO)nM(OH)I m-l-n)-~

[SOH]n {M m+ } exp

= KM(OH)/ x

leIn+l-mt~'!o kT

(n +/)H + * n is the number of surface sites in mol/L, which reacts with one mol/L of M m§ in solution; n = 1, 2, ..., N. l defines t h e / - t h hydrolysis product of Mm+; l = 0, 1, 2 ..... L; l = 0 corresponds to the simple ion. **Expressions include electrostatic interactions; if electrostatic interactions are neglected, exponential terms equal to one.

293 The modified equilibrium constants in Equations (la)-(lc) include activity corrections. Because of the EDL adjacent to the sorbent surface, electrostatic components c o and cfl are included in Equation (lb). The EDL consists of the charged surface and an excess of counter-ions over co-ions t h a t are diffused in the solution as shown in Figure 1. The electrostatic components are energy terms related to the electrical potentials ~'o and ~fl at o and ~-layers of the EDL as follows: co = exp(- e~T~) ,

c#=exp -

e ~'fl) kT

(2a) (2b)

'

where e, k a n d T are the electron charge, Boltzmann constant and absolute temperature, respectively. The electrical potentials, ~uo , U fl and ~d, in the three layers of the EDL are related to the corresponding surface charge densities, cr o , crfl and cr d , through the charge/potential relationships defined from the specific model chosen for the EDL. For the triple layer model illustrated in Figure 1, these relationships are [27]:

O"d = - ~ / 8 c R T I

sinh[e~dl[~ 2kT)

(3c) '

where C1 and Cz are the capacitances of the EDL, c is permittivity of the medium, R is the gas constant and I is the ionic strength. From the electroneutrality condition we have: O"o + O'fl + O"d = 0.

(4)

The surface charge densities, cro and ap, can be obtained from the sorbed species concentrations as: My o- o = ~

i=1

aYyi,

(5a)

294

My (5b)

cr fl = Z a ~ Y i .

i=1

Equations (1)-(5) may be solved to give the total sorbed concentration of the j-th aqueous component, S j , given by:

My j = 1,2..... N a .

S j = Z aY yi ,

i=1

(6)

Based on the various reactions listed in Table 1 and on the electrostatic relations and mass balance equations, the distributions of ions in both solid and liquid phases can be easily obtained by using the KINEQL model [9], which allows equilibrium calculations similar to MINEQL [30], in addition to kinetic calculations as will be discussed in Section 3.

2.2.3. Application of surface complex formation model In order to obtain parameters for the various sorption reactions, two kinds of experiments have to be performed and used as an input for determining the parameters. They include potentiometric titration experiments of sorbents and equilibrium experiments of metal ion sorption. The surface charge of sorbents obtained by the titration experiments provides information on the equilibrium constants of surface protolysis reactions (KH1 and Kin) and of electrolyte surface reactions (Kx and Ky ), as well as on the physical constants of EDL, i.e., the capacitances C 1 and C 2 (see Table 1 and Figure 1). Once these parameters are obtained, equilibrium constants of metal ion sorption reactions can be found based on sorption equilibrium experimental data and the parameters (KH1, KH2, Kx, Ky. C1 and C2) obtained from the modeling study of surface charge. The search of all these parameters can be carried out by comparing experimental data with model predictions and by using optimization techniques. All the estimations of parameters can be conducted by a computer program combining the KINEQL model [9] and the Sequential Quadratic Programming (SQP) optimization routine [31], with the objective function given as:

~__~I(Yj,measuredYj,yj, measuredCalculated/20.5 SHAT -

N P - NPAR- 1

(7)

where yj,measured and yj,calculated correspond to the experimental data and calculated results, respectively, NP is the number of experimental data and NPAR is the number of parameters to be found. When the parameters are obtained, the

295 speciation of ions at equilibrium is well-defined. Here, modeling of copper ion sorption equilibrium by activated carbon and by calcium alginate will be presented to illustrate the application of SCFM.

Example 1" Modeling of copper ion sorption equilibrium by activated carbon An experimental study of copper ion sorption by activated carbon was carried out by Chen and coworkers [5,11]. Representation of the experimental data can be performed by the two-pK triple-layer model, one of the SCFM approaches, as shown in Table 2. The reactions given in Table 2 are represented with the same way as in the KINEQL algorithm; c o m p o n e n t s are shown on the left-hand side of the reactions and species on the right-hand size of the reactions. Modeling of surface charge of activated carbon was conducted based on reactions (1) to (4) given in Table 2. A computer program with two major subroutines (KINEQL and SQP) was used to search for the equilibrium constants of reactions, the capacitances C 1 and C 2 and the concentration of functional groups. Values of these p a r a m e t e r s are found and are listed in Table 3. Detailed information of the p a r a m e t e r search can be found in the literature [5,11]. Table 2 Two-pK model and solution reactions for copper ion sorption Surface reactions: KH1 1. SOH + H + + e x p ( - y o ) r =SOH + KH2 2. S O H - H + - e x p ( - y o) r SOKCIO4 3. SOH + C10~ + H + + e x p ( - y o ) - e x p ( - y f l ) ~ SOH~-C10~ KNa 4. SOH + Na + - H + - exp(-y o ) + exp(-y/3) ,r SO-Na + 5. SOH+Cu 2 + + 2 e x p ( - y f l ) - H + - e x p ( - y o )

KSOCu ~ SO-Cu 2+

6. SOH+Cu 2+ + e x p ( - y f l ) - 2 H + - e x p ( - y o )

KSOCuOH ~ SO-CuOH +

Solution reactions: 7.

C u 2+ -

H+ r

8. Cu 2+ - 2H + r 9. 2Cu 2+ - 2 H + r 10.-H + r

CuOH + Cu(OI-I)2

2+

OH-

where Yo = e~'o / kT referred to o-layer and yfl = e g f l / kT to fl-layer

296 Table 3 Model p a r a m e t e r s for copper ion sorption by activated carbon and calcium alginate Name of p a r a m e t e r

Activated carbon

Calcium alginate

C1

31 gF/cm 2

25.3 ~tF/cm 2

C2

23 gF/cm 2

-*

KH1

107.26 M 1

100.49 M 1

KH2

10 "10"70M

10 -13.~176 M

Kclo4

103.56 M 2

-*

KNa

10-11.95

_r162

Ksocu

10 1"13

10 T M

KSOCuOH

10-14.55 M

10 .6.02M

Concentration of sorbent

10 g/L

20 g/L

Concentration of functional groups

0.10 M

0.924 M

Surface area

1.24x104 m2/L

lx104 m2/L

Dpcu

l x l 0 -6 cm2/s

l x l 0 5 cm2/s

kfcu

3xl0-3cm/s

5xl0-3cm/s

ap

0.2 mm

1.25 mm

IOp

1.316 g/cm 3

1.01 g/cm 3

* The two-pK Basic Stern model was employed in the modeling of copper ion sorption by calcium alginate and reactions 3 and 4 of Table 2 were not included.

The search of the equilibrium constants of copper ion sorption can be conducted, with the objective function, SHAT, given by Equation (7). Since the electrolyte was added to the system during the metal ion sorption, reactions (1) to (4) of Table 2 can be included in the model and the p a r a m e t e r s obtained from the modeling of surface charge are used in the modeling of sorption equilibrium. Copper ion solution reactions listed in Table 2 were also included in the modeling. Modeling by using the inner-sphere metal ion sorption shows t h a t no better fit could be obtained [11]. It was also assumed t h a t copper ion removal results from formation of single surface-metal complexes, SOCu 2+ or SOCuOH +. Modeling of the metal ion sorption shows t h a t a single surface-metal complex assumption (SOCu 2+ or SO-CuOH +) cannot be held because of larger values of SHAT. Some researchers [13,29] showed t h a t combinations of several single sorption reactions can give a better representation. In addition, Reed and Matsumoto [29]

297 reported that in cadmium ion sorption by activated carbon, the overall stoichiometric coefficient (moles H § released per mole cadmium sorbed) ranges from 0.89 to 1.24. This observation implies that there are more than one sorption reactions taking place. Therefore, a search is based on the consideration of two single sorption reactions (5) and (6) listed in Table 2. The modeling results shown in Figure 2 indicate that the combination of reactions (5) and (6) gives a good representation of the experimental data. The parameters found from the search are listed in Table 3. From the modeling, one can conclude that copper ion removal results from the formation of surface-metal complexes SO-Cu 2+ and SO-CuOH + in the outer layer. The disagreement between modeling results and experimental data in the low pH region shows that the proposed model has limitations. 0,12 Co=l.lxl0 4 M, I= 0.05M m =20 g/L Calcium Alginate 9 Experimental Data Modeling Results

O

0,1 O O

"~

0,08 0,06 0,04

O

0,02 ra~

0 1

3

5

7

9

11

pH Figure 2. Modeling of copper ion sorption equilibrium by activated carbon.

Example 2: Modeling of copper ion sorption equilibrium by calcium alginate Experiments of copper ion sorption by calcium alginate, one of the biopolymers, was also studied by Chen and coworkers [3,12]. This information provides important input data for the modeling of copper ion sorption. In this case, the two-pK Basic Stern (BS) model, a limiting case of the two-pK triplelayer model, was employed to describe metal ion sorption equilibrium by calcium alginate. The triple-layer model can be simplified by merging the three layers (o, [3 and d-layers; see Figure 1) in various ways [9]. In the BS model, [3 and d-layers coincide and the complexation reactions occur in the d-layer. Sorption reactions (1), (2), (5) and (6) and solution reactions (7) to (10) listed in Table 2 were used in

298 the modeling. Similar to the modeling process for activated carbon, two procedures were used in searching for the model parameters. First, titration of calcium alginate provided information on the surface charge which was used in the determination of such parameters as the capacitance of EDL (C1), KH1 and KH2, as well as the surface area and the concentration of functional groups. Values of these parameters are listed in Table 3. Based on equilibrium sorption experiments, on copper solution reactions available in Table 2, as well as on surface protolysis reactions found from the modeling of surface charge, searching for the equilibrium constants of sorption reactions was initiated. Similar to the modeling of copper sorption by activated carbon, it was assumed that copper ion removal results from the formation of such surface-metal complexes as SO-Cu 2§ and SO-CuOH § In the calculations, a computer program including KINEQL and SQP was used to find the equilibrium constants of reactions (5) and (6) of Table 2, with the objective function SHAT given by Equation (7). The constants Ksocu and KSOCuOHwere found and listed in Table 3 and the modeling of copper ion sorption is shown in Figure 3 together with the experimental data. As shown in Figure 3, the two-pK BS Model provides a good representation of the experimental data and describes well the effect of pH on the equilibrium uptake. The formation of the surface-metal complexes S O C u 2+ and SO-CuOH § results in the copper ion removal.

0,12 Co=l.lxl0 -4M, I = 0.05M m =20 g/L Calcium Alginate 9 ExperimentalData Modeling Results

O

0,1 o o

-~ ~D

o

0,08 0,06 0,04

0

0,02 r~

~

o 1

3

5

7

9

11

pH Figure 3. Modeling of copper ion sorption equilibrium by calcium alginate.

299 3. METAL ION SORPTION KINETICS In this section, factors affecting metal ion sorption kinetics and possible mechanisms of sorption kinetics by various sorbents, as well as kinetic models, are discussed. KINEQL, a kinetic model, is introduced, followed by a discussion on its applicability to describe copper ion sorption kinetics by activated carbon and calcium alginate. 3.1. Factors i n f l u e n c i n g metal ion sorption k i n e t i c s Metal ion sorption kinetics experiments show that sorption follows two phases. Rapid sorption occurs initially followed by a much slower phase, which takes from a few hours to several weeks [11,32]. The sorption rate depends on the mixing strength of solution, type of both the sorbent and the sorbate, pH and ionic strength of solution, initial concentration of metal ions, as well as concentration of sorbents. Lo and Leckie [33] demonstrated that sorption of cadmium ions by aluminum oxide reaches equilibrium in one day. Chen et al. [11] showed that copper ion sorption by activated carbon reaches equilibrium quickly at higher pH values and lower initial concentration. Most of the sorption is completed in the first hour and equilibrium is reached within four hours approximately. Huang and Smith [34] investigated cadmium sorption kinetics by using two types of activated carbon, Nuchar SN and Nuchar SA, at various values of pH and sorbent concentration and in the presence of cyanide (CN-) and fluoroborate (BF4). They found that maximum sorption is reached in the first two minutes. Thirty minutes is sufficient for the complete sorption of lead and copper by powdered activated carbons as reported by Netzer and Hughes [35]. Complete sorption of cobalt by the same sorbents is done in two hours. Reed and Matsumoto [17] used two types of activated carbon, Darco HDB and Nuchar SN, to remove cadmium and found that sorption of cadmium is completed in six hours. In these studies, the removal kinetics was found to be slower for solutions containing high initial cadmium/activated carbon concentration ratios. Perez-Candela et al. [36] showed that there is no further sorption of chromium (VI) by activated carbon after nine hours. Wilczak and Keinath [32] reported that initial sorption of copper and lead onto two types of activated carbon, Nuchar SA and Filtrasorb 400, occurs rapidly, followed by a slow sorption step that takes about 35 days. Metal ion sorption with granular activated carbon is slower than that with powdered activated carbons [29]. Rubin and Mercer [37] reported that the equilibrium time for the 810 mesh activated carbons is 95 hours, while cadmium sorption is achieved equilibrium in six hours with 50-200 mesh carbons. Huang and Smith [34] concluded that pore diffusion is the rate limiting step in metal ion sorption. Metal ion sorption kinetics by biopolymers has extensively been investigated in recent years. Low et al. [38] observed that copper sorption by Eichornia crassipes is rapid in the first ten minutes. Corder and Reeves [39] reported that no further uptake of nickel by A. flos-aueae is observed after three hours. Konishi

300 et al. [40] showed that uptake of neodymium and ytterbium by alginate reaches equilibrium in two hours. Chen et al. [3] reported that copper ion removal by calcium alginate occurs rapidly in the first hour, followed by a slow uptake process that takes about 15 hours. Based on sorption kinetic experiments with various kinds of calcium alginate, it was found that diffusion plays an important role in the uptake rate. 3.2. M o d e l s for m e t a l ion s o r p t i o n k i n e t i c s Among a few modeling investigations, there are two major kinetic models, one formulating sorption kinetics empirically [32,33,41] and the other considering it mechanistically [3,9,11,40,42-44]. The first group of models are essentially based on data-fitting and are not able to interpret the sorption experimental data. The second group of models can be catalogued into reaction-controlled and diffusioncontrolled models and are based on the relative importance of the chemical reaction rate to the diffusion transport rate [3,9,11]. Reaction controlled or diffusion controlled models can be identified experimentally [3,5]. Diffusioncontrolled models can be employed to describe metal ion sorption kinetics once the reactions rate is found to be negligible. Similarly, reaction-controlled models are used when diffusion of metal ions is less important. Among the various models, the diffusion-controlled models are often used for activated carbon and biopolymers. Most of the diffusion-controlled models discussed in the literature use a simplified expression (e.g., the Freundlich and Langmuir models) to describe the local equilibrium relationship. As mentioned in Section 2.1, these models and their parameters are system-specific and cannot be applied to other conditions, such as different pH and ionic strength. Yiacoumi and Tien [9] developed a kinetic model, called KINEQL, in which SCFM is used to express the local equilibrium relationship. KINEQL can solve not only equilibrium problems as discussed in Section 2.2.2, but also kinetic problems in reaction-controlled and diffusion-controlled cases. It has been shown that KINEQL works well in the representation of metal ion sorption kinetics by hydrous oxides [9,42,43], activated carbon [5,11] and biopolymers [3,5,12]. The advantage of KINEQL is that all the important factors, such as pH, ionic strength, temperature, metal ion concentration and sorbent concentration, are considered. 3.2.1. K I N E Q L m o d e l Single-species batch sorption kinetic experiments for the sorbents can be analyzed to determine the rate parameters of the uptake process. The first step in this analysis is to determine the rate-limiting step of the sorption process by various experimental techniques, such as using varying sorbent sizes and mixing intensities. Once the rate-limiting step is identified, the appropriate formulation can be followed. Many studies show that diffusion plays an important role in the sorption kinetics by activated carbon [34,37] and biopolymers [3,12,45,46]. On the other hand, sorption reactions in most systems were considered

301 i n s t a n t a n e o u s and, therefore, the diffusion mechanism is a s s u m e d to be the controlling step in the metal ion sorption process and is discussed in detail below. Neglecting any electrical potential gradients in the internal surface of the sorbent particles, the intraparticle equation can be written as:

617j Ep - -

D pj ~ r 2 - ~ ] =

r2

p p Oqj

c~

m c2

,

(8)

where cj is the local concentration of ion j in the pore solution within the particle pellet and qj is the local concentration of ion j in the sorbed phase based on bulk

m, top, c p and D~j are the mass sorbent concentration, the

solution volume,

sorbent density, the sorbent porosity and the diffusivity, respectively. The independent variables are the time t and the radial distance r. When the sorption reaction step has no effect on the rate of the overall uptake process, equilibrium exists between the pore liquid phase and the solid phase across the sorbent particle. The local concentrations of ion j in the solution and sorbed phases of the sorbent pellets, cj and q j , are related with the equilibrium relationship which is defined here by the function f, i.e.,

qj = f (cj ) ,

(9)

The initial and boundary conditions are:

cj=O

at

t=O,

(lOa)

&J =0 &

at

r=O,

(lOb)

at

r=ap,

(10c)

&j

k j j ( C j - c j ) = Dpj cTr where

k~ is the external

mass transfer coefficient of the ion, Cj is the

concentration of ion j in the bulk solution and a p is the radius of the sorbent particles. The macroscopic conservation equation for the concentration of ion j in the bulk solution, Cj, is described as: .

dt

.

.

.

,Op a p

.

.

.

r =ap

,Op

ep

--~

+

,Op dt J

,

(11)

where cj and q j are the average quantities of c i and qj within the sorbent pellet.

302

By applying the parabolic boundary approximation [47,48] and assuming that

pp--

@cj <<~qj, one can obtain the equations describing the diffusion of N a m aqueous components inside of the sorbent as: dSj ~= dt

Ej 3m B j ( c j - C j s ), pp

j = 1,2..... N a ,

(12)

Ej - D p j / a p2,

j = 1,2..... Na,

(13a)

Bj = apk y] / Dpj ,

j = 1,2,...,N a .

(13b)

where

The symbol cjs in Equation (12) defines the concentration of ion ] at the external surface of the sorbent, given by:

Cjs=

E(

(1- Bj)c] +-1 1- Bj)2--cj2 + 8~-5-j(75-j+ Bjcj 4 4

)11/2

, j : 1,2.... ,Na,

(14)

where

cjSj cj= My

j = 1,2,...,N a .

(15)

X ai~Yi i=1

The system of Equations (12)-(15) is an approximate representation of intraparticle diffusion equations in a batch system based on the mathematical scheme introduced by Yao and coworkers [47,48]. The solution of this system of equations, in association with the required equilibrium relationships, provides Sj as a function of time. The mathematical method to solve the system of Equations (12)-(15) in connection with Equations (1)-(5) should include a combination of a chemical equilibrium algorithm with a differential equation solver. Such an algorithm was developed by Yiacoumi and Tien [9] in the KINEQL model for the case of sorption of metal ions from aqueous solutions under conditions where either reaction or mass transfer controls the process. The rate parameters can be estimated from correlation relationships available in the literature [49]. The first-order ordinary differential equations corresponding to the rate expression given by Equation (12) can be solved numerically by using an ordinary differential equation solver, such as the EPISODE package [50].

303 A similar mathematical formulation can be derived for the reaction-controlled limiting case, as well as for the more generalized one in which both diffusion and reaction are important. Formulations for the latter case are available in the literature to describe sorption of nonionic organic compounds by porous particles [51-54].

3.2.2. Application of KINEQL model To illustrate the application of the KINEQL model, modeling of copper ion sorption kinetics by activated carbon and by calcium alginate is presented here. The modeling assumes that diffusion is the controlling step for metal ion sorption. Experimental data can be found in the literature [3,5,11].

Example 1: Modeling of copper ion sorption kinetics by activated carbon To carry out the modeling, the equilibrium relationships as well as the identity of the ions that diffuse into the interior of the sorbent have to be defined. Surface reactions (1) to (6) and solution reactions (7) to (10) of Table 2 are used in the kinetic calculations as discussed in Section 2.2.3. The parameters required in the calculations are listed in Table 3. It is assumed that only Cu 2§ ions diffuse from the bulk of the solution to the interior of the sorbent. The surface-metal complex SO-Cu 2§ is formed by the sorption of Cu2+, while the species SO-CuOH § is formed by the hydrolysis of the surface species SO-Cu 2§ Such an assumption was necessary in order to give physically realistic model parameters [9]. The rate expression for sorption occurring in the micropore region can be derived from Equation (12) as:

d[so-cu2+]+[so-cuo.+]) cu dt

3m Bcu[{Cu2+ } _ {Cu2+ }s],

(16)

Pp

where

_

-

+1 E

+

)2

2

}

8{Cu2+ x

(17)

The sensitivity analysis of the model showed that the external mass transfer coefficient affects the initial sorption phase and results in most of the metal ion removal, while the diffusivity influences the rate of sorption in the second phase [11]. Increasing these two parameters results in the rapid sorption of metal ions. Wilczak and Keinath [32] reported that the slow sorption phase takes several weeks and may result from lower diffusion of metal ions in the activated carbon.

304 Modeling of copper ion sorption kinetics was carried out with the diffusivity, external mass transfer coefficient and other parameters listed in Table 3. The external mass transfer coefficient of 3x10 3 cm/s is close to the value obtained from correlations available in the literature [49]. The diffusivity of 10 .6 cm2/s is one order of magnitude lower than that in water, which can be explained from the nature of the porous structure of the activated carbon pellets. Modeling results are plotted in Figure 4 against experimental data, where the solid line and the points represent the modeling results and the experimental data, respectively. It shows that the model describes well the concentration histories for pH 4 and 5. The poor description for the concentration history of pH 3 is because of the poor description of the sorption equilibrium at the same pH (see Figure 2). The prediction of the sorption kinetics strongly depends on the accuracy of the equilibrium calculations. Better description of sorption kinetics cannot be accomplished without an improved sorption equilibrium model.

~, 0,08 Co=7xl 0-5 M, I=0.05M 0,06 pH=3.0

0,02

o

o

"~ 0,00 0

20

40

60

80

Timel/2 , m l .n 1/2 Figure 4. Modeling of copper ion sorption kinetics by activated carbon.

Example 2: Modeling of copper ion sorption kinetics by calcium alginate Similar to the modeling of copper sorption kinetics by activated carbon, the model used here is the intraparticle diffusion model where the equilibrium relationship is expressed by the sorption and solution reactions (1), (2), (5), (6) and (7) to (10) of Table 2 and the EDL is described by the two-pK BS model.

305

Similar to the case of sorption by activated carbon (see Example 1), it is assumed in the modeling t h a t free copper ions diffuse from the bulk of the solution into the interior of calcium alginate and that the surface species SO-Cu 2§ and SOCuOH § are formed. The external mass transport affects the initial sorption rate and results in most of the metal ion removal, while the diffusion influences the rate of sorption in the second phase. It is assumed that the diffusivity is similar to t h a t in water, which is reasonable, given the high water content in the calcium alginate gel beads. Here a diffusivity of lx10 5 cm2/s is used, very close to that in water. Based on an empirical equation available in Tien's monograph [49], the external mass transport coefficient was found to be 5• .3 cm/s. The mass concentration, radius and density of beads used in the calculations can be found in Table 3. Modeling results are plotted in Figure 5 against experimental data, where the solid line and the points represent the modeling results and the experimental data, respectively. As one can see, the model gives a good representation of experimental data and successfully predicts the effect of pH on the kinetics of copper ion uptake. 0,12 Co=l.03xlO -4 M, o

~-, 0,10 9

0,08

nlq=9

.~ 0,06 0,04

o

oH=3

0,02 oH=5.

"~ 0,00

i

i

0

10

20

30

40

50

60

Timel/2 , mln . 1/2 Figure 5. Modeling of copper ion sorption kinetics by calcium alginate.

It can be concluded that the diffusion-controlled model presented here, in association with the surface complexation approach, can successfully describe the metal ion sorption equilibrium and kinetics by activated carbon and calcium

306 alginate. The rate parameters used in the model, i.e., the diffusivity and external mass transport coefficient, have a physical significance and their values obtained from the experimental data agree with existing correlations and the physics of the process. Based on the model, the mechanism of copper ion removal is diffusion of free copper ions from the bulk solution to the external and internal surfaces of the sorbents, resulting in the formation of two surface-metal complexes: SO-Cu 2§ formed by sorption of Cu 2§ and SO-CuOH § formed by hydrolysis of SO-Cu 2§

"

METAL ION T R A N S P O R T IN T R E A T M E N T AND N A T U R A L SUBSURFACE SYSTEMS

In Sections 2 and 3, the discussion on metal ion sorption equilibrium and kinetics was based on simple batch reactor systems. More complex reactors, such as fixed-bed columns and fluidized beds, are usually employed for metal ion removal. In addition, pollution of metal ions in natural subsurface systems has increased dramatically in the recent decades. Eventually, information from experimental and modeling studies of batch sorption equilibrium and kinetics is not sufficient. Experimental studies with various reactors, as well as transport modeling under certain conditions, have to be carried out. Microscopic and macroscopic modeling are necessary for the description and practical applications of metal ion sorption. Microscopic mathematical models, which are essentially the kinetic models discussed in Section 3, describe the uptake rate of metal ions from the solution to the sorbent, while macroscopic modeling incorporates the microscopic models with mass transport and hydrological transport equations to predict the behavior of the process. 4.1. F a c t o r s i n f l u e n c i n g m e t a l ion t r a n s p o r t in t r e a t m e n t a n d n a t u r a l subsurface systems Metal ion transport can be categorized into two major groups, one in treatment systems, e.g., fixed-bed columns and another in natural surface and subsurface systems. The major difference between these two groups is that the flowrate of the first is much larger than that of the second. In addition to the factors discussed in Sections 2 and 3, the flowrate of solution also plays an important role in metal ion transport. 4.1.1. T r e a t m e n t s y s t e m s As a treatment approach, sorption by fixed-bed columns has been widely used in water and wastewater treatment. While it is well-documented that this technique provides a highly effective remediation method for organic contaminants, there are not many studies on metal ion removal by fixed-bed columns. Research on metal ion sorption in the past has focused on batch equilibrium and kinetic studies. Nonetheless, a few studies reveal valuable information on metal sorption in fixed-bed columns by various kinds of sorbents.

307 These investigations include the influence of influent pH, concentration of metal ion and chelating agents, flowrate of the influent solution, bed depth and the effect of pre-treatment/pre-washing of fixed-bed columns. Among different chemical and physical conditions, column pH is a critical parameter that influences fixed-bed sorption performance. Fixed-bed experiments with activated carbon and at various values of influent pH, ionic strength and metal ion concentration, investigated by Chen and coworkers [5,55], showed that with an increase of influent pH, the breakthrough time increases. An overshoot of effluent concentration was observed due to desorption of copper ions from surface-metal complexes. A dramatic increase of pH at effluent was also found in the experiments. The breakthrough time increases as ionic strength and empty bed contact time are increased and influent metal ion concentration is decreased. The removal of lead by granular activated carbon column was investigated by Reed and Arunachalam [56]. Fifty bed volumes (BV) of stream with pH 5.47 and concentration of 10 ppm were treated by virgin carbon. The effluent pH decreases from pH 7 at the beginning of the operation to pH 5.4 at the end. The breakthrough occurs at 300 BV by using activated carbon pretreated with HNO3, or a combination of HNO3 and NaOH. The effluent pH also decreases from 11 to 8 at breakthrough. Additionally, the presence of acetic acid or EDTA significantly decreases the BV treated. Sorption and surface and pore liquid precipitation were assumed to be the dominant removal mechanisms. Shay and Etzel [57] reported that a stream with nickel and zinc, as well as citrate and EDTA, was treated by granular activated carbon column. Thirty-six minutes is sufficient to reach complete metal ion removal. Bowers and Huang [58] demonstrated that the breakthrough of granular activated carbon column occurs after 600 BV of Cr (VI) were treated. In the study, the fixed-bed column was washed previously with 150 BV of pH 2.5 and 0.1 M NaC1 solution. Huang and Wirth [59] conducted a study of fixed-bed sorption of cadmium with powdered activated carbon. Over 1500 BV of wastewater with pH 7.0 and concentration of 10 .4 M were treated. Removing lead from drinking water with granular activated carbon fixed-bed columns at various values of pH and in the existence of other metal ions was also investigated by Kuennen et al. [60]. It was found that for lead sorption, the breakthrough takes place at 10, 900 and 7000 BV for pH 5, 6 and 9, respectively. At pH 7, breakthrough occurred at 1000, 1000, 3200 and 6400 BV for zinc, cadmium, copper and lead, respectively. These results illustrate that the activated carbon used is favorable for copper and lead sorption, but not for zinc and cadmium. Interest in the usage of biopolymers as sorbents in fixed-bed columns has increased in recent years since biopolymers can be inexpensive, have high selectivity and high capacities for sorption of many heavy metal ions. Copper ion sorption by calcium alginate was investigated by Chen and Yiacoumi [12]. It was found that the breakthrough occurs in the beginning of the experiment and the column reaches saturation in 300 hours. Low et al. [38] conducted sorption of

308 electroplating waste containing copper and nickel in a column by using 2.5 g nonliving dried water hyacinth roots as sorbents. Solution pH was 5.24 and the concentrations of copper and nickel were 7.98 ppm and 14.73 ppm, respectively. The flowrate of the waste and the internal diameter and length of the column were 10 mL/min, 1.2 cm and 12.5 cm, respectively. It was found t h a t about 700 mL nickel waste and 3100 mL copper waste were completely removed from the column. Watson et al. [61,62] reported that over 50 BV of strontium solution with 10 ppm of initial metal concentration and pH 7 were treated by using Micrococcus luteus (ATCC-4698). The usage of granules of crosslinked biopolymers of brown alga A. nodosum (FCAN) for fixed-bed sorption of cadmium with pH 5.5 and concentration of 10 ppm was investigated by Volesky and Prasetyo [63]. They found that there is a linear relationship between the breakthrough time and the depth of column at any given flowrate.

4.1.2. S u b s u r f a c e s y s t e m s Heavy and radioactive metals resulting from metal milling, incineration and military operations have caused serious impact on natural environments mainly because of their toxicity to h u m a n and their longer half-lives. For example, in recent years, the concern on h u m a n health from depleted u r a n i u m (DU) is increasing. A facility for producing uranium metal and uranium tetrafluoride began operation in 1951 and was shut down in 1989 [64]. Estimates suggest t h a t 2,000,000 to 4,000,000 m 3 of soil near this facility may have unacceptable levels of uranium contamination. The current technologies for the remediation of DU-contaminated soil include: (i) excavation and transportation of the soil to repository, (ii) immobilization of the DU in place, (iii) physical separation and removal of the more highly contaminated soil fractions from the rest of the soil and (iv) soil washing. Field tests indicate that metal ions may transport hundreds of miles after some years, depending on the pH and other factors of soil. 4.2. Models for m e t a l ion t r a n s p o r t in t r e a t m e n t and n a t u r a l s u b s u r f a c e systems The development of metal ion transport models was initiated for designing fixed-bed columns for the t r e a t m e n t of metal ions, as well as for determining the extent of metal ion transport in natural subsurface systems. Many of the existing models use empirical methods to express the chemistry of the system and cannot describe accurately a system in which aqueous speciation, precipitation, redox, ion-exchange and sorption reactions are involved. These reactions can be correctly described using the equilibrium model MINEQL [30] discussed in Section 2.2. HYDROGEOCHEM [4,65] comprises MINEQL and a hydrologic transport model and can predict metal ion transport in subsurface systems. The model was developed for simulation of metal ion transport in subsurface systems. One of the assumptions in HYDROGEOCHEM is t h a t all the chemical reactions are at equilibrium, which may be valid for subsurface

309 systems as the flowrate of groundwater is very slow, but may not be true in fixedbed columns where the flowrate is usually higher. 4.2.1. I n t r o d u c t i o n of H Y D R O G E O C H E M m o d e l HYDROGEOCHEM is a coupled model of HYDROlogic transport and GEOCHEMical reactions in saturated-unsaturated media. The model is designed to simulate transient and/or steady-state transport of aqueous species and transient and/or steady-state mass balances of sorbent and ion-exchange sites. Thus, HYDROGEOCHEM is a general-purpose model useful for the simulation of reactive multispecies-multicomponent chemical transport through saturatedunsaturated media. The model includes hydrophysical and chemical processes. The hydrophysical processes, which include advection, dispersion and diffusion, are described by a set of partial differential equations. The chemical processes, which include aqueous speciation, sorption/desorption, ion-exchange, precipitation/dissolution, redox and acid-base reactions, are assumed to be at equilibrium and are described by a set of nonlinear algebraic equations. Additionally, the intraparticle mass transport resistance of metal ions is neglected since groundwater flowrate is low. The transport of aqueous components, the mass balance of sorbent components and the mass balance of ion-exchange sites are included to describe the mass transport processes in subsurface systems, with detailed information available in the literature [4,65]. These transport equations, together with the appropriate flow equations and the chemical reactions of the system, are solved by a finite element method in order to give the concentrations of various species. 4.2.2. A p p l i c a t i o n of H Y D R O G E O C H E M m o d e l Modeling of copper ion sorption in fixed-bed columns and simulation of depleted uranium transport in subsurface systems are demonstrated here as examples for the application of HYDROGEOCHEM. The detailed information of these modeling procedures has been discussed elsewhere [5,6,55]. Example 1: Modeling of copper ion transport in activated carbon fixed-bed columns In copper ion sorption experiments with activated carbon fixed-bed columns, carried out by Chen and coworkers [5,55], it was found that pH increases dramatically once the activated carbon is put into the solution [11]. In order to eliminate this pH variation, deionized water with ionic strength of 0.005 M and pH 3.1 or 5.8 was used to wash the fixed-bed columns. Modeling of the sorption in fixed-bed columns was performed by using the HYDROGEOCHEM model. The chemistry of aqueous reactions, sorption reactions and their constants were obtained from Tables 2 and 3. The physical parameters, such as influent flowrate, were obtained from the experimental study [5,55]. Similar to the modeling performed for the sorption kinetics experimental study, the simulation was conducted with fixed pH as 3.1 and 5.8. The modeling results, together with experimental data, are presented in Figure 6. As one can see, the model gives a

310 reasonable representation of the experimental data. It also can predict correctly the exhaustion or saturation of fixed-bed column for metal ion sorption. There are several factors resulting in the inaccuracy of the model for the representation of experimental data. One of the factors is that sorption kinetics is not considered in the formulation of the model. From the kinetic study, it is demonstrated that the sorption takes about four hours to reach equilibrium [11]. The assumption of instant equilibrium in the HYDROGEOCHEM model and the less accurate equilibrium model, therefore, are the major reasons for the disagreement.

0,12 pH=3.1 0,1 r

0,08

pH=5.8

9149

0,06 0,04 0,02

0_ 0

9

Co=lxl0-4 M, I=5x10 3 M L=4.0 cm, Q= 15 mL/min

~ 10

20

30

40

50

Time, hr Figure 6. Modeling of copper ion sorption in activated carbon fixed-bed columns at two different pH values.

Example 2: Modeling of depleted uranium (DU) transport in subsurface systems. In order to obtain an understanding of DU transport in n a t u r a l subsurface systems in terms of space and time and under different chemical and physical conditions, simulations of DU transport were conducted by using HYDROGEOCHEM. DU transport resulted from a continuous DU source of constant concentration and the groundwater flow. Different types of reactions, such as aqueous speciation, precipitation and sorption, were considered in the simulations. These reactions were obtained from the literature [65-68]. Depleted uranium transport was studied with the consideration of aqueous speciation, precipitation and sorption reactions. The hydrologic conditions and

311 background and incoming concentrations of the chemical components are shown in Table 4. The results in Figure 7 show t h a t DU concentration decreases to 0 at 1500 m after 90 days and at 4100 m after 720 days for influent pH 4.0.

Table 4 DU t r a n s p o r t with aqueous speciation, precipitation and sorption Component

pH

C a 2+

C O 32-

Background

4.0

lxl0 4 M

l x l 0 -4 M

Incoming

4.0

lx10 -a M

lxl0 4 M

SO24 -

Sorbent

0

lx10 4 M

lx10 -4 M

5x10 -4 M

l x l 0 -4 M

0

DU

Velocity = 0.3 m/d, dispersivity = 0.1 m, moisture content = 0.3

0,6 pH=7.9, t=790 d ,, pH=7.9, t=80 d

0,5 <~N

g• g

--o- pH=4.0, t=720 d

0,4

0,3

0,2

0,1

0

1000

2000

3000

4000

5000

Distanc e, m Figure 7. Effect of pH on DU transport in groundwater.

It is well-known t h a t pH plays an i m p o r t a n t role in metal ion sorption. With an increase of pH, metal ion u p t a k e increases, as discussed in Section 2. The data of Table 4 were used, but the background pH and t h a t of the incoming s t r e a m were changed to 7.9. The results in Figure 7 show t h a t DU concentration decreases to 0 at 700 m after 80 days and at 2000 m after 790 days. Comparing the results for pH 4.0 with those for pH 7.9, one can see t h a t a higher pH can

312

effectively retard DU transport in a subsurface system since at lower pH values sorption is negligible as can be seen from the equilibrium studies (see Figures 2 and 3). A conclusion from this example is that sorption reactions play a major role in the transport of metal ions in subsurface systems. 5.

SUMMARY

A systematic approach for modeling metal ion sorption phenomena in engineered and natural environmental systems is presented in this chapter. In addition, all the important factors that affect metal ion sorption equilibrium, sorption kinetics and transport are discussed based on relevant studies in the literature. Solution pH, ionic strength and initial concentration of metal ions play important roles in metal ion sorption equilibrium. Metal ion sorption increases with increasing pH and decreasing concentration. For activated carbon and hydrous oxides, the metal ion uptake increases when ionic strength is increased, but it decreases for biopolymers. The surface complex formation model is able to describe well the metal ion sorption equilibrium. Metal ion sorption rate varies with pH, ionic strength, metal ion concentration and type and concentration of sorbents. It has been shown that the KINEQL model represents successfully the metal ion sorption kinetics. Metal ion transport in treatment and subsurface systems is a complex process, involving important parameters, such as influent pH, ionic strength and flowrate of solution. The HYDROGEOCHEM model can simulate the metal ion transport in fixed-bed systems, as well as in subsurface systems. ACKNOWLEDGEMENTS

Financial support for this work, provided by the Georgia Institute of Technology, the Army Environmental Policy Institute and the National Science Foundation through a Career Award (BES-9702356) to S. Yiacoumi, is highly appreciated. The authors are also thankful to Dr. C. Tsouris, K. Subramaniam, F. Tendeyong, J. Yoon and T. Blaydes for their help and comments during this work. NOTATION ap

radius of sorbent particles

ap ik

stoichiometric coefficient of the k-th aqueous component in the i-th precipitated species

aik

stoichiometric coefficient of the k-th aqueous component in the i-th aqueous species

ay ik

stoichiometric coefficient of the k-th aqueous or electrostatic component in the i-th sorbed species

x

313

biky

stoichiometric coefficient of the k-th sorbent component in the i-th sorbed species parameter defined by Equation (13b) stoichiometric coefficient of jth sorbent component in ith sorbed species

cj

concentration of j-th aqueous component, or local concentration of ion j in the particle, based on pore solution volume

cj

average quantity of cj

Cjs

concentration of ion j at the external surface of the sorbent

cO

electrostatic components defined by Equation (2a)

cz

electrostatic components defined by Equation (2b)

Co

initial concentration of metal ion

C1

first capacitance of EDL

C2

second capacitance of EDL

c,

concentration of ion j in the bulk solution diffusivity

e

electron charge

Ej

parameter defined by Equation (13a)

I

ionic strength

k

Boltzmann constant (1.38x10 2a J/K)

koo

external mass transfer coefficient

Kf

modified equilibrium constant of the i-th precipitated species

Kx

modified equilibrium constant of the i-th aqueous species

Ki"

modified equilibrium constant of the i-th sorbed species

L

fixed-bed column length

m

mass concentration

Mp

number of precipitated species

Mx

number of aqueous species

314

My

number of sorbed species

Na

number of aqueous components

NP

number of experimental data

NPAR

number of parameters to be found

qj

local concentration of ion j in the adsorbed phase, based on bulk solution volume

qj

average quantity of qj

O

flowrate radial distance

Sj

concentration of j-th sorbent component

Sj

total sorbed concentration of jth aqueous component

SHAT

objective function in Equation (7) time

T

absolute temperature

Xi

concentration of ith complexed species

Yi

concentration of ith adsorbed species

Yj,measured

experimental data

Yj,calculated

calculated results

g

permittivity of the medium

gp

porosity

Pp

density of sorbent surface charge densities at o-layer surface charge densities at ~-layer

cr d

surface charge densities at d-layer

~'o

electrical potential at o-layer electrical potential at ~-layer

~gd

electrical potential at d-layer

315 BS

Basic Stern

BV

bed volumes

DU

depleted uranium

EDL

electrical double layer

SCFM

surface complex formation model

SQP

sequential quadratic programming

REFERENCES

1. D. Faust and O.M. Aly, Adsorption Processes for Water Treatment, Butterworth Publishers, Stoneham, MA, 1987. 2. B. Volesky, Biosorption of Heavy Metals, CRC Press, Boca Raton, FL, 1989. 3. J. Chen, F. Tendeyong and S. Yiacoumi, Environ. Sci. Technol., 31 (1997) 1433. 4. T. Yeh and T.S. Tripathi, Water Resour. Res., 27 (1991) 3075. 5. J. Chen, Sorption of Metal Ions from Aqueous Solutions: Equilibrium, Kinetics and Transport, Ph.D. Thesis, Georgia Institute of Technology, Atlanta, GA, 1997. 6. J. Chen and S. Yiacoumi, Transport Modeling of Depleted Uranium (DU) in: Subsurface Systems, submitted for publication in Waste Management (1997). 7. K. Subramaniam, S. Yiacoumi and C. Tsouris, Effect of Sorption on Flocculation Kinetics: Experimental Studies, submitted for publication in Sep. Sci. Technol. (1997). 8. A. Dzombak and F.M.M. Morel, Surface Complexation Modeling: Hydrous Ferric Oxide, John Wiley & Sons, New York, 1990. 9. S. Yiacoumi and C. Tien, Kinetics of Metal Ion Adsorption from Aqueous Solutions: Models, Algorithms and Applications, Kluwer Academic Publishers, Norwell, MA, 1995. 10. O. Corapcioglu and C.P. Huang, Wat. Res., 21 (1987) 1031. 11. J. Chert, S. Yiacoumi and T.G. Blaydes, Sep. Technol., 6 (1996) 133. 12. J. Chen and S. Yiacoumi, Sep. Sci. Tech., 32 (1997) 51. 13. A. Davis and J.O. Leckie, J. Colloid Interface Sci., 67 (1978) 90. 14. A. Davis and J.O. Leckie, J. Colloid Interface Sci., 74 (1980) 32. 15. F. Hayes and J.O. Leckie, J. Colloid Interface Sci., 115 (1987) 564. 16. F. Hayes, C. Papelis and J.O. Leckie, J. Colloid Interface Sci., 125 (1987) 717. 17. E. Reed and M.R. Matsumoto, J. Environ. Eng., 119 (1993) 332. 18. R. N. Rao, L. Iyengar and C. Venkobachar, J. Environ. Eng., 119 (1993) 369. 19. Y. Hao, Y. Zhao and G.J. Ramelow, J. Environ. Sci. Health, Part A: Environ. Sci. Eng., 29 (1994) 2235.

316 20. D. Ke, W.L. Anderson, R.M. Mocrief and G.D. Rayson, Environ. Sci. Technol., 28 (1994) 588. 21. O. Corapcioglu, The Removal of Heavy Metals by Activated Carbon Process from Water and Wastewater, Copper(II), Lead(II), Nickel(II) and Zinc(II), Ph.D. Thesis, University of Delaware, Newark, DE, 1984. 22. J. Chang and J. Hong, Biotechnol. Bioeng., 44 (1994) 999. 23. K. Jang, S.L. Lopez, S.L. Eastman and P. Pryfogle, Biotechnol. Bioeng., 37 (1991) 266. 24. D. Cho, S. Lee, S. Park and A. Chung, J. Environ. Sci. Health, Part A: Environ. Sci. Eng., 29 (1994) 389. 25. N. Kuyucak and B. Volesky, Biotechnol. Bioeng., 33, (1989) 809. 26. A. Davis and J.O. Leckie, J. Colloid Interface Sci., 63 (1978) 480. 27. C. Westall and H. Hohl, Adv. Colloid Interface Sci., 12 (1980) 265. 28. H. Bold and W.H. Van Riemsdijk, Chapter 13, Soil Chemistry B. PhysicalChemical Models, G.H. Bolt (ed.), Elsevier, Amsterdam, 1982. 29. E. Reed and M.R. Matsumoto, National Conference on Environmental Engineering, ASCE (1990) 420. 30. C. Westall, J.L. Zachary and F.M.M. Morel, MINEQL: A Computer Program for the Calculation of Chemical Equilibrium Composition of Aqueous Systems, Technical Note No. 18, Massachusetts Institute of Technology, Cambridge, MA, 1976. 31. T. Biegler, Comput. Chem. Eng., 9 (1985) 245. 32. A. Wilczak and T.M. Keinath, Wat. Environ. Res., 65 (1993) 238. 33. L. Lo and J.O. Leckie, J. Chinese Inst. Eng., 12 (1989) 579. 34. C.P. Huang and E.H. Smith, Removal of Cd (II) from Plating Waste Water by an Activated Carbon Process, Chemistry in Water Reuse, vol. 2, W.J. Cooper (ed.), Ann Arbor Science Publishers Inc., Ann Arbor, MI, 1981. 35. A. Netzer and D.E Hughes, Wat. Res., 18 (1984) 927. 36. M. Perez-Candela, J.M. Martin-Martinez and R. Torregrosa-Macia, Wat. Res., 29 (1995) 2174. 37. J. Rubin and D.L. Mercer, Sep. Sci. Technol., 22 (1987) 1359. 38. S. Low, C.K. Lee and C.H. Tai, J. Environ. Sci. Health, Part A: Environ. Sci. Eng., 29 (1994) 171. 39. L. Corder and M. Reeves, Appl. Biochem. Biotechnol., 45 (1994) 847. 40. Y. Konishi., S. Asai, J. Shimaoka, M. Miyata and T. Kawamura, Ind. Eng. Chem. Res., 31 (1992) 2303. 41. L. Lo and J.O. Leckie, Proceeding of National Science Council, ROC, 13 (1989) 109. 42. S. Yiacoumi and C. Tien, J. Colloid Interface Sci., 175 (1995) 333. 43. S. Yiacoumi and C. Tien, J. Colloid Interface Sci., 175 (1995) 347. 44. C. Papelis, P.V. Roberts and J.O. Leckie, Environ. Sci. Technol., 29 (1995) 1099. 45. D. Chen, Z. Lewandowski, F. Roe and P. Surapaneni, Biotechnol. Bioeng., 41 (1993) 755.

317 46. L. Apel and A.E. Torma, Can. J. Chem. Eng., 71 (1993) 652. 47. C. Yao, A Study of a Few Problems in Adsorption, Ph.D. Thesis, Syracuse University, Syracuse, NY, 1991. 48. C. Yao and C. Tien, Chem. Eng. Sci., 48 (1993) 187. 49. C. Tien, Adsorption Calculations and Modeling, Butterworth-Heinemann, Boston, MA, 1994. 50. C. Hindmarch and G.D. Byrne, EPISODE: An Effective Package for the Integration of Systems of Ordinary Differential Equations, Lawrence Livermore Laboratory Report UCID-30112, 1976. 51. R. Aris, Ind. Eng. Chem. Fundam., 22 (1983) 150. 52. D. Do and R.G. Rice, Chem. Eng. Sci., 42 (1987) 2269. 53. L. Riekert, AIChE J., 31 (1985) 863. 54. D. Do and R.G. Rice, Chem. Eng. Sci., 43 (1988) 2897. 55. J. Chen, J.-T. Yoon and S. Yiacoumi, Investigation of Copper Ion Sorption in Activated Carbon Fixed-Bed Columns, submitted for publication in Wat. Res. (1997). 56. E. Reed and S. Arunachalam, J. Environ. Eng., 120 (1994) 416. 57. A. Shay and J.E. Etzel, Treatment of Metal-Containing Wastewaters by Carbon Adsorption of Metal-Chelate Complexes, Proceeding of the 46 th Purdue Industrial Waste Conference, Lewis Publishers, Inc., Chelsea, Mich, 1992. 58. R. Bowers and C.P. Huang, Wat. Res., 9 (1987) 1031. 59. C.P. Huang and P.K. Wirth, J. Environ. Eng., 108 (1982) 1280. 60. W. Kuennen, R.M. Taylor and K.V. Dyke, J. Am. Water Works Assoc., 84 (1992) 91. 61. S. Watson, C.D. Scott and B.D. Faison, Appl. Biochem. Biotechnol., 20/21 (1989) 699. 62. S. Watson, C.D. Scott and B.D. Faison, Emerging Technologies in Hazardous Waste Management, 173, D.W. Tedder and F.G. Poland (eds.), Am. Chem. Soc., Washington, DC, 1990. 63. B. Volesky and I. Prasetyo, Biotechnol. Bioeng., 43 (1994) I010. 64. R.J.R. Turney, D.A. York, C.F.V. Mason, C.J. Chisholm-Brause, D.C. Dander, P.A. Longmire, D.E. Morris, R.K. Strait and J.S. Brewer, Heap Leach Studies on the Removal of Uranium from Soil, LA-UR-94-1361, Los Alamos National Laboratory, Los Alamos, NM, 1994. 65. T. Yeh, HYDROGEOCHEM: A Coupled Model of HYDROlogic Transport and GEOCHEMical Reactions in Saturated-Unsaturated Media, Department of Civil Engineering, The Pennsylvania State University, University Park, PA, 1990. 66. D. Langmuir, Geochim. Cosmochim. Acta, 42 (1978) 547. 67. L. Lu and T.N. Narasimhan, Water Resour. Res., 25 (1989) 869. 68. L. Lu and T.N. Narasimhan, Water Resour. Res., 25 (1989) 883.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

319

T r a c e m e t a l c a t i o n a d s o r p t i o n i n soils" s e l e c t i v e c h e m i c a l e x t r a c t i o n s and biological availability J.H. Rule Department of Ocean, Earth and Atmospheric Sciences, Old Dominion University, Norfolk, VA 23529 USA 1. I NTR ODUCT I O N Metal cations in soil systems may occur in many geochemical forms: free or complexed ions in soil solution, ions held to charged surfaces (exchangeable or specifically sorbed) and metal (Cd, Cu, Pb, etc.) hydroxides and carbonates. Metals may also be associated with Fe and Mn oxides (or oxyhydroxides) and A1 hydroxides, bound within organic matter, incorporated into sulfides and bound within lattice structures of phyllosilicates and resistant primary minerals. The speciation of trace metals in soils determines the availability of metals for plant uptake as well as the potential for ground water contamination. This chapter addresses the nature of metal ions adsorbed to soil solids as interpreted by selective chemical extractions. Adsorbed metals may be classified as being bound by processes of general adsorption or specific adsorption, known as chemisorption. Metals in soil geochemical phases other than exchangeable (adsorbed) ions are in chemical equilibrium with the adsorbed metals and serve as a reservoir for metals in the soil solution and sorbed forms. Outside of the soil solution species, adsorbed metals are the most bioavailable geochemical forms in the soil and are of great interest with respect to bioavailability and potential ground water contamination. Metals in geochemical phases other than adsorbed (exchangeable) ions are often available to living organisms and also need to be considered. One of the most recent geochemical applications is the use of chemical extractions to provide rapid assessment of the bioavailable cations in contaminated soils and to estimate the long-term reservoir of potentially bioavailable and leachable metals. 2.

SOIL SOLIDS

The solid phase of soils consists of primary and secondary minerals, amorphous compounds and organic matter. The secondary minerals and amorphous compounds result from the weathering of primary minerals, and to a certain extent, reflect their composition. Weathering and soil formation from diverse

320 types of rocks and other p a r e n t materials give rise to the compositional differences of various soil types. Extreme w e a t h e r i n g may, of course, produce similar mineralogical and compositional suites from a diverse group of p a r e n t materials. The n a t u r e and extent of m a n y physico-chemical reactions in soils will vary as a function of the soil type due to these compositional differences. As the p r i m a r y minerals w e a t h e r and the abundance of the secondary minerals increases, a corresponding reduction in average particle size occurs (Figure 1). One of the most i m p o r t a n t chemical reactions in soils is ion exchange which m a i n l y occurs on clay (< 2 pm) and colloidal sized (< 1 ~m) particles. 1

secondary minerals

Z

9 9

\ \

primary

W

W O2

minerals .......

\

~

quartz, feldspars, m!cas, _resistant

Sand

i

50

phyllosi licate clays, AI, Fe, Mn oxides, non-crystalline

aluminosilicates \ \

Silt

i

2

Clay

PARTICLE SIZE(#m)

Figure 1. Standard abundance of primary and secondary minerals in various size fractions of soil (redrawn after [6]).

Ion exchange in soil materials occurs due to the surface electrical charges of clays and organic particles. These materials m a y carry either a net negative or net positive charge. Most of the phyllosilicate clays, Mn oxides, some amorphous aluminosilicates and organic m a t t e r in soils generally have net negative charges while hydrous A1 and Fe oxides are usually net positively charged. The amphoteric hydrous Fe oxides have a net positive charge at pH generally less t h a n 7 and a net negative charge above pH 7. The specific pH at which the net charge changes, the isoelectric point, is a function of the type and concentration of

1The terms "clay" or "colloid" will be used in this chapter to refer to all materials in the respective size class unless specific minerals or compounds are mentioned.

321 complimentary ions in the specific soil environment. Net negative charges commonly predominate, except in very old soils, notably those from the tropics. The negative charge of phyllosilicate clays arises from two sources: (1) constant or p e r m a n e n t charges: ionic substitution of a cation of lower valence within the structural layers or (2) variable or pH dependent charges: dissociation of H § from -OH groups on the edges of the clay particle. Dissociation or exchange of H § from -OH groups are thought to be the major source of negative sites for cation exchange in hydrous oxides of A1, Fe and Mn and amorphous aluminosilicates. Negative charges on organic m a t t e r originate primarily from H § dissociation of phenolic (-OH) or carboxylic (-COOH) groups. The degree of H § dissociation is strongly influenced by the pH and therefore, the number of sites available for cation exchange varies with pH. At lower pH values, the H § ions are held much more strongly t h a n at higher pH levels and are not easily replaced by other cations. As a result, the capacity for cation exchange for all these exchangers is generally less at lower pH t h a n at higher pH. Ions in the soil solution may be attracted to these electrically charged surfaces. Since negative charges predominate in most soils, cation exchange dominates anion exchange and will be considered in detail in this chapter. Cation exchange in soils occurs when cations in the soil solution move close to the exchanger surface and, due to a stronger force of attraction, displace an ion already held near the surface. The strength of adsorption of metal cations to different soil constituents varies as a function of both ion and exchanger characteristics as well as the chemical nature (pH, ionic strength, etc.) of the soil environment. Soluble cations may arise from several sources in a soil system: weathering of primary and secondary minerals, addition of fertilizer and liming materials, atmospheric precipitation, decay of organic m a t t e r or anthropogenic inputs. Cation attraction to negatively charged surfaces is depicted in Figure 2. An electric double layer, consisting of the surface of the exchanger and the first row of cations, is partitioned from the remaining ions in the soil solution by the Stern layer. The cationic density decreases with distance away from the surface in approximately a logarithmic relationship. Various intermolecular reactions occur at the solid-solution boundary and may be involved in "sorption" reactions [1]: 1. van der Waals' forces 2. Ion-dipole forces 3. Hydrophobic and hydrogen bonding 4. Charge transfer 5. Ion and ligand exchanges 6. Chemisorption (specific adsorption) 7. Magnetic bonding. All of the above mentioned soil components are reactive and affect the soil solution ion concentration by either ionic interactions at phase surfaces or by

322

Stern layer 44-

4-

4-

4-

.

4444-

44-

4-

-8 iJJ

4-

4-

44-

03 t-

_ 444-

4-

4-

44-

4-

444-

4-

44-

4-

4-

4-

4-

4444-

Diffuse ion layer Electric double layer (EDL)

Figure 2. Interaction of ions with a negatively charged clay particle and ionic distribution with distance away from the surface.

Figure 3. Dynamic trace metal equilibrium between major soil phases depicting ionic transference through the soil solution.

323 precipitation-dissolution reactions. Soil components that are involved in sorption of trace metal cations are: 1. Phyllosilicate clays 2. Hydrous, amorphic oxides of Fe and Mn, and to a lesser degree, A1 and Si 3. Organic m a t t e r 4. Carbonates, hydroxides, phosphates and sulfides. The clay minerals, organic m a t t e r and hydrous metal oxides are the most important groups that participate in various sorption reactions of metal cations. There is competition and equilibrium between all charged solid phases and cations in the soil system. The soil solution is the medium through which ions migrate to and from the various surfaces (Figure 3). Bulk ions in the soil solution generally determine its ionic strength and may influence the behavior of the less abundant trace metal cations. 3.

CATION S O R P T I O N

3.1. Specific and nonspecific sorption Sorption of a cation is the result of the various forces described above, the nature of the cation which is generally best characterized by the ionic potential (charge/radius), the nature of the sorbing surface, charge density or intensity and physical nature of the surface. Each of the solid phases in a soil reacts differently with an individual cation or group of cations and the cation exchange capacity (CEC) of these phases also varies (Table 1). The units for the CEC values are centimoles of positive charge per kilogram of soil (cmolc kg-1).

Table 1 Representative cation exchange capacities of common exchange materials in soils as measured at pH 7.0 Cation exchange capacity (CEC) Exchanger (Soil Phase) cmolsc kg -1 Organic m a t t e r Vermiculite Allophane Smectite (montmorillonite) Chlorite Illite Kaolonite Hydrous oxides

100 100 100 60 20 20 2 2

-

300

-

150

-

150

-

I00

-

40

-

40

-

16

-8

324 When exchangeable ions are measured, the major portion of the sorbed ions on all solid surfaces is included. However, some sorbed ions are held very strongly through specific adsorption, also termed chemisorption, and are not extracted as exchangeable ions. General or non-specific adsorption is attributed to the formation of outer-sphere complexes of the metal and exchanger surface. Specific sorption involves formation of inner-sphere complexes [2,3]. In uncontaminated soils, most trace metal concentrations in solution are very low, much below what is predicted by the ionic strength and composition of the soil solution and mineral solubility. The most common reason is chemisorption of the metal cations. The limited number of chemisorption sites is generally sufficient to bind the available trace metal cations, even against a large concentration imbalance with the major cations in the soil solution. Dominant exchangeable cations in soils of humid regions are Ca ++, H § A1§ and complex Al-hydroxy ions while Ca ++, Mg §247 and Na § dominate in low-rainfall area soils [4]. When concentrations of trace metals in soils are increased notably above background levels, specific adsorption sites become saturated and the remaining amounts of trace metal cations will not compete well for simple adsorption sites against high levels of the major cations. During recent experiments using serial batch leaching of Cd, Pb and Zn from contaminated surface soil downward into a non-contaminated, low CEC soil, the three trace metals quickly saturated available specific sorption sites and then migrated, without further attenuation, due the presence of high concentrations of Ca and Mg in the soil solution. Although Cd, Pb and Zn were initially removed from solution by chemisorption, the sites available for the three trace metals in this situation comprised a small percentage of the total sites available for sorption. Addition of organic matter to the contaminated soil caused the formation of soluble organic Cd, Pb and Zn complexes which reduced the amount of trace metal sorbed by the underlying soil [5]. Specifically sorbed ions may also be displaced by more strongly held ions or may be released to the soil solution if the solid phase to which they are attached is affected by dissolution processes. Chemisorption of metal cations is sometimes difficult to distinguish from precipitation processes, especially at high pH levels [6]. The method by which cation adsorption is traditionally measured contributes to this difficulty. Usually, the pH is adjusted over a given range and the amount of metal removed from solution is determined from the change in solution concentration. Using this method, the solid phase is not examined and sorption cannot be distinguished from precipitation of metal oxides or hydroxides. Trace metal cations hydrolyze increasingly as pH increases and the hydrolyzed forms are apparently sorbed to exchange sites. The presence of oxide and hydroxide phases in the soil may serve as sites for hydrolyzed metals and/or as nuclei sites for precipitation. The combination of sorption, hydrolysis and hydroxide precipitation creates a process that usually appears as a "sorption continuum" which is described as a smooth sorption isotherm [6]. One technique that gives some insight into distinguishing between sorbed and precipitated metal cations is selective sequential extraction

325 following the sorption step of the isotherm process. Results of such an experiment are presented in Section 7.1. Since many of the trace metal cations that are soil contaminants are bound by mechanisms other t h a n chemisorption, study of phases other t h a n exchangeable is necessary to understand their geochemical behavior and fate. Several studies have shown that metal bioavailability in soils is commonly related to the amount of exchangeable ions while other research has discovered that metals in other geochemical phases are also correlated to biological uptake. In order to characterize and model both the geochemical behavior and bioavailablity of the trace metal cations, study of phases in addition to exchangeable is very important.

3.2. Ion e x c h a n g e of soil phases Trace metal cations are sorbed to different solid phases with variable strengths. Understanding the nature of differential sorption of trace metal cations in soils of variable composition is important in predicting contaminant behavior. Preference order of sorption of cations for the major soil phases is difficult to predict and varies notably with different exchangers (Table 2). The more electronegative metals theoretically form the strongest covalent bonds with O atoms on any mineral surface. However, on the basis of electrostatics (ionic potential), the strongest bond should be formed by the metal with the greatest charge/radius ratio. Predicting the order of bonding via chemisorption is difficult because of these two contrasting characteristics (Table 2; [6]). Electrostatics also predict that M § metals would chemisorb in preference to the divalent metals. Markedly different sorption selectivities are evident between different soil components and few sequences are similar to the ones predicted by either electronegativity or ionic potential (Table 2). Sorption preferences for the trace metal cations onto the silicate clays is somewhat variable and does not seem to reflect the type of clay (Table 2). Sequences for sorption onto goethite varied in different studies and differed from that reported for a freshly precipitated Fe oxide gel (Table 2). One would expect that the charge, size and surface area differences between the freshly precipitated Fe oxide and the more aged goethite would account for some of the observed deviations. The order of selectivity of divalent cations for soil organic m a t t e r does not follow the sequence as predicted from electronegativity or any other single factor. The order of stabilities of complexes between divalent ions and soil fulvic and humic acids vary somewhat from that of whole organic m a t t e r but the same general pattern is reported for all ([19]; Table 2).

326 Table 2 Selectivity of Trace Metal Cations in Different Soil Materials (modified after Yong et al., 1992) Material

Selectivity Order

Refs.

G e n e r a l - based on electronegativity

Cu>Ni>Co>Pb>Cd>Zn>Mg>Sr

[6]

G e n e r a l - based on ionic potential

Ni>Mg> Cu> Co> Zn> Cd> Sr> Pb

[6]

Kaolinite clay (pH 3.5-6)

Pb>Ca>Cu>Mg>Zn>Cd Cd>Zn>Ni

[7] [8]

Kaolinite clay (pH 5.5-7.5) Montmorillonite clay (pH 3.5-6)

Ca>Pb>Cu>Mg>Cd>Zn

[7]

Montmorillonite clay (pH 5.5-7.5)

Cd=Zn>Ni

[8]

Illite clay (pH 3.5-6)

Pb>Cu>Zn>Ca>Cd>Mg

[7]

A1 oxides (fresh, amorphous)

Cu >Pb>Zn>Ni>Co>Cd>Mg>Sr

[9]

Fe oxides (fresh, amorphous)

Pb> Cu> Zn> Ni> Cd> Co> Sr>Mg

[9]

Goethite

Cu> Zn> Co> Pb>Mn

[10]

Goethite

Cu>Pb>Zn>Co>Cd

[11]

Goethite

Pb>Zn>Cd>T1

[12]

Mn oxides

Mg> C a>B a>Ni> Zn>Mn> Co

[13]

Soil organic matter

Cu>Ni>Pb>Co>Ca>Zn>Mn>Mg

[6]

Fulvic acid (pH 3.5) Fulvic acid (pH 5.0)

Cu>Fe>Ni>Pb>Co>Ca>Zn>Mn>Mg Cu>Pb>Fe>Ni>Mn~Co>Ca>Zn>Mg

[14,15] [14,15]

Humic acid

Cu> Zn> Ni> Co>Mn

[16]

Humic acid (pH 4-6)

Cu>Pb>Cd>Zn

[17]

Mineral soils (pH 5, no organics) Mineral soils (20-40 g kg -1 organics)

Pb>Cu>Zn>Cd

[18]

Pb>Cu>Cd>Zn

[18]

Selectivity of metal complexation with soil organic matter depends upon several factors other than properties of the metals themselves, including [6]: 1. The chemical nature of the organic ligands (types of functional groups) 2. The amount of sorption on the organic matter 3. The pH at which sorption is measured (variable competition of M §247for H § on functional groups) 4. The ionic strength of the soil solution (influences competition by other cations for complexing sites).

327 Although the above cited studies reported generally similar findings, variations in selectivity exists even for the "same" substance. Differences in the geochemical conditions, e.g., ionic strength (I), counter ions, matrix cations, of the system affects ionic competition and sorption. The pH of the system significantly influences sorption on all types of surfaces present, especially oxides and organic matter where the charges are predominately pH dependent. The ionic strength influences sorption processes on all types of surfaces due to competition for sites between the dominant cation(s) and trace metal cations. When multiple solid phases are present, the competing factors and interactions become even more complex.

3.3. Ion e x c h a n g e r e a c t i o n s Although cation exchange is not a chemical reaction in the conventional sense, the process of ion exchange is represented in an analogous manner. The exchange of aqueous Cd § in the soil solution for Ca § that is adsorbed onto a soil particle is written as CaX + Cd++ ~ - Ca++ + CdX where X represents a solid phase exchanger. Applying the law of mass action, the ion exchange equilibrium constant, Keq, can be written as _ (Ca ++ ). (CdX) K e q - (Cd++). (CaX) or expressed as ICa++{ : ke x (CaX) (Cd ++ (CdX) where kex is the selectivity coefficient [20]. The value of kex indicates the preferential adsorption of one cation over the other. Where kex is equal to unity, equal amounts of both ions are sorbed. This simple approach is a very useful technique in binary cation systems to measure selectivity of an exchanger for a cation in very low solution concentrations as compared to the second cation present at much higher concentrations. This is an excellent method to determine the amount of trace metal cations attracted to soil solids via chemisorption. By adding low concentrations of the trace metal of interest to a solution containing a much greater concentration of a major cation such as Ca ++ , the sorption of the trace metal via chemisorption should be measurable. An excellent summary of several other important mathematical equations used to describe cation exchange of different types of exchangers and ions of unequal valence is given in [20]. The

328

general adsorption isotherms of three classical types are shown in Figure 4. Significant chemisorption of metal cations by an exchanger would be described by the high affinity type of curve. Most cation exchange is described by the Langmuir, or a similar relationship. The constant adsorption curve is generally considered a theoretical line for adsorption. However, a linear relationship may exist for many adsorbers at low cation concentrations and for u n u s u a l c a s e s .

High affinity type

"5

'/

~,o~

0 '4-0 0 CO ~J

E :3

..0 23 OLU

Equilibrium conc of solute in solution

---

Figure 4. General adsorption isotherms of three classical types (modified from [77]).

Ion exchange studies are complicated for multi-element systems due to the interaction and competition between the ions. The major problem is the near impossibility of measuring isotherms for all combinations when there are three or more competing ions [22]. The required amount of exchange data increases as the number of species increases. For a binary system, there is only one exchange isotherm for a given total solution concentration and measurement in this system is not difficult. As the number of competing ions increases, the data requirement rapidly expands. For a ternary system (three competing cations), there are three unique binary combinations of ions and for n cations, there are n(n-1)/2 unique binary combinations. One study compared several computational methods for predicting ternary exchange from binary isotherms [22]. The methods were ranked with respect to their predictability on the basis of the root mean squares of differences between measured and predicted ternary exchange data in NHn-Ba-La or Na-K-Ca systems. The conclusions ranked the performance of the methods relative to each other but the suitability of any method to reliably predict ternary exchange was not established. This is obviously an area that warrants further research.

329

4.

SOIL GEOCHEMICAL PHASES

Behavior and interactions of cations with solid phase surfaces vary with the different types of solids and cause difficulties in understanding metal behavior in soils due to these differences in interactions. In an ideal situation, each of the soil phases would be isolated and the amount and nature of the metal cations present determined. Several studies have been conducted using artificially prepared compounds, such as Fe and Mn oxides, or purified soil extracts of organic matter, i.e., humic or fulvic acids [23,24]. However, these prepared compounds are sufficiently different from natural soil components to give only an approximate understanding of soil reactions. When all components are present, solid phase competition and interactions further complicate the evaluation. Another approach, discussed below, is to selectively extract metals associated with specific solid geochemical phases from soils. Most commonly, the geochemical phases are "operationally defined" and target the following phases (modified from [1]): 1. Free or complexed ions in soil solution 2. Ions bound to charged surfaces (sorbed or exchangeable) by processes of general adsorption or chemisorption 3. Metals (Cd, Cu, Pb, etc.) present as hydroxides and carbonates 4. Metals associated with Fe and Mn oxides or oxyhydroxides and A1 hydroxides 5. Metals bound within organic matter, excluding exchangeable cations 6. Metal sulfides in reduced soils and as sulfide minerals 7. Metals bound within lattice structures of phyllosilicates and resistant primary minerals.

4.1. Water soluble phase Metals exist as free cations or as organic or inorganic complexes in the soil solution. The predominant form of a given cation depends upon the organic or inorganic ligands available, competing ions and stability coefficients of possible forms. For example, Cu has an extremely high stability coefficient with most forms of soluble soil organic m a t t e r and most solution Cu exists as organic complexes. Generally the concentrations of trace metal cations in soil solutions are very low, due to chemisorption processes as previously discussed. 4.2. E x c h a n g e a b l e / s p e c i f i c a l l y adsorbed phase Most trace metal cations are sorbed to charged surfaces by chemisorption and in uncontaminated soils, only small amounts are readily exchangeable. However, when significant amounts of anthropogenic metals are present in soils, proportions in the exchangeable fraction may be unusually high. After all of the chemisorption sites are filled, the trace metals must compete with the major cations for the general adsorption sites. When a high proportion of a trace metal is present, a notable amount may be sorbed to the general sites. With a significant excess of major cations, this competition may result in minimal

330 significant excess of major cations, this competition may result in minimal adsorption of the trace metals. These metals may, however, react with specific soil phases via a variety of mechanisms. For this reason, determination of the metal distribution in other geochemical phases is extremely useful in ascertaining the nature and degree of soil contamination.

4.3. Carbonate phase This phase may be an important reservoir for trace metals at near neutral soil pH levels. The major carbonates in most soils are those of Ca and Mg. Carbonates contain various amounts of trace metal cations incorporated into the structure, generally by coprecipitation, of the various mineral forms present in soils. Hydroxide and oxide forms are also important in governing trace metal behavior in the soil system. Most trace metal cations precipitate as oxides or hydroxides at higher pH levels. Selective extraction of the carbonate fraction in soils should also remove the oxides/hydroxides of the trace metals. 4.4. Fe-Mn oxide phase The most predominant oxide/hydroxide phases in most soils are those of A1 and Fe with Mn oxides being less abundant. In old and ancient soils, these oxides are the dominant clay minerals. Forms of Fe vary from highly hydrated amorphous hydroxides, such as ferrihydrite, to crystalline oxides such as goethite. The term oxy-hydroxide is often used to describe the variably hydrated amorphous forms of Fe that are prevalent in most soils. The predominant form of Mn in soils is MnO2. All of these oxides rarely have isomorphous substitution of cations in their structures and consequently have low CECs, especially relative to their large surface area. The greatest degree of reactivity is from the Fe and Mn oxide forms. The Fe oxide/hydroxide phases generally lose hydration water, decrease in solubility and increase in crystallinity with age. Differentially extracting Fe oxides of varying solubility/crystallinity may sometimes distinguish between more recent and historical inputs of trace metals into the soil system. Metals entering the soil system thousands of years in the past should be mostly associated with crystalline or moderately crystalline forms while "recent" additions should be mostly incorporated with the easily reducible or amorphous forms. This technique works best in well drained upland soils that have not undergone periods of oxidation-reduction during profile development. Trace metal cations are associated with the Fe and Mn oxide/hydroxide phases in soils by the following processes: (a) sorption onto external surfaces, (b) solidstate diffusion of metal cations and (c) cation binding and fixation at sites within the solid particles [25]. Variable charges at oxide surfaces may also promote sorption of anions which, in turn, may attract trace metal cations. Metal hydroxide formation at Fe-Mn hydroxide surfaces may also be of importance, as discussed above.

331

4.5. Organic phase Humic substances comprise about 60-80% of soil organic matter with the balance consisting of non-humic materials. Humic substances are characterized by aromatic, ring-type structures that include polyphenols and comparable polyquinones. Humic substances in soil are classified into three chemical groups based on solubility. The first group is fulvic acid, with lowest molecular weight (1000 to 5000) and lightest color, soluble in both acid and alkali and most susceptible to microbial attack. The second group is humic acid, of medium molecular weight (10000 to 100000) and color, soluble in alkali but insoluble in acid and intermediate in resistance to degradation. The third group is humin, with highest molecular weight (>100000), insoluble in both acid and alkali and most resistant to microbial decomposition and darkly colored[4,6]. The nonhumic portion is composed of (1) polysaccharides, polymers which have sugarlike structures and a general formula of Cn(H20)m, where n and m are variable; (2) polyuronides, not found in plants but synthesized by soil microorganisms; and (3) low molecular weight organic acids and some protein-like materials, present in relatively small quantities [4]. Moderate amounts of most trace metals are associated with the organic fraction in most mineral soils unless the texture is very sandy, or the organic content is unusually high. In these soils, a high proportion of the trace metals is bound to the organic materials. 4.6. Sulfide p h a s e Significant concentrations of trace metal sulfides are not common in soils, especially those having good drainage and aeration but could be expected in reduced or partially reduced soils that contain significant amounts of S. Thus sulfides would not be expected in normal agricultural soils but may be encountered in rice paddy or wetland soils. Common forms might be pyrite or sulfides of any of the trace metals present in notable concentrations. 4.7. R e s i d u a l p h a s e The most important primary minerals that might be present, especially in young soils, are olivine, pyroxenes, amphiboles, micas, feldspars and the silica group, including quartz. Important secondary minerals are the silicate clay minerals (phyllosilicates): illite, chlorite, vermiculite, smectites, especially montmorillonite, and kaolinite. As mentioned above, while soils age and weather, the proportion of primary minerals decreases and the proportion of secondary minerals and amorphous forms increase. 5.

SELECTIVE SEQUENTIAL EXTRACTIONS

Numerous schemes have been devised to selectively extract different geochemical phases from soils and determine metals present in each phase. Most commonly used are schemes of sequential selective chemical extractions which

332 were first developed for use with sediments in geochemical exploration and contaminated sediment characterization [26-29]. Other methods were developed directly for soil extractions [30-33] but a sediment-developed extraction sequence [29] is commonly cited as the basis for most soil extraction procedures. These methods are ideally designed to affect only the target phase or species with minimal influence on other phases, which is difficult to accomplish. A few studies have shown that although multiple phases are affected by each extractant, the influences are often minimal and not sufficient to preclude beneficial use of this technique [34,35]. The order in which the extraction reagents are used is critical to obtaining the most phase specific data since some of the chemicals will affect more than one phase if added to a previously untreated soil sample [33,36,37]. For example, the strong H202 used to oxidize organic matter will also oxidize MnO2 unless the Mn02 has been removed with a prior extractant. One of the most complete extraction schemes was outlined in [38] as: 1. soluble in water (aqueous phase) 2. exchangeable or unspecifically adsorbed 3. specifically adsorbed 4. bound to carbonates 5. bound to organic matter 6. bound to Mn oxides and hydroxides 7. bound to amorphous Fe (and A1) oxides and hydroxides 8. bound to crystalline Fe (and A1) oxides and hydroxides 9. bound to sulfides 10. bound to silicates (residual fraction). A discussion of the utility and general procedures for determination of each of these phases is helpful in evaluating the applicability of determining a given fraction for particular studies. Following this general discussion, a practical extraction scheme for evaluating contaminated soil is presented. 5.1. Water s o l u b l e p h a s e Although the water soluble fraction is the most bioavailable and subject to leaching, most researchers do not include this aqueous phase in their extraction schemes. The rationale is that concentrations of trace metals in the soil solution are extremely low and generally not of significance when compared to the exchangeable fraction. Additionally, the water soluble phase is co-extracted with the exchangeable phase metals. While this is true for most soils, contaminated soils may have notable water soluble concentrations and this fraction should be determined when investigating soils that might have been subjected to anthropogenic inputs. For example, an unusual characteristic of the soils at a contaminated site may be uncommonly high levels of Pb in the aqueous phase which should trigger a detailed investigation of the abundance and forms of Pb in the soils. Some researchers [39,40] have reported significant concentrations of Cd, Ni, Pb and Zn in the water soluble fraction of trace metal contaminated soils. This simple step involves extraction of the sample with deionized water.

333

5.2. Exchangeable phase Exchangeable cations are measured after displacement with a cation from a neutral salt solution equilibration. Several different types of salts t h a t were used in earlier procedures have largely been discarded due to various difficulties. Reagents should be carefully evaluated depending on specific objectives of the extraction scheme. One of the first extractants was NH4OAc because of its widespread use as a soil test reagent. This extractant is now rarely used to determine exchangeable cations due to the analytical interferences and the possibility that carbonates and hydroxides may be affected [29]. Calcium and Mg chlorides are commonly used to determine exchangeable cations but due to the formation of chloride complexes of Cd and Pb, studies suggest t h a t phases other that simple exchangeable forms are affected [41]. Mg is a harder Lewis acid (a species that can accept an electron pair) than Ca which indicates that Mg may displace specifically adsorbed trace metals [42]. This consideration that may discourage its use as an extractant. Both Mg(NOa)2 and NH4NOa were proposed as extractants because of their efficiency for displacing adsorbed cations as well as volatility which decreases background interferences during AAS analysis [37,43]. Extraction with NH4NOa seems to be the most overall desirable relative to extraction efficiency and low analytical interferences.

5.3. Specifically adsorbed phase Very few studies methods are reported for determination of specifically adsorbed (chemisorbed) metals [33,44-46]. Most of these methods utilize reagents (NH4OAc, NH4NOa, Na2EDTA, HOAc) that were also used for other phases and are not specific for extraction of this phase. A few researchers [33] and [45], who modified the method of [33], utilized Pb(NOa)2 in dilute CaC12 to determine specifically adsorbed metal cations. The Pb §247 is an appropriate cation for displacing most other trace metals due to its low pK (7.7) and large atomic radius which suggest that Pb would be effective in determining specifically sorbed ions. If Pb were of interest in the soil, another cation such as Cu (pK = 7.7, smaller atomic radius than Pb) would have to be utilized for the extraction.

5.4. Carbonate phase The extractants often employed to remove trace metals bound to carbonates are the acids HC1 and HOAc (pH 3-3.5), a buffer solution of HOAc/NaOAc (pH 5) and a buffered complexing agent Na2EDTA (pH 4.6) [29,47,48] The most frequently used is the buffer solution of HOAc/NaOAc (pH 5) which is thought to have minimal effect on other soil phases. Hydroxide phases of alkali, alkali-earth and trace metals are also affected by this extractant and may explain why notable amounts of trace metals are extracted in this phase in slightly acidic soils.

334

5.5. Organic/sulfide phase Three reagents most frequently used for extraction of the organic phase in soils are: acidified 30% H202, K4P207 and NaOC1 [29,49]. The H202 procedure has long been used for extraction or removal of organic matter from soils and sediments and is the most often reported method used for speciation of trace metals in soil organic materials [36,37,40]. F o l l o w i n g the oxidation step with the H202, extraction with NH4OAc is conducted to prevent re-adsorption of the liberated trace metals onto the remaining solid phases. Sulfides are solubilized during the same extraction process as organic matter in oxidative procedures. Methods to separately extract organic forms from sulfides are difficult and their use has rarely been reported with the exception of the use of K4P207. The most simple manner to estimate the proportion of metals associated with each of these two phases is to measure organic carbon and apply a correction. However, the very low solubility metal sulfides are likely to predominate over organic forms in soils where both are present. Another method is to separately determine the acid volatile sulfides and associated trace metals and assign the balance of the metal concentration to the organic form.

5.6. Mn oxide phase The Mn oxides can be either extracted simultaneously with the Fe oxides or as a separate phase. When extracted as a separate phase the reagent of choice for Mn oxides is 0.1M NH2OH-HC1 in 0.01M HNO3 [32,33,40,50].

5.7. Fe oxide phase When Fe-Mn oxides are simultaneously extracted, commonly used reagents are: (1) a heated mixture of 0.04 M NH2OH'HC1 in 25% (v/v) HOAC or (2) 0.175 M (NH4)2C204 + 0.1 M H2C204 (acid ammonium oxalate). The use of Na2S204 in a heated Na-citrate buffer solution for determination of the combined oxides is limited because of analytical interferences. Amorphous Fe oxides are generally extracted with either (1) 0.175 M (NH4)2C204 + 0.1 M H2C204 in dark; (2) 0.04 M NH2OH-HC1 in 25% (v/v) HOAC (heated); or (3) 0.25 M NH2OH-HC1 + 0.25 M HC1. To distinguish crystalline Fe oxides, the reagents of choice are usually (1) acid ammonium oxalate + ascorbic acid (heated); (2) acid ammonium oxalate under UV radiation; or (3) Na2S204 in a Na-citrate buffer solution [29,34,51]. The assortment of reagents, their concentrations and specific procedural conditions used for extraction of the Mn and Fe oxides are rather inconsistent and lead to variable results. Despite these differences in extraction parameters, the methods are all fairly selective as long as each procedure is carefully followed and the proper sequence of extraction is observed: Mn oxides prior to amorphous Fe oxides prior to crystalline Fe oxides.

5.8. Residual phase The residual phase is generally described as silicate structures and resistant minerals that remain after selective removal of the more susceptible geochemical

335 phases. In fact, this phase may contain highly r e s i s t a n t m a t e r i a l s from other phases t h a t were not completely removed. This is especially true for m a n y of the commonly used sequences t h a t do not include the extraction step for crystalline Fe oxides. The Fe content in the residual fraction is usually high in the selective extraction procedure. The types/strengths of mineral acids utilized for extraction/dissolution of the residual phase vary with two p r i m a r y schools of thought. Some researchers deem it necessary to conduct complete dissolution of the residue in order to establish a complete inventory of trace metals in the soil. Other investigators claim t h a t use of a very strong mineral acid will strip away all metals from any mineral residue t h a t may be released to the environment even over short geologic time [29]. The procedures used for these digestions also play a role in the choice of acids. For total dissolution, highly corrosive mixtures of HNO3, HF and usually HC104 are employed, either in open vessels, P a r r bombs or microwave digestion vessels. For strong acid digestion, conc. HNO3, HC1 and HC104 are used singularly or in combination and a few procedures use 30% H202. An abbreviated version of a general procedure for extraction of contaminated soils is as follows. 6.

SELECTIVE EXTRACTION FOR SOILS

1. W a t e r soluble P h a s e (WSP): Add 24 mL of ASTM Type I deionized H20 to 3.0 g sample, shake tubes end-to-end for I hour. 2. Exchangeable P h a s e (EP): Add 24 mL of 1.0 M, pH 7 NH4NO~ to 3.0 g sample, shake tubes end-to-end for 1 hour. 3. Carbonate P h a s e (CP): Add 24 mL of 1.0 M, pH 5.0 NaOAc and place on shaker for 5 hours. 4. Easily Reducible P h a s e (ERP) - predominantly Mn oxides: Add 24 mL of 0.1 M NH2OH-HC1 in 0.01 M HNO3 and place on shaker for 1 hour. 5. Moderately Reducible P h a s e (MRP) - predominantly Fe oxides: add 24 mL of 0.04 M NH2OH-HC1 in 25% HOAc and place samples in a 96 ~ w a t e r b a t h for 6 hours. 6. Organic (Sulfide) P h a s e (OSP): a) add 5 mL 0.02 M HNO3 + 5 mL 30% H202, pH 2.0 and place in 85 ~ w a t e r bath. Over the next 5 hrs, add 3 x 5 mL 30% H202. b) Extract with 7 mL of 4.0 M NaOAc in 20%(v/v) HNO3 for 30 minutes. NOTE: the reaction with peroxide can be very vigorous. 7. Acid Extractable P h a s e (AEP): Digest residue in 150 mL tall form Pyrex beaker with 15 mL conc Trace Metal Grade HNO3 and 5 mL 30% H202 at ~ 100 ~ for four hours on a hotplate. 8. General comments: a) Centrifuge the samples at 7,000 rpm in the HS-4 rotor (rcf = 9600g) in a Sorvall RC2-B refrigerated centrifuge 15 minutes, after each e x t r a c t a n t and after the w a t e r wash.

336 b) Wash the samples with 15 mL of ASTM Type I deionized water after each extractant. c) Remove s u p e r n a t a n t extract with a 5-10 mL Macro-Set or similar pipet.

B

P H A S E D I S T R I B U T I O N O F M E T A L S IN N A T U R A L AND CONTAMINATED SOILS

The origin and forms of trace metals strongly influence their behavior and availability in soils. Lithogenic metals are only slightly mobile and are potentially available to plants only under specific conditions. Pedogenic metal actions reflect particular soil geochemical conditions and anthropogenic metals are generally the most mobile of these three groups [52]. Knowledge of the distribution of trace metals in all of the geochemical phases is often extremely valuable in determining the amount and impact of anthropogenic metals in soils. In uncontaminated soils, a major reservoir for many trace metals is the residual phase while for anthropogenically influenced soils, a much greater percentage of the metals occurs in more mobile forms. Exchangeable metals are the most bioavailable and mobile of the solid phase species, but metals in other geochemical phases are in equilibrium with the exchangeable phase. Multiple and complex chemical forms may be added as contaminants to soils and these chemicals will undergo alterations over time as the soil phases alter towards a new state of equilibrium. Contaminant alteration may produce a supply of metal ions that can be sorbed to exchange sites over time. Measuring only the exchangeable ions at one point in time does not provide a sufficient evaluation of the contaminate status of the soil system. 7.1. M e t a l s in m u l t i p l e f r a c t i o n s Simultaneous examination of trace metals in soil phases is very useful in evaluating contaminant behavior in soils. Very often, patterns of soil enrichment of metals from similar sources or processes, as well as lithogenic or pedogenic metals can be determined. A recent study of the trace metal distribution in soils of four national parks in Poland did not reveal any concentrations high enough to classify the soils as "contaminated". However, enriched concentrations of Cd, Cr, Cu, Ni, Pb and Zn were discovered in soils near industrialized regions. Much of the Cd and, especially, Pb resided in the exchangeable and carbonate fractions indicating recent deposition and relatively high mobility of these two metals. The proportion of exchangeable metal fraction for Pb was greater t h a n for Cd, Ni and Zn indicating greater enrichment of Pb relative to the other three metals. Most of the soil Cu and Ni occurred in the acid extractable (residual) phase suggesting that lithogenic materials were the primary source of these two elements [53]. One new technique that offers unique insights into sorption processes is combined sequential extraction-sorption analysis (CSSA) and the method is well worth utilizing for trace metal sorption studies. Soil or sediment samples are

337 first subjected to standard sorption isotherm methods after which they are sequentially extracted determine metal partitioning into the different geochemical phases. In one study, calcareous clay sediments (pH 7.6-8.0) were individually treated with Cd, Ni or Pb at solution concentrations from 50 to 11000 mg L -1 [54]. Each sample was then sequentially extracted for the following phases: exchangeable, carbonate, Mn oxide, organic and Fe oxide; bulk metals were separately measured. For the untreated samples, Cd was detected only in the oxide phase, an unusual occurrence for this metal, while Ni and Pb were found mainly in the Fe oxide and carbonate phases. After metal sorption, the exchangeable and carbonate phases were dominant for Cd, with the carbonate phase prevailing at sediment concentrations <10000 mg kg -1 but exchangeable Cd dominant at 10000 to 200000 mg kg -1. The carbonate phase was the major reservoir for both Ni and Pb, containing as much as 90% of the total Pb and 75% of the total Ni. Data from the sequential extractions for each of the phases of all metals fit a Langmuir sorption isotherm. This pattern suggests that mechanisms for sorption for each phase operated independently of the other phases, within the realm of sorption competition. The very high proportion of Ni and Pb partitioning into the carbonate phase was expected due to the high pH of the sediments. However, the predominance of exchangeable Cd at the highest soil concentrations was unusual. One reason may be the reportedly slow rate of CdCO3 formation [55]. It should be noted that solution and sediment concentrations in the cited study were extremely high and may not reflect trace metal behavior at low to moderate contamination levels. 7.2. M e t a l s in the e x c h a n g e a b l e fraction As mentioned previously, the proportion of trace metals in the exchangeable phase is generally very low for uncontaminated soils. A major exception is Cd where notable proportions may exist in the exchangeable phase in both natural and contaminated soils [56,57]. In most cases, the proportion of exchangeable Cd is greater in contaminated soils than in those containing only lithogenic or pedogenic metals. However, the chemical forms of the contaminants may control the metal phase distribution, especially in the short-term, and the proportion of exchangeable Cd may be less than for corresponding native soils. The exchangeable Cd in a variety of native soils, including agricultural soils from Norway and Poland and forested soils from Switzerland, ranged between 23-64% of the total Cd [36,56,57]. Other studies have found as little as 13% exchangeable Cd in sandy textured native soils [58]. Exchangeable Cd in several soils contaminated by inorganic sources ranged from 30-48% of total Cd [56,59] while soils treated with biosolids had only 2-29% exchangeable Cd. While it is evident t h a t there is considerable range to the proportion of Cd existing as exchangeable, the average percentages are unusually high when compared to the other divalent trace metal cations. Relative abundances of trace metals in the geochemical phases in background and contaminated soils are presented in Table 3.

338 Table 3 Metal abundances in soil geochemical phases Element

Abundance Order

Background Soils ....... residual>exch>organic~carbonate>Fe-Mn oxide Fe-Mn oxide>residual>>organic>carbonate>exch residual>Fe-Mn oxide>organic>carbonate>exch residual>Fe-Mn oxide>exch~organic>carbonate residual>>carbonate=Fe-Mn oxide>organic>exch carbonate>Fe-Mn oxide>exch>organic>residual exch>Fe-Mn oxide>residual>organic>carbonate residual> Fe-Mn oxide>organic>carbonate>exch ....... Contaminated Soils ....... Cd # exch>residual>Fe-Mn oxide>carbonate>organic Pb # oxide>>carbonate~organic~residual>exch Zn # Fe-Mn oxide~residual>organic>carbonate~exch Cd* residual>carbonate>Fe-Mn oxides>organic>exch Ni* residual>>carbonate>Fe-Mn oxide>organic>exch Zn* carbonate>Fe-Mn oxide>exch>organic>residual Cd residual>>exch>carbonate>Fe-Mn oxide>organic Cu residual>organic> Fe-Mn oxide>carbonate>exch Ni residual>>exch> Fe-Mn oxide>organic>carbonate Zn residual>Fe-Mn oxide>carbonate>organic>exch Cd* carbonate>exch>Fe-Mn oxides>residual>organic Cu* Fe-Mn oxide>carbonate>residual>organic>exch Ni* residual>Fe-Mn oxide>carbonate>organic>exch Zn* Fe-Mn oxide>carbonate>residual>organic>exch Cd # exch>fe-Mn oxide>residual>carbonate>organic Cu # Fe-Mn oxide>organic>carbonate~residual>exch Ni # residual>>Fe-Mn oxide>organic>carbonate>exch Zn # Fe-Mn oxide>residual>carbonate>exch~organic #smelter contaminated soils; *sludge-treated soils Cd Pb Zn Cd Ni Zn Cd Zn

Refs [57] [57] [57] [59] [59] [59] [58] [58] [57] [57] [57] [59] [59] [59] [61] [61] [61] [61] [60] [60] [60] [60] [60] [60] [60] [60]

Very small proportions of Cu, usually < 1%, are found as exchangeable in both native and contaminated soils. Exchangeable Cu ranged from 0.2 - 1% in nine different contaminated soils [60]. About 2% of the total Cu present in contaminated soils near Sudbury, Ontario was measured as exchangeable [61], and the same proportion was noted in the surface horizon of two forested soils in a virtually unpolluted area of Switzerland [36]. Less than 5% of the total Cu was reported to be in the exchangeable form in a study of biosolids treated soils [62]. Relatively small percentages of exchangeable Ni are found in soils, but the proportions are often greater than those for Cu. Exchangeable Ni ranged from

339 8-11% in nine soils that received contaminants from five different sources [60]. Exchangeable Ni was reported to have varied only from 5-8% in both an untreated and sludge treated soils [58]. Even though the concentration of exchangeable Ni doubled in the sludge treated soil, the proportion on the exchange sites remained constant. In one study, after long-term application of biosolids, it was reported that about 10% of the total Ni in soil existed as exchangeable [62]. In uncontaminated and many contaminated soils the fraction of exchangeable Pb is very low, about 5-6% [52,58,62], but in some contaminated soils the exchangeable fraction may be substantial, as high as 25% [39,40]. The proportion of exchangeable Zn varies notably in both native and contaminated soils. These percentages are usually greater t h a n for Cu, Ni and Pb but less than for Cd. No clear pattern of controlling factors is evident from the studies reported in the literature. In an investigation of nine soils that received contaminants from five different sources, a range of 1-10% exchangeable Zn was reported [60]; A range of 3-12% exchangeable Zn was found in both native and contaminated soils in southwest Poland [58]; while another investigation noted that exchangeable Zn was < 15% of the total for a variety of biosolids treated soils [62]. However, results of one study noted that exchangeable Zn increased from 3% to 26% after the soil was treated with sewage sludge [58]. One review paper [63] reported that the order of prevalence of metals in the Easily Soluble/Exchangeable fraction of natural soils was: Cd > Zn > Ni > Pb > Cu. 7.3. M e t a l s in the c a r b o n a t e f r a c t i o n Carbonates, hydroxide and oxide forms of trace metals are extracted by the Carbonate Phase reagent. This phase has been found to contain notable proportions of Cd, Ni, Pb and Zn in both uncontaminated and contaminated soils, even at moderate soil pH values [40,58]. Many soil contaminants originate from various incinerator processes (e.g., fly ash and flue dust) and the metals occur in the forms of oxides which obviously enriches this soil phase. Depending upon the counter ions available, some of these metals may accumulate as sorbed ions as the original forms are altered. Many studies, however, have found rather low concentrations of several of the trace metals in the carbonate fraction. This is the most pH sensitive soil phase and has a notable presence at near neutral or higher pH values and a minor abundance in acidic soils. One study reported 1-3% Cd and 1-10% Zn occurred in this phase in cultivated soils, averaging near pH 6, in Norway [57]. It was noted that < 2% of Cu, Ni and Zn was found in the carbonate phase in a study of contaminated Sudbury soils but these were very acidic soils [61]. The following results for metals in the carbonate phase was reported from a study of nine different contaminated soils, with a pH range of 5.8 to 7.5: Cd: -10%, Cu: 3-11%, Ni: 0.5-2% and Zn: -10% [60]. Biosolids are often lime-stabilized and therefore soils which have received applications of biosolids may have notable proportions of trace metals in the

340 carbonate phase. The following increases of soil metals in this fraction were reported after application of sewage sludge: Cd, from 7 to 25%; Ni, from 14 to 16% and Zn, from 16 to 41% [58].

7.4. M e t a l s in t h e Fe-Mn o x i d e f r a c t i o n Although there is a difference in the selectivity of trace metals by Mn and Fe oxides (Table 2), most researchers use a reagent t h a t simultaneously extracts both oxides. The low abundance of Mn oxides in m a n y soils, especially those of sandy textures, m a y be a reason t h a t this phase is not extracted s e p a r a t e l y from the Fe oxides. The combined oxide phase is generally an i m p o r t a n t reservoir for most of the trace metal cations. Moderate a m o u n t s of Cd and Cu are associated with the oxide phase, in both u n c o n t a m i n a t e d and contaminated soils. Moderate to high proportions of Ni, Pb and Zn are common in the oxide fraction in native and c o n t a m i n a t e d soils. The Fe-Mn oxide fraction in several native soils contained from 29-48% Cd, ~ 4 8 % Pb and 19-39% Zn [56,57]. Distribution of most metals in soils c o n t a m i n a t e d by metal smelters or from biosolids applications was similar for: C d - 20-25%; Cu - 28-43%; and Zn= 34-44%. The proportions of Cd and Cu were approximately 10% less in the smelter affected soils but there was no difference in the proportion of Zn between the two groups of soils. However, the proportion of Ni in the oxide phase was much higher (34%) in the biosolids t r e a t e d soils t h a n those c o n t a m i n a t e d by a smelter (13%). Differences in the chemical forms of Ni a p p a r e n t l y influenced its distribution in the two soil groups, but there was no influence on the distribution of Zn. The identical distributions for Zn m a y have resulted from similar chemical forms in the two sources, which is less likely, or as a result of the strong affinity of Zn for the oxides. 7.5. M e t a l s in t h e o r g a n i c f r a c t i o n The association of metals with soil organic m a t t e r is generally low for Cd and Ni, low to moderate for Pb and Zn and moderate to high for Cu. This relationship would, of course, be different for organic soils or sludges where the majority of the solid m a t e r i a l is organic. In such cases, a high proportion of all metals would occur in the organic phase. No major differences were noted in the distribution of Cd, Ni, Pb and Zn in the organic fraction between a variety of background and c o n t a m i n a t e d soils [56-58,60]. There was little difference in the distribution of Cd, Ni and Zn in the organic fraction of soils affected by two different c o n t a m i n a n t sources but a notable difference in the distribution of Cu between the soils. The following proportions of metal in the organic phase were reported: Cd < 6%; Ni < 9%; Pb and Zn 7-15%. Proportions of Cu were generally 17-33%. There is an effect of added organic m a t t e r on trace metal adsorption and mobility t h a t is not evident from the phase distribution data. Research has indicated t h a t increased levels of dissolved humic materials in leachates from soil or sand columns increased the solubility of Cd, Cu and Zn [64-66]. Simultaneous field and laboratory studies were conducted to investigate the effects on Cd

341 mobility induced by a single liquid sewage sludge application onto a soil [67]. Low-level movement of Cd and soluble organic C from the sludge application site was observed in comparison to the control plot during several weeks following the sludge application. The conclusion was that the mobility of Cd was enhanced by the increased soluble organic m a t t e r from sewage sludge disposal, especially during the period immediately following liquid sludge application. The results of these studies emphasizes the fact that understanding all on-going soil processes is very important in evaluating the exchange, mobility and availability of soil trace metals. 7.6. M e t a l s in t h e r e s i d u a l f r a c t i o n The major concentrations of many metals in native or uncontaminated soils are found in the residual fraction. The proportion in this fraction often decreases significantly when contaminants, usually containing more soluble forms, are applied to soils. High, but variable proportions of trace metals have been reported for the residual fraction for a variety of native soil types: Cd - 23-59%; Ni ~ 80%; Pb ~ 30%; and Zn - 29-70% [56-58]. The proportions of trace metals in the residual phase was usually less in contaminated soils in the same studies" Cd = 16-34%; C u - 11-20%; N i - 50-80%; Pb ~ 10% and Z n - 6-30%. [56-59]. In a study of the phase distribution of metals before and after the application of sewage sludge, the proportion of Cd and Ni in the residual phase changed only slightly but the change in Zn distribution was notable" the proportion of residual Zn decreased from 36 to 6% after sludge application [58]. Similarities exist in the proportion of metals in the residual fractions of native as compared to contaminated soils, but the variation between contaminated soils may be especially notable. A s u m m a r y of metal abundance in the geochemical phases from the cited studies is given in Table 3.

8.

E V A L U A T I O N O F M E T A L B I O A V A I L A B I L I T Y IN S O I L S U T I L I Z I N G SELECTIVE EXTRACTIONS

Plant growth involves interaction of the plant system and soil. Soil is the normal medium for plant growth and the plant's roots absorb nutrients, other elements and water from the soil. The plant's absorption of elements involves processes occurring in both the plant's root and in the soil. Characteristics or mechanisms of either of these two media may notably influence processes of the other. Active ion uptake by the plant roots allows the a t t a i n m e n t of plant metal levels that are in excess of soil concentrations in the mobile metal phases. The depletion of soil concentrations of some elements around the growing roots may induce a redistribution in the soil solution-solid phase partitioning of metals. The mass flow of soil water into the plant may increase the overall concentration of soluble cations, or salt content, at the soil-root interface. This increased concentration will, in turn, affect the distribution of exchangeable cations and

342 could increase the availability of trace metal cations which occupy a low proportion of the exchange sites [68]. There has been considerable discussion over the past several years t h a t metals in the exchangeable phase and a few of the other so-called mobile geochemical phases are the most bioavailable. While this is proving to be correct, m a n y studies have been limited to chemical extractions and have not included plant uptake [47,60,69]. The exceptions have been studies on the use of soil test extractants, many of which measured exchangeable cations [70]. However, this type of extractant was used primarily for determination of the major cations (Ca § Mg § and K § and rarely trace metals. There are several recent investigations which have determined metals in the various soil geochemical phases and compared the values with metal concentrations in the plant. As previously discussed, there are many different selective extraction schemes which make data comparison often difficult. However, the one extraction t h a t is almost universal to all schemes is the one for the exchangeable (adsorbed) cations. As noted in Section 7.1, the proportion of several trace metals in the exchangeable phase may be very small which may explain why plant uptake is often related to phases other t h a n exchangeable. Some researchers have attempted to relate metal uptake to soil solution concentration because they believe that the greatest proportion of metals are absorbed from the soil solution. In an experiment to test this hypothesis, swiss chard (Beta Vulgaris) was grown in soil which had been treated with sewage sludge and various additions of P and N fertilizer [71]. The Cd and Zn levels in the plants were not related to the concentrations in soil saturation extracts (soil solution), P or N application rates. Although metals in the soil phases were not measured, the researchers postulated that the availability and plant uptake of Cd and Zn were reflective of metal desorption from the solid phases. It is reasonable to assume that a high proportion of these metals were desorbed from the solid phase directly into the plant root system. The rhizosphere consists of the soil zone within 1-2 mm of the root surface. There is a much higher soil acidity, bacterial activity and organic content in the rhizosphere than in the bulk soil and there is little question of the significant effect of the rhizosphere on the solid phases in the soil [4]. Low-molecular-weight organic acids (LMWOA) secreted by plant roots were found to modify the mobility of Cd through formation of soluble complexes in the rhizosphere of uncontaminated soils [72]. In a study where organic acids identified in plant root exudates, viz., 0.02 M acetic, citric, fumaric, oxalic or succinic acid were equilibrated with soil samples, the released Cd increased over that from the control soil. The results showed that: (1) Cd was brought into soil solution from the soils as Cd-LMWOA complexes by the LMWOAs secreted by the plant roots; (2) the kinetics of Cd release by LMWOAs was diffusion controlled, and (3) the dynamic release of LMWOAs from the plant roots into the soil rhizosphere would continuously release Cd from the soils, as indicated by the renewal of the LMWOAs. The average diffusion coefficients of Cd release from the soils by

343 LMWOAs and the Cd release by renewal of the LMWOAs followed the same trend as the Cd availability index of the soils. In an earlier study, these workers suggested that NH4Cl-extractable Cd should be used as a Cd bioavailability index based on its highly significant correlation with Cd concentration in the grain of durum wheat [73]. The NHnC1 extraction would measure the exchangeable Cd which is the most likely phase affected by the LMWOAs. For contaminated soils in Sudbury, Ontario, a strong linear relationship was found for Cu and Ni concentrations in birch twigs (dwarf birch, Betula pumila L. var. glandulifera Regel and white birch, Betula paprifera Marsh.) and soil exchangeable metal [61]. There was also a highly significant relationship between Ni concentrations in the twigs and total soil Ni. Levels of Zn in the plants and soil were within normal ranges and there was no significant association between soil and plant concentrations. The Cd concentration in sudax [Sorghum bicolor (L.) Moench] correlated strongly (r 2 = 0.91) with exchangeable Cd where sewage sludges containing high Cd levels were land applied [74]. It was also found that a good correlation existed between Cd plant levels and a Ca(NOa)2/EDTA extractant which was postulated to remove both exchangeable and chelation-bound Cd in the sludge solids. The best correlations between plant content and soil extractions were with soils where the concentration of metal in the applied sludge was relatively high. Associations between Cd in the sudax and soil phases for low-Cd sludges were much poorer and the overall correlation coefficient (r 2) for exchangeable Cd and the sudax concentration was only 0.56. This lack of a strong relationship between plant metal content and soil phases at low soil concentrations of the metal seems to follow a proposed "Threshold Level" theory [75]. The uptake of Cu, Ni and Zn by wheat (Triricum aestivum L.) and soybean (Glycine max L.) grown on soil treated with co-composted sewage sludge under greenhouse conditions was investigated [76]. The selective extraction scheme used by these researchers was somewhat different t h a n the procedure given above. After extracting the exchangeable and organic fractions, EDTA was utilized to remove metals from inorganic precipitates. The results determined that exchangeable Zn correlated with plant uptake but t h a t there was no such correlation with either Cu or Ni. The best uptake models were obtained from a stepwise multiple regression procedure where pH, exchangeable and inorganic precipitate-bound metals were the most important parameters. The authors pointed out t h a t the small increases in tissue concentrations of metals as a function of sludge application rate may have resulted in the poor relationship between metals in plant tissue and in soil phases. However, a relationship could exist even without a notable increase in metal content as a function of t r e a t m e n t rate. Other cited studies [39,61] have noted correlations between soil phase concentrations and plant uptake where there were no t r e a t m e n t effects. Another explanation for the lack of correlation between metals in plant tissue and soil phases is t h a t the overall soil and plant concentrations were below a level where this relationship is clearly expressed.

344 Romaine lettuce (Latuca sativa L.) was grown on soils t h a t had received varying rates and frequencies of biosolids applications over a 20-year period [62]. Applied biosolids where Cd concentrations were high and the Cd was in forms that were easily extracted from soil were readily available for uptake by the lettuce more t h a n 15 years following application. Concentrations of Cd, Cu, Ni and Zn in the lettuce leaves were positively correlated to the total concentrations of respective metals in the soil by either a linear or quadratic regression model. Using plant uptake slopes from the regression analysis equations, the authors suggest that the relative bioavailability of the biosolids-applied metals followed the trend: Cd>>Zn>Ni_>Cu>>Pb. For Cd, the best correlation (r 2 = 0.99) was obtained by including in a regression model the exchangeable, carbonate and Feoxide fractions. The correlation for Cd uptake and the exchangeable fraction alone was r 2 = 0.73. Plant Cu, Ni and Zn concentrations were correlated primarily with the exchangeable soil fraction. Plant Pb levels were generally not correlated to any of the soil geochemical fractions.

0

U S E O F S E L E C T I V E E X T R A C T I O N S TO E S T I M A T E M E T A L L E A C H A B I L I T Y F R O M THE SOIL TO THE G R O U N D WATER.

Few studies have directly investigated the relationship between metal concentrations in geochemical phases and amounts that can potentially leach into the ground water. The exception is the incorporation of exchangeable phase metal concentrations in some transport models. Some of the transport models for u n s a t u r a t e d soils utilize the cation exchange capacity, Freundlich equation exponent, distribution coefficient, Kd, and metal solubility as inputs [77]. The data are often generated from adsorption isotherms or leaching column studies, which are mono-element systems as discussed above. The importance of ion sorption reactions has long been recognized and ion exchange data is included in several models. Many researchers consider that sorption reactions are best described by a cation exchange model. The most rapid reactions are attributed to non-specific ion exchange while slower retention reactions are related to specific sorption of metal ions onto soil solid surfaces [78]. These reactions may also be interpreted in terms of formation of outer- and inner-sphere complexes with soil surfaces [2,3]. Two researchers [79,80] incorporated ion exchange reactions into the tworegion (mobile-immobile) concept. Their approach was generally successful in predicting the overall shape of breakthrough curves for Ca ++ and Mg §247which were obtained from miscible displacement columns that held different sized soil aggregates. Other transport models consider several mechanisms including ion exchange, complexation, dissolution-precipitation and competitive adsorption. Examples of such models include FIESTA [81], CHEMTRAN [82], and TRANQL [83]. Because of their complexity, several of these models have not been fully validated [78]. Results for Cd and Ni breakthrough on a sandy soil utilizing the FIESTA model have been described [84]. The model predictions provided higher

345 retardation of Cd and lower retardation of Ni t h a n empirically observed. Improved predictions were obtained when a kinetic approach was used with data from batch experiments. Model development and validation for movement of trace metal cations in soils has yet a significant period of evolution ahead. A thermodynamic model was developed for Cd, Cu and Zn concentrations/solubilities in soil solutions of sandy textured soils that had received applications of cattle-manure slurry [85]. With the assumption that organic m a t t e r was the dominant exchanger phase for the metals, the model accounted for metal complexation with dissolved organic carbon ligands and with the solid exchanger phase. The authors cautioned that their model is site specific and has yet to be verified however it provides a basis for further research to test assumptions and improve the model. A study of Cd, Pb and Zn in soils that had received surface applications of a metal-rich flue dust revealed that high proportions of Zn were present in the exchangeable phase of one of the soils to a depth of 105 cm [86]. The flue dust was applied annually for six years to raise soil pH and the metal measurements were conducted four years after the last application. Approximately 20% of the total Zn was found in the exchangeable phase in the surface samples, but the exchangeable phase was the dominant fraction at depths greater than 30 cm. Earlier research had reported that Zn in the exchangeable phase in subsurface soils was not the primary fraction in most southeastern soils [43]. The high proportion of exchangeable Zn in the subsurface was caused by the downward migration of the Zn in the soil profile. There were notable amounts of Zn in the organic, amorphous Fe oxide, crystalline Fe oxide and residual phases in the 0-30 cm portion of the profile but only minimal amounts in these phases at greater depths. These results note the stability of metals in the other phases as compared to exchangeable Zn. There also was a slight downward movement of Cd noted in the study, but much less than for Zn. Only about 11% of the Cd was present in the exchangeable form in the surface soil and much smaller proportions at greater depths. 10. S U M M A R Y AND C O N C L U S I O N S The importance of ion exchange reactions to nutrient dynamics in soils was first recognized in the mid-nineteenth century. The ability of soil colloids to reversibly adsorb cations from the soil solution was understood as an ion exchange reaction long before the origin of the negative charge on clay minerals and organic m a t t e r was known. Both non-specific adsorption, rapid ion exchange, and specific adsorption, chemisorption, are important reactions that influence trace metal cation behavior in the soil system. Exchangeable metals are usually the most reactive solid phase metals - most available for plant uptake and most easily leached downward through the soil profile. Other geochemical soil phases are important influencing factors of trace metal behavior in soils. Trace metal cations interact differently with the various soil phases based on

346 both solution and solid phase characteristics. The proportions of metals associated with the individual phases can often provide information pertinent to the contaminant status of the soil. Often, the absolute and relative concentration in each phase, as compared with the total metal concentration, can determine the presence and extent of anthropogenic input as well as the degree of environmental concern over the metals present. With the exception of Cd, the proportion of the trace metal cations in the exchangeable phase is small compared to the total concentration. When larger proportions of exchangeable metals are found, it is a prompt to conduct a detailed investigation of the trace metal geochemistry of the site under study. Exchangeable trace metal cations have been correlated to plant uptake and, in a few instances, related to transport of metals in the soil water. Since all of the geochemical phases are connected through the soil solution in a dynamic state of equilibrium, evaluation of all components is important to fully characterize trace metal cation behavior in soils.

REFERENCES

1. A. Kabata-Pendias and H. Pendias, Trace Elements in Soils and Plants, Second Ed, CRC Press, Boca Raton, FL, 1992. 2. G. Sposito, The Surface Chemistry of Soils. Oxford University Press, New York, NY, 1984, 234. 3. K.B. Krauskopf and D.K. Bird, Introduction to Geochemistry, Third Ed., McGraw-Hill, Inc., New York, 1995. 4. N.C. Brady and R.R Weil, The Nature and Properties of Soils, Eleventh Ed., Prentice Hall, Upper Saddle River, NJ, 1996. 5. S. Lewis and J.H. Rule, Unpublished research. 6. M.B. McBride, Environmental Chemistry of Soils, Oxford University Press, New York, 1994. 7. H. Farrah and W.F. Pickering, Water, Air and Soil Pollut., 8 (1977) 189. 8. R.W. Puls and H.L. Bohn, Soil Sci. Soc. Am. J., 52 (1988) 1289. 9. D.G. Kinniburgh, M.L. Jackson and J.K. Syers, Soil Sci. Soc. Amer. J., 40 (1976) 796. 10. U. Schwertmann and R.M. Taylor, Iron Oxides, in: J.B. Dixon and S.B. Weed (eds.), Minerals in the Soil Environment, Soil Sci. Soc. of Am., Madison, WI, 1977, 145. 11. E.A. Forbes, A.M. Posner and J.P. Quirk, J. Soil Sci., 27 (1976) 154. 12. R.R. Gadde and H.A. Laitinen, Environ. Lett., 5 (1973) 91. 13. J.W. Murray, Geochem. Cosmochem. Acta, 39 (1975) 505. 14. M. Schnitzer and S.I.M. Skinner, Soil Sci., 102 (1966) 361. 15. M. Schnitzer and S.I.M. Skinner, Soil Sci., 103 (1967) 247. 16. S.U. Khan, Soil Sci. Soc. Am. Proc., 33 (1969) 851. 17.F.S. Stevenson, Soil Sci. Soc. Am. J., 40 (1975) 197.

347 18. H.A. Elliot, M.R. Liberat and C.P. Huang, J. Environ. Qual., 15 (1986) 214. 19.M. Schnitzer, Soil Sci. Soc. Am. Proc., 33 (1969) 75. 20. M. Schnitzer and H. Kemdorff, Water, Air Soil Pollut., 15 (1981) 97. 21. K.H. Tan, 1993. Principles of Soil Chemistry. Second Ed., Marcel Dekker, New York, NY, 362. 22. W.J. Bond and V. Verburg, Soil Sci. Soc. Am. J., 61 (1997) 444. 23. R.G. McLaren and D.V. Crawford, J. Soil Sci., 24 (1973) 443. 24. C. Kheboian and C.F. Bauer, Anal. Chem., 59 (1987) 1417. 25. G.W. Br~immer, J. Gerth and K.G. Tiller, Soil Sci., 39 (1988) 37. 26. R.M. Engler, J.M. Brannon, J. Rose and G. Bigham, A practical selective extraction procedure for sediment characterization, in: T.F. Yen (ed.), Chemistry of Marine Sediments, Ann Arbor Science, Ann Arbor, MI, 1977, 163. 27. S. Gatehouse, D.W. Russell and J.C. van Mort, J. Geochem. Expl., 8 (1977) 483. 28. S.J. Hoffman and W.K. Fletcher, Selective sequential extraction of Cu, Zn, Fe, Mn and Mo from soils and sediments, in: J.R. Watterson and P.K Theobald, (eds.), Proceedings of the Seventh Internatioal Geochemical Exploration Symposium, Assoc. Expl. Geochem., 1978, 289. 29. A. Tessier, P.G.C. Campbell and M. Bisson, Anal. Chem., 51 (1979) 844. 30. M.J. Gibson and J.G. Farmer, Environ. Pollut., 11 (1986) 117. 31. L.M. Shuman, Soil Sci., 127 (1979) 10. 32.L.M. Shuman, Soil Sci. Soc. Am. J., 50 (1986) 1236. 33. W.P. Miller, D.C. Martens and L.W. Zelazny, Soil Sci. Soc. Am. J., 50 (1986) 598. 34. N. Belzile, P. Lecomte and Tessier, A. Environ. Sci. Technol., 23 (1989) 1015. 35. K. Wallman, M. Kersten, J. Gruber and U. F6rstner, Int. J. Environ. Anal. Chem., 51 (1993) 187. 36. C. Keller and J.C. V6dy, J. Environ. Qual., 23 (1994) 987. 37. W.W. Wenzel and W.E.H. Blum, Effect of sampling, sample preparation and extraction techniques on mobile metal fractions in soils, in: D.C. Adriano, Z.-S. Chen, S.-S. Yang and I.K. Iskandar (eds.), Biogeochemistry of Trace Metals, Advances in Environmental Science, Science Reviews, Norwood, 1997, 121. 38. L.M. Shuman, Chemical forms of micronutrients in soils, in: J.J. Morvedt, et al., (eds.), Micronutrients in Agriculture, 2nd Ed., Soil Sci. Soc. Am., Madison, WI, 1991, 113. 39. J.H. Rule, B.C. Comstock and C.I. Impellitteri, Proc. and Extended Abstracts of the Fourth International Conference on the Biogeochemistry of Trace Elements, Berkeley, CA, 1997, 549. 40. J.H. Rule and D. Adriano, Proc. of the Third International Conference on the Biogeochemistry of Trace Elements, Paris, France, 1995, in press.

348 41. W.W. Wenzel and G. Wieshammer, Extractability of mobile A1, Co, Cu, Fe, Mn, Ni, V and Zn from soils. Submitted to J. Environ. Qual., 1997. 42. K.G. Tiller, J.L. Honeyset and M.P.C. de Bries, Aust. J. Soil Res., 10(2) (1972)165. 43. L.M. Shuman, Soil Sci., 140 (1985) 11. 44. J. Kotuby-Amucher, R.P. Gambrell and M.C. Amacher, The distribution and environmental chemistry of lead in soil at an abandoned battery reclamation site, in: I.K. Iskandar and H.M. Selim (eds.), Engineering Aspects of MetalWaste Management, Advances in Trace Substance Research, Lewis Publishers, Boca Raton, FL. 1992, 1. 45.J. Liang, J.W.B. Stewart and R.E. Karamanos, Can. J. Soil Sci., 70 (1990) 335. 46. Y.K. Soon and T.E. Bates, J. Soil Sci., 33 (1982) 477. 47. G. Sposito, L.J. Lurid and A.C. Chang, Soil Sci. Soc. Am. J., 46 (1982) 260. 48. G. Rauret, R. Rubio, J.F. L6pez-Sfinchez and E. Casassas, Int. J. Environ. Anal. Chem., 35 (1989) 89. 49. J.L. Sims and W.H. Patrick Jr, Soil Sci. Soc. Am. J., 42 (1978) 258. 50. T.T. Chao, Soil Sci. Soc. Amer. Proc., 36 (1972) 764. 51. W.P. Miller, D.C. Martens and L.W. Zelazny, Soil Sci. Soc. Am. J., 49(1985) 856. 52.A. Kabata-Pendias, Applied Geochem., Suppl. Issue No. 2 (1993) 3. 53. J.H. Rule, 1997. Unpublished manuscript. 54. I.A. Salim, C.J. Miller and J.L. Howard, Soil Sci. Soc. Am. J., 60 (1996) 107. 55. M.B. McBride, Soil Sci. Soc. Am. J., 44 (1980) 26. 56. A. Chlopecka, J.R. Bacon, M.J. Wilson and J. Kay, J. Environ. Qual., 25 (1996) 69. 57.A.S. Jeng and B.R. Singh, Soil Sci., 156 (1993) 240. 58. S. Dudka and A. Chlopecka, Water, Air Soil Pollut., 51 (1990) 153. 59. M.G. Hickey and J.A. Kittrick, J. Environ. Qual., 13 (1984) 372. 60. L. Ma and G.N. Rao, J. Environ. Qual., 26 (1997) 259. 61. S. Dudka, R. Ponce-Hernandez, G. Tate and T.C. Hutchinson, Water, Air Soil Pollut., 90 (1996) 531. 62. J.J. Sloan, R.H. Dowdy, M.S. Dolan and D.R. Linden, J. Environ. Qual., 26 (1997) 966. 63. A. Kabata-Pendias, Agricultural problems related to excessive trace metal contents of soils, in: W. Salomons, U. F6rstner and P. Mader (eds.), Heavy Metals: Problems and Solutions, Springer-Verlag, Berlin, Germany, 1995, 3. 64.M. Boyle and W.H. Fuller, J. Environ. Qual., 16 (1987) 357. 65. F.M. Dunnivant, P.M. Jardine, D.L. Taylor and J.F. McCarthy, Environ. Sci. Technol., 26 (1992) 360. 66. C. Amrhein, J.E. Strong and P.A. Mosher, Environ. Sci. Technol., 26 (1992) 703. 67. I. Lamy, S. Bourgeois and A. Bermond, J. Environ. Qual., 22 (1993) 731.

349 68. S.A. Barber, Soil Nutrient Bioavailability: A Mechanistic Approach, Second Ed., John Wiley & Sons, Inc., New York, 1995. 69.M.P. Levesque and S.P. Mathur, Soil Sci., 142 (1986) 153. 70. B. Zhu and A.K. Alva, Soil Sci., 156 (1993) 251. 71. J.R. Villarroel, A.C. Chang and C. Amrhein, Soil Sci., 155 (1993) 197. 72. G.S.R. Krishnamurti, G. Cieslinski, P.M. Huang and K.C.J. Van Rees, J. Environ. Qual., 26 (1997) 271. 73. G.S.R. Krishnamurti, P.M. Huang, K.C.J. Van Rees, L.M. Kozak and H.P.W. Rostad, Commun. Soil Sci. Plant Anal., 26 (1995) 2857. 74.J. Jing and T.J. Logan, J. Environ. Qual., 21 (1992) 73. 75. J.H. Rule, Use of Small Plants as Phytomonitors with Emphasis on the Common Dandelion, Taraxacum Officinale, in: D.C. Adriano, Z. Chen and S. Yang (eds.), Biogeochemistry of Trace Elements, Environ. Geochem. and Health, Special Issue, 16, 1994, 627. 76. J.T. Sims and J.S. Kline, J. Environ. Qual., 20 (1991) 387. 77. R.N. Yong, A.M.O. Mohamed and B.P. Warkentin, Principles of Contaminant Transport in Soils, Elsevier Science Pub. B.V., Amsterdam, The Netherlands, 1992. 78. H.M. Selim and M.C. Amacher, Reactivity and Transport of Heavy Metals in Soils, Lewis Publishers, Boca Raton, FL, 1997. 79. H.M. Selim, R. Schulin and H. Fliihler, Soil Sci. Soc. Am. J., 51 (1987) 876. 80. R.S. Mansell, S.A. Bloom, H.M. Selim and R.D. Rhue, Soil Sci. Soc. Am. J., 52 (1988) 1533. 81. A.A. Jennings, D.J. Kirkner and T.L. Theis, Water Resour. Res., 18 (1982) 1089. 82. C.W. Miller and L.V. Benson, Water Resour. Res., 19 (1983) 381. 83. G.A. Cederberg, R.L. Street and O.J. Leckie, Water Resour. Res., 21 (1985) 1095. 84. D.J. Kirkner, A.A. Jennings and T.L. Theis, J. Hydrol., 76 (1985) 107. 85. D. Hesterberg, J. Bril and P. del Castilho, J. Environ. Qual., 22 (1993) 681. 86. Z. Li and L.M. Shuman, Soil Sci., 161 (1996) 656.

Adsorption and its Applications in Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski(Editor) 9 1998Elsevier Science B.V. All rights reserved.

351

Application of environmental colloid science in the soil systems J. Szczypa a, I. Kobal b, and W. Janusz a aDepartment of Radiochemistry and Colloid Chemistry, UMCS, 20-031 Lublin, Poland b Josef Stefan Institute, Jamowa 39, 1000 Ljubljana, Slovenia and School of Environmental Sciences, Vipavska 13, 5000 Nova Gorica 1.

INTRODUCTION

Soil is, beside air and water, the most important environment for the living organisms including h u m a n beings. Between these systems, some components exchange, influencing the development and condition of organisms. Up till now most of the h u m a n food comes directly or indirectly from soil cultivation. That is the reason for the special care of soil system quality. The main way is to fully understand complex processes that rule the ecosystem. When air and water systems are simple, the soil is very complicated and variable one, concerning the mineral, grain-size distribution and composition of surrounding solution. Soil is multicomponent, polydispersed system, usually treated as a three-phase system (solid, liquid and gas), although some authors consider organisms the fourth phase. Tens thousands of soils can be distinguished. They can differ by their provenance, parent minerals and age. For example later soils, less weathered, are rich in silica, alumina and iron minerals whereas old weathered ones are without many soluble minerals of alumina and iron and consists of silts and rusty oxides [1]. Looking at composition of soil there can be distinguished primary and secondary minerals. Quartz, orthoclase, plagioclase, muscovite, biotite, pyroxenes and olivine are in the first group and hydrolyzed silicon oxides, aluminum and iron hydroxide, carbonates (calcite and dolomite), hydroxides and silt minerals are in the second one. Half of the most soils is formed by minerals, the rest consists of water solution, air and organic substances (about 5%). In sandy silts the organic substance contents is small and reaches few percent whereas in peat soils or muck the organic substance contents may reach 100% [1]. Among many organic substances in soil very important role play humines, humine acids, humatomelitic acids and fulvic acids. These compounds dissolve in water in a different degree. The solid particles of soil show various size from bigger t h a n l m m treated as gravel and stones, then sand 0.05-lmm, silt 0.002-0.05mm to clays < 0.002mm. The sandy soil, formed by fine particles is concise, plastic, sticky and impermeable. This

352 type of soil is defined as heavy soil. The soil with sand is porous and defined as light soil. In this soil the transportation processes of air and water with nutrient substances are easier. The presence of humus provides proper structure of soil, decides degree and mechanism of soil particle aggregation. Adsorbing on mineral particles, h u m u s provides spongy structure of soil that promotes a transportation process. Soil is a source of heavy metals and their adsorbent also. These factors, which have an influence on the total contents of metals and their amount assimilated by organisms is essential for h u m a n beings life and soil fertility. Some of heavy metals are known as microelements (Cu, Zn, Co, Mn) that are indispensable for growth and their life at specific concentrations and toxic at higher, and other metals that are potentially toxic elements (PTE) (As, Hg, Pb, T1 and U) [2]. Radioactive isotopes, existing in the environment may be divided according to their origin: primary, cosmic and anthropogenic ones. Primary radioisotopes, which came into being during nuclear synthesis of elements have a half-life time comparable or bigger than the age of our planet. To this group belong K-40 and isotopes of U-235, U-238 and Th-232 series. The isotopes of cosmic origin form in atmosphere of the Earth as a result of cosmic radiation for example H-3 and C-14. Man has inconsiderable influence on the formation and spreading of the isotopes from these groups. Unfortunately there is a group of isotopes that were introduced to environment by man. Some of them, by the occasion of nuclear energy production (from mining of radioactive ores to closing down the exploited reactors), other as a result of nuclear weapon tests. The main radioisotopes introduced to environment in this way include the uranium family nucleus' fission products (U, Pu) and products of activation of reactor materials. The list of radioactive isotopes, important for the environment contains elements of the most groups of the periodic table. The list of elements and their origin was presented by Lieser [3]. A physico-chemical behavior of the most radioactive isotopes in soil is the same as their stable equivalents, only for H-3 some differences may be distinguished. Radioisotope concentration may be similar or considerably smaller than stable isotope in environment, depending on its half-life time. The concentration of the hydrolyzable element is connected with a form it exists in the environment. When the concentration is smaller than the solubility product of the respective hydroxide or other insoluble salt and there are no conditions to adsorption on the dispersed particles of other phase, the isotope may form non-ideal solution. When the isotope concentration is so high that exceeds the solubility product then the dispersed phase may form. When the dispersed phase exists and physical conditions enable the adsorption on the solid, then isotope forms pseudocolloid, that means exists in the dispersed phase not forming crystalline structure. From the chemical and physical behavior at the solid/solution interface there is no difference between stable and a radioactive isotope. The only difference is their harmfulness to living organisms. To fully understand the transport phenomena of nutrient substances in such complicated system, the adsorption measurements of the selected element on the

353 soil are not sufficient method. It is necessary to learn surface properties of individual components of soil and on this basis to work out the model of incorporated processes. There are some papers that treat processes running in soil as a physical chemistry of dispersed systems that is ion adsorption and colloid transportation in water [2-7]. In present paper the essential processes, having influence on the ion distribution between solid and solution are presented. They may be related to soil minerals/ aqueous solution system and thus may be useful for the understanding of the transportation and adsorption phenomena in such complex system as the soil is. In the subsequent chapters of this paper the problems of the ion adsorption and surface charge formation are presented.

2.

METAL OXIDE/AQUEOUS SOLUTION SYSTEM

Among metal oxides found in soil, beside silica, there are oxides, hydroxides and hydrated oxides of iron, aluminum and mixes hydroxides. The electrolyte ions may accumulate on the metal oxide surface as a result of nonspecific adsorption, caused by electrostatic interaction, complexing, ion exchange. Another mechanism of accumulation of ions on the soil minerals is heterocoagulation of colloids which have been formed by ions. The adsorption process, connected with the influence of coulombic interaction, is defined as nonspecific adsorption. It is caused by the electric charge on the surface of the metal oxide, which caused distribution of the ions in the surrounding solution layer. In consequence, the ions of the same sign as the sign of surface charge of the particle of the solid will be removed from that region of the solution, whereas the ions of the opposite charge will be accumulated. The layer of the solution that is under the influence of coulombic forces of the solid surface charge is called the diffusion layer. The whole range of the solid with accumulated charge, together with the part of solution with compensating charge is defined as an electrical double layer (edl). The distribution of the charge in the diffusion layer is relatively well described by Gouy-Chapman theory of the edl, where the charge density near the flat plane of the solid is equal:

~d = - ~ / 8 ~ ~

sinh/F~d)2RT

(I)

where: ~d - diffuse layer charge density, ~- relative dielectric constant for water = 78.25, ~o-absolute dielectric constant = 8.85*10~2CemU-~, c - electrolyte concentration, R - g a s constant - 8.314 J*mol~*K-~, T - temperature, ~gd- diffuse layer potential, F - Faraday constant 96500 C'mole 1.

354 Electric charge on the metal oxide surface is formed because of acid-base reactions of the surface hydroxyl groups. In colloid chemistry there are two approaches to an electric charge formation on the surface of the oxide. The first one by the reactions of the ionization surface group, defined as 2-pK model [8]: -SOH~

~

-SOH+H

(2)

§

- S O H +_~ - S O - + H +

(3)

where - S represents some surface of metal oxide irrespective of metal atom n u m b e r that coordinates the oxygen atom of the hydroxyl group. The reaction 2 and 3 constants may be calculated from the charge density data as a function of pH or from ~ potential versus pH dependence. The alternate attitude to the formation of the charge on the metal oxide is proposed by MUltiSIte Complexation model (MUSIC model) where hydroxyl group on the oxide surface are gifted by the charge that depends on the degree of saturation of the oxide valence by coordinating metal atoms and hydrogen, connected by hydrogen bound by donor or acceptor bonding [9]. (--MekO(H)m(HH20)n~ init + H + ~

( - M e k O ( H ) m + l ( H H 2 0 ) n - l ~ fin

(4)

where: Sinit(fin) - charge of the surface group before or after adsorption of hydrogen, which depends on the number and charge of coordinated metal atoms (k and SMe) and number (m) of hydrogen atoms (donor type connection) and (n) number of hydrogen atoms (acceptor type connection), s = ~ SMr + m* s H + n* (1 - s~) + V, SH = 0.8, whereas V = -2. For the number of free orbitals of the oxygen there is restriction for the hydrogen bonds for k = 1 ~ n + m = 2, for k = 2 ~ n + m =2 or n + m = 1 whereas for triple coordinated oxygen atoms (k = 3), n + m = 1. That means that hydrogen atom may be bound to surface oxygen in the acceptor or donor way [9]. Reaction constants of surface groups in MUSIC model are calculated theoretically from crystallographic data. First, from Brown theory, the valence of metal in the lattice of metal oxide is calculated [10], next, the charge of the surface group and finally, on this basis, the constants of the surface groups. The increase of the electrolyte concentration in the metal oxide/electrolyte system causes the increase of the charge density at the interface due to the following reactions: - SOH~An-SOH+Ct

~

+ ~

- SOH + H § + An-

(5)

-SO-Ct ++H +

(6)

According to the site binding theory, anions, which reacts with the hydroxyl group, produce surface complex type compound. The positive charge of this group is

355 located in the surface layer. that occupies position in the concentration of -= SOH ~Ansolution and increase with Similarly, for reaction 6, the

It is compensated by the negative charge of the anion inner Helmholtz plane (IHP). Following reaction 5, the groups should decrease with the increase of pH of the the increase of the concentration of the electrolyte. adsorption of the cation results in the formation of the

surface compounds - S O - C t +- type complex, where negatively charged part is in surface plane of edl, whereas the cation is in the IHP. Because the ions, adsorbing according to the reactions 5 and 6, form the complex type connections not only by electrostatic but also chemical forces, this type of adsorption is called specific adsorption of the ions [11]. The Cs-137 and Cs-134 isotopes adsorb in the same way as the stable caesium. However, because of their low concentrations (for example 1Bq/dm 3 Cs-137 equals to 1"10 -15 mole/dm 3) towards 1:1 salts present in soil water, the processes of specific and nonspecific adsorption will have minor importance. The excess of monovalent cation Na § or K § will lower the adsorption of Cs § because of the competitive adsorption on the same site. The concentration of potassium ions, in aqueous solution of the soil, reaches 2mg/dm 3 (about 50 ~mole/dm~). These ions will adsorb on hydroxyl groups according to reaction 6, whereas the adsorption of Cs § ions from such dilute solution will be limited [12]. For example complexation constant of Na § for TiO2, pKNa=8.2 and Cs § pKc~=7.2. For the concentrations Na+=lmmole/dm ~ and Cs+=lpmole/dm 3 [lkBq of Cs-137/dm 3] the [-TiO-Na ~] to [-TiO-Cs § relation, calculated with neglecting of the radius difference and activity coefficient was 10-s, based on the following equation:

I TiO-Cs TiO-Na-

Kcs * [Cs+]= 10 -7,2 ,10-12 KNa [Na + ]

= 10 -8

(7)

10 .8,2 , 1 0 - 3

That confirms the above opinion of the negligibly low specific adsorption of the Cs-137 or Cs-134 from the soil solutions containing other also alkaline metal cations Na § or K +. Only contamination by Rb-87, whose concentration (1Bq/dm 3 =1 mole/dm~), may cause appreciable specific adsorption. After all, caesium isotopes may adsorb on metal oxides by the exchange reaction with respective ions, presented in the oxide for example as contamination. Another process, responsible for the deposition of the caesium on the solid surface may be heterocoagulaton of the pseudocolloidal form of Cs [13]. This mechanism will be discussed later. The exchangeable adsorption of ions on the metal oxides occurs in the presence of the ion type contaminations. On the surface of the oxide, beside the adsorption of the cation according to reaction 6, the substitution of the contamination for to the Cs takes place. = C t s + C s + ~-

-Cs s+Ct +

(8)

356 As far as the concentration of the cation on the surface of the oxide does not change with pH, the adsorption, according to reaction 8, is independent on the pH of the solution. On the other hand, because in the exchange reaction the H § ions may take part, then a small pH dependence of the Cs sorption can be observed. The investigation of caesium sorption on the titanium, aluminum and silicon oxides, performed by Hakem et al., revealed that the increase of the concentration of the electrolyte lowers the adsorption of the Cs-137 or 1-131 [14]. The pH dependence of the sorption of these radionuclides is typical for the adsorption of the ions on oxides. However, the higher adsorption of the cation at pH>pHpzc and anion at pH>pHpzc was observed. This behavior suggests that ion exchange process has a vivid share in the ion adsorption on the surface of the oxides. The authors of discussed papers characterize applied oxides by mentioning the size of the particles and specific surface, without telling about the existence of the ion contamination. Ionic impurities of metal oxide may have influence on the mechanism of the ion adsorption, especially from very diluted solutions-10-Smole/dm 3. The investigations made by Kosmulski et al. showed that porous glass, containing on its surface borsodium phase, is good adsorbent for Cs-137115-17]. The adsorption of this isotope is promoted by alkaline pH and low ionic strength. Some adsorption of the Cs-137 was observed on the silica gel [15]. Appropriately prepared four component glasses, Vycor-type, showed the good adsorption of Cs-137 [18-19], also not only in the alkaline pH, as it happened for three component glasses. Although the examination of the adsorption on the porous glasses focused on the obtaining the adsorbent for the removing the Cs radioisotopes from water, the achieved results showed that the presence of ions or the ion exchangeable layer on the surface of the oxide increases the adsorption of monovalent ions from the solution. Ion exchange character of the Cs adsorption on the soil sample that mainly consists of the sand was observed by Shenber and Johanson [20]. The adsorption of multivalent ions or monovalent hydrolyzable ions is specific adsorption. Because of the valence of the ion, more than one adsorption site may be occupied. The adsorption of hydrated form may go through dissociation of the hydrogen cation from the hydroxyl group of the adsorbed complexes as well as from the surface hydroxyl group. Because the adsorption of the metal cations on the surface hydroxyl group goes with dissociation of H § then the adsorption of cations in the some range abruptly increases with pH. This effect is called the edge of adsorption. The parameters that characterize the edge of adsorption [22], are explained in Figure 1. The specific adsorption may lead to the formation of inner or outersphere complexes [23,24]. As an innersphere complex is treated surface compound where the cation is directly connected with oxygen from the surface of the metal oxide (Figure 2a). The outersphere complex is formed when the adsorbed cation maintains the hydrated layer of water, Figure 2b. From the pH dependence on the adsorption, one cannot conclude, whether the inner or outersphere complex is formed.

357 ApHlo.9o%

20.0 --

6.4

9 -- ,.L

_--

o~ 15.0 - -

-- 6.0

E "O

_o O

E E

E N

10.0

0 e.0 0

<

e--

III tn(

_

"10

~+

!r

--

--

5.6

--

5.2

No-

a ) = d___Me

5.0-I

, 0.0

~

~-i

PHso% I 4

I

I 5

'

I

'

6

I

'

7

I

'

I

8

9

,

i

4.8

10

pH

Figure 1. Adsorption of Zn(II) (cricles) and concentration of Zn ions (triangles) in TiO2(Rutile)/electrolyte system as a function of pH (data from [21 ]). pHs0o/o- pH of the solution with 50% cation adsorption, April0-90 - the pH range where adsorption changes from 10 to 90%, dpMe/dpH - the parameter that shows how the activity of the cation must be changed with the changes of pH to leave the cation adsorption on the same level.

a)

b) (~

(Do 9 o ~I

0

o @

o~

o ~

o @

o

@

o @ ~t

o 9

o@

o ~

o9

0 ' e

000 0o~

k, ;

Hq H

o o

~I

o@

( D o @

o@

o@

sp ip

sp

IHP

Figure 2. Inner- (a) and outer (b) sphere complexes at the metal oxide surface, sp - surface plane, ip - innersphere complexes plane, IHP - Inner Helmholtz Plane.

358 The structure of the complex may be estimated by the spectroscopic measurements, as was proposed by Robertson and Leckie [22]. Especially, very useful are the spectroscopy methods that allow in situ investigations, i.e. Fourier Transformation Infrared Spectroscopy (FTIR) electron or proton resonance spectroscopy, M6sbauer spectroscopy or X-ray adsorption spectroscopy (XAS) [24]. These methods allow also to decide other mechanisms such as polynuclear surface complex formation or surface precipitation. The polynuclear complexes or clusters of the new phase may form at higher concentrations of the adsorbing ion. According to James and Healy, the surface precipitation of the hydroxide occurs at lower concentration of the cation than in bulk. That is because of the lower value of the dielectric constant in the compact region of the edl, than in the solution [25]. The coverage of greater and greater surface of metal oxide by the new phase, with the increase of pH (increase of "adsorption"), changes the surface properties of the oxide. At high enough coverage degree, there may occur the change of the surface charge from the negative to positive one in CR2 point (charge reversal). The following increase of the pH, results in the succeeding reverse of the charge sign CR3, at pH characteristic for pzc of the adsorbing metal cation hydroxide. Similar effect, of the charge reversal during specific adsorption of hydrolyzable cation, was described by Schindler [26]. The specific adsorption of an anion on the surface of oxides goes through the exchange one or two hydroxyl groups for the anion: k(- SOH)+ L n- +-~ (- S)kL n-k + k O H -

(9)

The greatest adsorption of the anion is at low pH and decreases from certain value of pH with the increase of pH of the solution. The weak acid anions may adsorb on the metal oxide surface from the solutions of the low pH as a molecule of appropriate acid with liberation of the water molecule: k(_ SOH)+ LHn-m m ~

n-m-k + k H 2 0 (- S)k LH m-k

(~o)

At the increase of the pH the respective ion forms of acid may be adsorbed. This type of adsorption occurs for As, Cr(VI), Mo and V ions. The cation complexation constants, on the surface of the oxide, may be calculated from the pH dependence, with the method proposed by Schindler [26]. Recently, the complexation constants are found with application of the numerical optimization, based on the chosen model of the edl (DLM, TLM). They allow to fit the model parameters to the experimental data (FITEQL [27], HYDRAQL [28], SURFEQL [29], GRIFT [30], SURCOM [31]). Fitting of the edl models with numerical optimization procedures allows to regard the energetic heterogeneity of the adsorption sites on the surface of the oxide [32,33]. A survey of data, concerning specific adsorption of cations on the metal oxides, was presented by Schindler [26], Schindler and Stumm [34], Kinniburgh [35] and Huang [36]. Some values for the complexation reaction

359 constants of the important for the environment heavy metals are presented in Table 1 and 2. Table 1 Negative logarithm of the apparent stability constants of surface complexes of heavy metal ions with one surface site Pb(II)

SiO2 5.1 [36]

Cd(II)

6.1

Hg(II) Co(II) Cu(II) H3AsO4 H2AsO4AsO43H3AsO3

[36]

5.52 [34]

FeOOH 3.8 [36] 4.65 [37] 4.9 [36] 0.47 [37] 7.76 [37] -0.46 [37] 2.89 [36] 29.31 [37] 23.51 [37] 10.58 [37] 5.41 [37]

A12Os 2.2 [35]

2.1 [33]

TiO2 0.2 [36] 0.44 [26] 3.2 [36] 3.32 [26] 4.3 [34] 1.43 [34]

Table 2 Negative logarithm of the apparent stability constants of surface complexes of heavy metal ions with two surface sites fl2'ii,,) Cu Cd Pb Co

SiO2 11.19 [34] 14.2 [34] 10.68 [34]

FeOOH 1.7 [38] 4 [38] 1.6 [38]

A12Os 7.0 [34] 8.1 [34]

TiO2 5.04 [34] 9.00 [34] 1.95 [34] 10.6 [34]

Other cation presence, for example alkaline metal cations, may change the adsorption of multivalent cations, by changing their activity in aqueous solution and competitive adsorption on the same surface sites. Because the adsorption of monovalent and multivalent ions on the surface of the oxides goes through the hydrogen ion exchange in hydroxyl groups, the competition of the adsorption may occur. For the adsorption that gives innersphere complexes, such competition may not occur. That is despite the reactions of hydrolyzable and background electrolyte cations, as the process is ruled by different mechanism. Moreover, the adsorption with innersphere complex formation is characterized by far greater adsorption IS OS constant. The relation I~Me(II) > > K.Mr > > I~Na can be observed [24]. Cations, which adsorb specifically with the formation of innersphere complex, reveal a shift of the adsorption edge (adsorption-pH dependence) for different concentrations of the

360 electrolyte solution. For example adsorption of Ba 2§ on TiO2 (anatase) [21], is depicted in Figure 3. In the case of innersphere complex, the adsorption edges for respective ion strength covers. This plot is demonstrated by the the adsorption of Cd on TiO2 (rutile) Figure 4. The adsorption of heavy metals on the metal oxides is the localized type adsorption and its isotherm (as a function of the concentration in the solution) is described by modified Langmuir equation. This modification includes interaction with surface potential: F,~/ FmaxKMe [H+ ~1 [MeZ+ ]* exp - RT FMe

(11)

=

1 + KMe [H+ ~1 [MeZ+ ]* exp(- ~-~/ For the settled pH value, the surface potential ~ maintains constant, so the components of the equation may be included, with ZMe, into adsorption constant Kad s = KMe

$[H+] -I *exp -

. Then the equation (11)is the same as Langmuir's

one. 0.05 - -

0.04 - -

r

E

"13 0 m o

E

0.03 - -

~l m

0

E o Q.

0.02

L

o (/) "13 <

0.01

--

0.00

' 3

I 4

'

I 5

'

I 6

'

I 7

'

I 8

'

I 9

'

I 10

'

I 11

pH Figure 3. Adsorption of Ba2§ ions at the TiO2(anatase)/solution of NaC1 as a function of pH, rectangles 0.001 mole/dm 3; triangles 0.01 mole/dm 3, circles 0.1 mole/dm 3 of NaC1, (data from [21 ]).

361

20.0

~E

15.0

+t% 10.0

O

o

<~

5.0

0.0

i

3

4

i

5

I

6

7

i

8

i

9

I

10

I

11

pH

Figure 4. Adsorption of C d 2+ ions at the T'O2~Rutile)/solution] ~ of NaC1 as a function of pH, rectangles 0.001 mole/dm3; triangles 0.01 mole/dm 3, circles 0.1 mole/dm 3 ( data from [21]).

Surface properties of soil are not uniform because of its composition, presence of different components and properties of crystal faces of each component. About this, for everyone component of the solid phase, one should use a separate set of the constants of the isotherm equation. Sposito showed, that it is possible to obtain the approximate Bemmelen-Freundlich isotherm, by using logarithmic-normal distribution of adsorption constants of the Langmuir isotherm [5]. In the case of very low concentrations of the cation, the adsorption may be described by Henry's equation. It may be applied thus for the adsorption of radionuclides having a short half-life time. At the relatively high concentrations, beside adsorption, the precipitation processes may take places. In the presence of the solid, for the cation concentrations higher or equal to the solubility product of the metal hydroxide, the adsorption isotherm does not break, that reflects the precipitation of the hydroxide of other phase but the isotherm changes the shape, which suggests the increase of the adsorption and is characteristic for the multicore complex formation or surface precipitation [24]. In the aqueous environment, there are some ligands beside adsorbing cations, which may form complexes with these cations. Then, in hydration sphere one or more water molecules may be substituted by a ligand. Depending on the complex type, the adsorbing complexing cation may be bind directly with the surface (A type complexes)

362

n - S O H + MeL~ + ~

(- SO- )nMe(L)I z-n)+ + nil+

(12)

or they may adsorb through a ligand (B type complexes) n - S O H + L M e z+ + nH20 +_~ (- SOH~)n LMe(n-z)+ + n O H -

(13)

In A-type complexes (reaction 12), as far as the formed complex does not adsorb or adsorb more weakly than the metal cation, the presence of L-ligand results in lowering of the metal adsorption. The possibility of the complex adsorption through a ligand, (for B type complexes) ease cation adsorption at lower pH values because the adsorption runs according to the mechanism characteristic for anions. Such influence of the ligand presence on the cation adsorption was observed for Ag§ 2 system. In the natural water environment, the following anions exist as complex anions: CI-, SO42-, HPO42-, F-, OH- and COa 2. Monovalent anions may influence the cation adsorption, according to reaction 13, whereas multivalent ones may increase or decrease the adsorption as well. The important role, as a complexing anion, is played by carbonate ions. Actinides and another elements which form with these ions the relatively stable complexes. Unfortunately, in the bibliography there is no data concerning cation adsorption in the presence of carbonates. Among many anions, potentially complexing heavy metal cations, present in a gmole amount in the soil many organic acids [39]. Their role was discussed in details by Hartera and Naidu [39]. 3.

CATION A D S O R P T I O N ON CLAY M I N E R A L S

Beside oxides, clay minerals, products of weathering of rocks, are important component of soil. They reveal lamellar structure, consisting of tetraheders sheet of XO4 ,,t" (X = Si, A1, Fe) and octaheders sheet of XO6 ,,o" (X = A1, Fe, Mg). Moreover, some places may be occupied by other cations of small size. Depending on the number of layers, the phyllosilicates may be divided into two, three and four layer types. Among these types in the soil are present kaolinite as twolayers ,,t-o" (Figure 5) and montmorylonite (vermiculite), smectite or mica, illite as threelayers ,,t-o-t" (Figure 6 )[39]. Generally, ideal phyllosilicates do not exist. The Si(IV) atoms in tetrahedrons may be substituted for Al(III) and AI(III) or Fe(III) atoms in octahedrons for Fe(II) or Mg(II). This substitution produces the resulting negative charge of phyllosilicate lattice, that is compensated by Na § K § or Ca 2§ metal cations, located in a position around ,,t-o-t" or ,,o-t" layers [40]. In comparison to the other minerals of this type, kaolinite has the relatively low specific surface (5-40m2/g) and low cation exchange capacity (CEC)[2]. The

363 threelayer minerals, containing water molecules or ions between ,,t-o-t" layers have greater surface for example: illites 100-200m2/g, vermiculites 300-500m2/g or smectites 700-800m2/g. The CEC value is slightly different and may be arranged in a following order illite < smectite < vermiculite [2].

o

0

o

o

,,~si o

~-

o

Si

0

0

A[

o,

O

sO

8i

8i

/

0

o-

8i

O O

\

0

9

OO AI

o .

H H

~

H

Figure 5. The model of the kaolinite structure.

t

O

t Figure 6. The model of structure of the three layer phyllosilicate.

364 According to Sposito, the electric charge on the mineral surface may form because of the isomorphic substitution of crystalline lattice atoms, or the reaction of the surface functional groups. The first one is called permanent charge[5]. For the hydrated oxides or twolayer phyllosilicates (kaolinite) is lower than 0.02 mole/kg, whereas for multilayer phyllosilicates (illite, smectite, vermiculite) the permanent charge is a hundred times higher. Because of acid-base reactions of the soil surface with H § or O H - i o n s the net proton charge is formed (determined as ~H). Innersphere and outersphere complexes also participate in the formation of the electric charge [41]. On the kaolinite, one can distinguish three kinds of the layers exposed to the surface: -the ,,o"-layer of the hydroxyl group connected with aluminum atoms, -siloxane ,,t"-layer, and edge plate of the ,,t-o" layer where =A1OH as well as --SiOH groups can exist. For the substitution of one Si atom by one atom of A1 in the siloxane layer, ,,t", the (=AI-O-Si-) group is formed, donated with negative charge. As it was previously mentioned, this charge is compensated by alkaline or alkaline earth cations. According to Stumm, the hydroxyl groups on the edge layer are characterized by constants pKs = 6.3 and pK~ 2 8.7, like hydroxyl group of A12Oa, =

whereas octaehedron display properties similar to gibsite pK~,~ ~-4. The surface charge comes from both kinds of groups, and pHp~c=7.5 [23]. Considering the solubility of kaolinite and the size of the permanent charge, Sposito found the surface charge density, point of zero net proton charge (p.z.n.p.c.) and point of zero net charge (p.z.n.c.) [42-45]. The determined value p.z.n.c.=3.5 is closer to iep value from electrophoretic measurements of kaolinite than 7.5. This position of p.z.n.c. suggests that the contribution of the hydroxyl group located on the edge in the formation of the charge may be smaller or the groups are more acid. On the basic of the surface charge density and zeta potential measurements, using the FITEQL program, D u e t al. found the ionization constants of surface groups of illite [46]. Because in the examined pH range the illite surface was negatively charged only pKa2 was determined. Good fitting was obtained using the model of energy homogeneous surface for constant pK=4.12-4.23 or for the surface of two kinds group. One of stronger acidity of pKa2(I)=4.17-4.44 and second type weaker acidity of pKa2(n)=6.35-7.74. Specific and nonspecific adsorption of monovalent cations (Cs § Rb+), in the face of high concentration of the ions in the soil solution, will be negligibly small, similarly to metal oxide/ aqueous electrolyte solution system. XPS and NMR examinations, also sorption of Cs § on the kaolinite, showed that considerable amount of caesium adsorb almost completely between the ,,t-o-t" layers, and only 1% of Cs adsorbs on the surface. The adsorption of weekly hydrated Cs § or Rb § ions in the region between layers is stronger in comparison with better hydrolyzed Li § or Na § ions, as these ions give lower potential formation of the edl [47]. The adsorption of cations of the higher radius (hydrated) promotes the swelling of the silt [3]. That explains the observed sequence of monovalent cation adsorption for silt type minerals [3,5,23]:

365 Li + < N a + < K + < R b + < C s +

(14)

The adsorption of the ions on the clay minerals is the exchange type process. The ions from the solution may substitute ions from interlayer for example: R - N a + +Csa+q ~

R-Cs + +Naa+q

(15)

where R - r e p r e s e n t s the phyllosilicate lattice with negative charge. The relative affinity of Cs § ions to the surface of phyllosilicate is described by the total selectivity coefficient: ZcsmNa Kc(Cs_~Na) = ZNamCs

(16)

where: Z- express the phyllosilicate fraction covered by ,,i"-ion (in relation to CEC), mi- molar concentration of ,,i"-ion (Cs § or Na § Though the equation for the selectivity coefficient resemble the equation for the reaction constant of the ion exchange, these coefficients are not thermodynamic constants. That is because the activity of ions in the crystalline lattice of the phyllosilicates are not known [23]. The adsorption of Cs § on the phyllosilicates is a complex process that runs through the succeeding stages. The adsorption of Cs § on K-illite and Ca-illite is well described by three-box model. Here the adsorption of the solution runs through two independent reactions followed by the third one, irreversible [48]. These mechanisms explain partial irreversibility of the Cs adsorption process on the phyllosilicates. It was proved that the adsorption on the illite covered with Ca ions (Ca-illite) goes better that on the mineral covered by potassium ions (K-illite). On the former also more Cs § adsorb in the irreversible way than on the latter. For this course of the reaction on both minerals, the higher distance between layers ,,t-o-t" in Ca-illite is responsible, that favors faster migration in the space between the layers. This migration is also responsible for the higher irreversible adsorption of Cs § on Ca-illite [49]. For divalent cation adsorption, the reaction of substitution runs through exchange of Me 2+ for two monovalent anions for example: 2 R - N a + +Me2q

~_ R2-Me 2+ +2Naa+q

(17)

and selectivity coefficient will be equal: ZMe m2Na Kc(Me-~Na) = Z~qam Me

(18)

366 The selectivity coefficients of the alkaline earth cation adsorption on the montmorylonite and vermiculite arrange in the following sequence [35]: Mg 2+
(19)

The investigations of the Sr 2§ and Ba 2§ adsorption on the kaolinite, illite and bentonite, covered earlier by K, Ca and AI(III) ions, proved that the process is well described by Freundlich or Dubinin-Radushkievich isotherms. The adsorptionconcentration dependence is poorly fit by Langmuir isotherm here. The adsorption of the ions is higher on the minerals covered K § than Ca 2§ or AI(III) [50]. For the adsorption of these ions the poor dependence from pH was noticed. The adsorption of heavy metal ions, such as Cd(II), Ni(II), Cd(II), Pb(II) is a more complicated process. Depending on the pH and concentration of the solution, beside the adsorption also the surface or bulk precipitation may take place. Cation adsorption on the silt type minerals may occur on hydroxyl groups of the layers exposed to the solution, on edges of the ,,t-o-t" layers (,,t-o" for kaolinite). It can run also by exchange with cations from the space between layers. It is considered that the siloxane layer (-Si-O-Si-) is not active towards the multivalent cations adsorption [35]. Schindler et al. found that the adsorption of Cd(II), Cu(II), and Pb(II) on kaolinite from aqueous solutions occurs on two types of sites, one of week acid character and another one on surface hydroxyl groups connected with aluminum [51]. The adsorption on week acid groups is ion exchange adsorption type and goes at low pH and ionic strength values. Higher values of these parameters favor adsorption on hydroxyl groups, resulting in the innersphere formation. The adsorption of Cu(II) and Pb(II) cations as a function of pH is similar for oxides/electrolyte systems, in certain pH range one can observe the increase of adsorption defined as the edge of adsorption. Differently, the adsorption of Cd(II) at increasing part of adsorption vs pH curve shows two distinct edges, one below and another above pH=6.5. Angove and coworkers, from the results of the Cd(II) adsorption and potentiometric titrations, found out the adsorption for pH_<4 is the exchange type adsorption on the permanent, negatively charged sites of the siloxane layer. This process is characterized by inconsiderable stoichiometry of a proton exchange (0.2) [52]. The adsorption on the hydroxyl groups takes place on the surface of the octahedron layer (A1) with characteristic, for alumina stoichiometry of the proton exchange. The -SiOH groups are in this adsorption of minor importance. In the similar way runs the adsorption of Co(II) in the system smectite/electrolyte solution [53]. Low values of pH and ionic strength promote the adsorption on permanent centers of the charge with formation of outersphere complexes, whereas increase of NaC1 concentration results in removing Co(II) from adsorption sites of permanent charge and favor multinuclear complex formation and surface precipitation. The size of these complexes increases with the increase of pH. XAS(X-ray Absorption Spectroscopy) investigations showed that distance Co-Co between atoms is smaller than in Co(OH)2. That may prove the adaptation of the

367 forming complexes to the crystalline lattice of the phyllosilicate. The smaller distances between adsorbed cations than between respective hydroxides were also observed for Ni(II) on many silty minerals [54]. In present paper, authors tend to opinion, that the smaller distance between adsorbed cations results from the formation of a mixing phase of the nickel-aluminum hydroxide. To the Pb(II) adsorption on the kaolinite data Majone et al. fitted different adsorption models. They stated that good fitting of the adsorption of Pb(II) as a function of pH and ionic strength of the electrolyte data is obtained using the threelayer model of the edl (TLM) characterized by two types of adsorption sites and their continuous distribution [55]. The adsorption of UO22§ on smectite runs in similar way to cation adsorption and is characterized by typical for them the adsorption edge [56]. The increase of Na § and Ca 2§ ions concentration in solution gives a shift of the edge towards higher pH values. In this system the adsorption model is very complicated because of many different ion forms of UO22§ and a solubility of smectite. In the investigations on the adsorption of Am-241 on the montmorylonite (the swelling mineral), illite (nonswelling) no significant difference was noticed, regarding their swelling behavior. That proves the actinides adsorption only on the surface of the minerals [57]. The adsorption of cations on the surface of silts may be described by the edl models characteristic for oxides. They should consider also the substitution adsorption processes on the interlayer surfaces of the silts, effects connected with energetic heterogeneity of the surface and dissolution of the solid. 4.

CATION A D S O R P T I O N ON C A R B O N A T E S

Calcium carbonates, beside oxides and silts are the main parts of soil components. Beside carbonates, recognized as minerals, other oxides may display surface properties of carbonates because of CO2 adsorption, though the structure of the bulk phase is oxide type. In this group may be included for example iron oxides [58]. A influence of carbon dioxide was also observed with the silt surface. Calcite and dolomite are the most popular carbonate minerals presented in soils, beside them siderite FeCO3 and rhodocrysite MnCO~ may be present. Because carbonates dissolve in aqueous solutions the determination of surface charge by potentiometric titration is difficult. The additional problem is caused by partial carbon dioxide pressure that influences the balance of the system. The behavior of the closed and opened to atmosphere systems is quite different. It is assumed that for the presence of anion and cation in the crystalline lattice of carbonates, on their surface the following reaction led to charge accumulation: -CO3H~=~

-CO~+H +

-C03H + Me 2+

~:~

-CO3Me+ + H +

(20) (21)

368 - M e O H ~ ~:t

(22)

-MeOH + H +

4:t -MeO-+ H + - MeOH+ CO2 ~ -MeCO~+H+

(23)

- MeOH

(24)

The charge density on the surface of carbonates is proportional to the algebraic sum of the concentrations of the respective forms bearing the charge [59], Go

: F{[-CO3H~]+[-C03Me+]+[-MeOH~1-[-MeO-]- [-MeCOg]- [-C03]}(25)

An electric charge on the surface of the carbonate, as of oxides, forms the electric double layer on the interface of the mineral. The presence of edl implies the occurrence of specific and nonspecific adsorption at the interface. Caesium ions may also adsorb on carbonates but because of higher concentrations of other cations of alkaline metals in the soil solution their sorption is probably minute. Divalent cations of heavy metals Co(II), Pb(II), Ba(II), Sr(II) show inclination to the formation of isomorphic crystals with Ca, Fe and Mn carbonates. Then, beside the adsorption in the edl a substitution adsorption may take place between Me 2§ ions from the solution and e.g. Ca e§ ions from a crystal lattice of the carbonate. This adsorption was observed for FeCOa/Mn 2+ aqueous solution and CaCO3/Cd 2+ solution systems [60,61]. In both, the adsorption process is more complicated the substitution adsorption is the first quick stage, then a diffusion to the hydrated layer of the carbonate proceeds and finally diffusion into the solid. (CaCO3XCaCO3 * H20)Cd 2+"

(26)

(CaCO3XCdCO3*H20)s+Ca2+

(27)

(CaCO3XCaC03 * H20) s + Cd 2+ ~ (CaCO3 XCaCO3* H20)Cd s2+~:t

(CaCO 3 XCdC03 * H20)s + c a 2+ ve

((Ca, c a ) c o 3 XCdC03 * H20)s + Ca 2+

(28)

In Davie's et al. opinion, the plot of Cd 2§ adsorption versus pH confirms the above mechanism of the process, the adsorption lowers with the increase of pH. Despite the presence of hydroxyl groups on the surface, the substitution adsorption process cannot run with hydrogen ion liberation. In these condition the increase of pH is accompanied with the increase of the adsorption, for example metal oxide/Me 2§ aqueous solutions. The observed decrease of adsorption Davies et al. explain by the exchange for Ca 2§ ions [61]. After all, this interpretation of adsorption versus pH dependence seems difficult to accept because in the examined pH range (pH6-8), the concentration of Ca 2§ ions decreases [62]. When assuming the Cd-Ca ion exchange mechanism, the increase of the adsorption with the increase of pH (decrease of H § concentration) should be observed. The increase of Me 2§ (Me=Cd, Zn, Mn, Co, Ni, Ba, Sr) adsorption on calcite was observed by Zachara et al. [63]. In their paper the following sequence of the cation adsorption on calcite was found:

369 Cd 2+ > Zn 2+ > Mn 2+ > Ni 2+ >> Ba = Sr

(29)

From the experiments on the divalent cation adsorption, the character of their substitution adsorption with Ca 2§ ions was proved [63, 64]. Additionally the following dependence was observed: the higher hydrating energy of the cation, the easier its desorption. Beside the Me 2§ ion adsorption, also the dissolution and recrystallization processes occur in carbonate/electrolyte solution system. These processes may be responsible for formation of solid solutions of carbonates (Me 2§ ion transportation into the solid phase). The influence of the recrystallization on the diminution of the cation from the solution is visible at longer adsorption time, and may be interpreted as "diffusion" into the solid phase. During recrystallization, the cations from the surface are covered and then the process lead to incorporation of adsorbed ion into crystal lattice of a solid phase which is one of the mechanisms typical for the system consists of cocrystallised micro- and macro constituents [13, 61, 65]. The description of Cd(II) adsorption process on CaC03, which considers recrystallization of the solid was proposed by Das and Van der Weijden [66]. 5.

A D S O R B I N G A N D C O M P L E X I N G P R O P E R T I E S OF ORGANIC S U B S T A N C E S OF S O I L

Considering their origin, organic substances presented in soil may be divided into two groups. The first one includes all substances, result from the natural, biological processes that happen in the environment, the second one contain all substances introduced by man and his industrial activity [3]. In the first group there are substances of small and big molecular weight such as acids, amines and aminoacids. The most important carboxylic acids are: oxalic, formic, citric, acetic, succinic, malonic, maleic, aconitic and fumaric. Their concentration in cultivated soils is lower than in forest ones [39]. The group of substances of the high molecular weight includes lignins, celluloses, simple proteins and products of their degradation. All these organic substances present in soil show various properties and different solubility in water solutions. An amount of the organic substance dissolved in water is defined as a dissolved organic carbon (DOC) and measure in mg of carbon/dm 3. To the group of organic substances of the anthropogenic origin may be included detergents, sulfonic acids, whitening agents, polymers, solvents, fuels and so on. Some of them dissolve well in water and form complexes or precipitates with heavy metals. Organic substances presented in soil and called "humic substances" do not have well-determined chemical constitution. They are formed mainly because of biological degradation of lignins, proteins and carbohydrates (mainly cellulose). Substances of smaller molecular weight, formed at the beginning and bonding

370 together gives humic macromolecules in the end [5]. For the different substances may react in this way the final particle may has the variety of functional groups: COOH, -OH (phenolic and alcoholic)=CO, -COH, -NH, -SOH, N in heterocyclic orimides and amides [5]. The average molecular weight of fulvic acid is 670 a.m.u. and molecule contains six carboxyl and five phenolic groups [4]. The average humic acid particle is bigger and may reach 25 000 a.m.u.. The humic substances reveal properties similar to polyelectrolyte gels because of their three-dimensional structure. The dissociation of carboxyl or phenolic group forms the electric potential around the humus molecule or their gel phase (for the insoluble humus). To neutralize this charge, some cations from the electrolyte accumulate at humus molecule [67]. The analysis of the H § affinity to the functional groups of humic acid, based on the model of continuous adsorption sites (energetic heterogeneity), showed two peaks on the distribution curve. One for carboxyl groups of p K H Int ~ 4 and second characteristic for phenol groups pKHInt -8-9 [68]. Ephraim and coworkers believe that humic acids behave in intermediate manner between simple electrolyte and polyelectrolyte; as oligoelectrolytes with discrete distribution of acidity constants of functional groups [69]. Marinsky et al., proposed the method for the calculation of ionization constants of carboxyl groups and ionization constants of metal cation complexation including chelate complexes [70]. They proved that fulvic acid contains three types of carboxyl groups (I,II, III), characterized by different dissociation constants (pZaI - 1.2, pZaII - 3.4, and pZaIII - 4.2) and acidic alcohol (enol) groups of the constant pKaIv =5.7. For metal complexation, there is possibility to obtain four unidenate and four chelatic species[70]. The organic substances of small molecular weight, show complexing behavior not only with metal cations in solutions but also may complex these cations that are in solid phase, for example A1 or Fe. This action promotes the solubility of oxides or other minerals [71]. The solubility of fulvic acid complexes with A1 or Fe cations, depends on their mutual ratio (metal cation/fulvic acid). For the ratio equal to one the complex is soluble. For higher ratio values (3-6) the solubility decreases [4]. Acid properties of humic substances may be learned by potentiometric titrations [23,72]. From these data the distribution of pK constants of humus functional groups may be found [73]. The complexing properties of humic and fulvic acids are the subject of many papers dealing with cation adsorption and the references can be find in many reviews [4,5,38,71]. Harter and Naidu presented values of complexation constants for some heavy metal cations: Co, Cd, Cu, Fe, Mn, Ni, Pb, Zn, by ligands existing in aqueous solutions of soil [39]. The metal cation adsorption on humic acid particles is complex process because of polyfunctional character of acid group, its polydispersity and existence in dissolved and colloidal form [38]. For simplification the adsorption reaction is assumed to run on quasiparticle with groups of the acid character: aSHn(aq or s)+ pMe 2+ + qL1- + xH + + y O H - ~-where: 5=2p-x-y-lq and b = n ' a ,

SaMep(OH)xLSq(aq or s) + bH+ (30)

371 Above equation may be reduced to simpler form regarding that humic acid is dissociated in aqueous solution [38]: a L H - + Me 2+ ~-

MeL2a(l-a) + a H +

(3~)

This equation is characterized by the constant KH: KH = [MeL2(1-a) ][H +

(32)

[LH- ~ [Me 2+ ] It can be noticed from the above equation, that the metal adsorption on humic acid should increase with the increase of pH. This behavior was observed for Pb(II), Cu(II), Cd(II) and Ca(II) [38]. The organic substances with groups of acidic (or alkaline) character are active towards molecules of opposite character or to surface groups having alkaline(or acidic) properties. Molecules of humic substances may adsorb on the soil minerals by the interaction of their functional groups with mineral surface. The presence of aliphatic chain segments or aromatic rings enables humus particles to the disperse interaction. Sposito distinguished following mechanisms of the organic substances of soil: - cation exchange, - protonation, - anion exchange, - water bridge formation, - cation bridges, - ligand exchange, - hydrogen bonding, - Van der Waals interaction. Humic acid adsorbs on the oxides and on kaolinite in the way characteristic for anions, the adsorption decreases with increase of pH [23,71]. Because of the size of particle and n u m b e r of functional groups the adsorption may run only for the part of molecule, such as for other macromolecules. Thus, h u m u s particles may immobilize the nonionic particles (organophosphates halogen derivatives of hydrocarbons) alkaline type organic particles or metal cations [4,23]. Humus particles compete with organic acid particles of small molecular weight in adsorption and complexation reactions [71]. The adsorption isotherm of humic acid on the kaolinite does not show the tendency to reach plateau, characteristic to complete coverage, so the humic acid adsorption is multilayer type. The reasons for this are nonpolar interactions of hydrophobic segments of acid molecules [74]. It was observed t h a t humic acid adsorbs in lower degree t h a n fulvic at the same concentration because greater particles of humic acids cover up neighbouring

372 adsorption sites. The adsorption of fulvic acid, which particles are smaller, is a Langmuir type (monolayer adsorption) [74]. The adsorption of Cu(II) on the kaolinite, covered previously by humic or fulvic acid showed that: - adsorption affinity of Cu(II) ions to humic acid is stronger than to fulvic acid - adsorption affinity of Cu(II) ions does not depend on the molecular weight of humic acid, and is always the same. - Cu(II) and H § ions adsorb completely on the same adsorption sites. In the soil/electrolyte system coexist: a solid inorganic phase, humic substances and solution that contains, among others, the metal cations. Model investigations of such systems are done mainly for metal oxide (or phyllosilicate) - humic acid electrolyte solution [75-85]. The presence of DOC changes the mechanism of cation adsorption because of cation complexation. On the one hand, the complexation leads to lowering the cation adsorption at higher pH values, according to the mechanism of reaction 12. On the other hand, the cation adsorption on the solid takes place, according to reaction 13, through the adsorption of ligands. For humic acid, because of the existence of many different functional groups, the adsorption may run according to both mechanisms (reactions 12 and 13). This effect was observed for the Cu(II) adsorption on A1203 and presence of humic acid [75]. Fulvic acid lowers the Zn(II) adsorption on goetite and hydroarargite by complexing the metal cations in the solution (reaction 11) and has no influence on this cation adsorption on SiO2 [85]. The adsorption of rare earth elements, in the presence of humic acid, is complex process and differs much from the solutions without this acid [76-80]. This adsorption is connected with humic acid complex formation with Eu(III) ions in the solution, adsorption of these complexes also the adsorption of ionic form of Eu(III) on the adsorbed humic acid on the oxide or on the phyllosilicate. Similar effect of humic acid on the adsorption, was observed for ions of V, Ag and other rare earth elements [85]. Beside discussed group of ions of rare earth elements (group A), Takahashi et al. distinguished three groups of ions of elements. They differ in their adsorption behavior in the environment humic acid/ dispersed oxides (phyllosilicates) [85]. Group B - (Mn,Zn, Co, Be, Sr, Ba, Fe, Cr) - humic acid causes desorption of these cations in neutral pH. Group C - (Ru, Rh, Ir, Pt, Ga, Zr, Hf) - humic acid does not have influence on the cation adsorption but earlier adsorbed acid molecules limit Hf and Zr ions adsorption through screening the adsorption sites. Group D - (Rb, As, Se, Te) - the presence of humic acid does not have any influence on the adsorption of these ions. The description of adsorption process in the metal cation - humic acid - mineral system is far more complicated than for the system without acid. Till now, there are some methods used for the description of adsorption in systems containing humic (fulvic) acid; LOGA J.C.M. de Wit et al. [86], NICA model (by Bendetti et al.) [87]. One can also adopt adsorption model that considers energetic heterogeneity of adsorption sites [33].

373 Another review dealing with actinides complexation by humic substances was presented by Maulin et al. [88]. In this paper some complexation constants of selected lanthanides and actinides were given. The complexed forms of actinides with humic acid dominate in solutions of pH<7 (sometimes 8) at concentrations 0.1 mg/dm 3. Whereas the presence of other cations (AI(III), Ca(II)) may change the contribution of h u m u s complexes of actinides in aqueous phase. To obtain the adequate opinion of the existing forms in the system, the calculations may include the concentrations of anions and cations in the aqueous environment of soil.

Q

P R O C E S S E S OF FORMATION, T R A N S P O R T A T I O N AND A D S O R P T I O N OF COLLOIDS IN SOIL SYTEMS

Heavy metal cations and actinides may often precipitate under the soil solution conditions. The average concentration of anions that form insoluble sediments in ground or surface water is often sufficient to form carbonates, hydroxides, sulfides, phosphates, fluorides or chlorides with respective cations. If in discussed system some colloids are formed, then processes characteristic for real solutions cannot explain the ion transportation in natural environment. Obtained as a results of radioactive isotope precipitation, fine sediments (insoluble salts) are called real colloids. It was mentioned earlier that radioisotopes under specific conditions may adsorb on fine dispersed oxides or hydroxides of colloid size. Then, the behavior of this system will not be the same as for real solutions, but rather as the colloid systems. Isotopes, adsorbed on colloid matrix form pseudocolloid [3]. Beside two mentioned colloid systems, Kim distinguishes in natural systems the third kind aqua colloids. They are formed as a result of succeeding reactions: dissolution of the mineral, hydrolysis of obtained product, polinucleation and colloid formation [89]. Depending on the solid contacted with water, the aqueous solution can contain even more than 100 ppm of colloids. Usually high concentrations of aqua colloids exist in water that contact with humus. Actinides form insoluble sediments with hydroxides, carbonates, sulfates, phosphates and fluorides [90]. There are no data concerning the solubility of actinides with silicate anion, though uranium forms with this anion several insoluble minerals. Because in natural conditions in ground water the hydrocarbon and carbonate anions play the dominant role, in the reactions of precipitation (in the 0.01 mole/dm 3 solution of Na § pH=7, pC02 =3.5) depending on the oxidation state of the actinides, the compounds presented in Table 3 may precipitate [90]. The behavior of actinides in natural environment was described in many papers reviewed by Lieser [3,89], Silva and Nitshe [90], Kim [91,92], Chopin and Stout [93], Newton and Sulivan [94], Larsen et al. [95] and Tanaka et al. [96]. The behavior of colloids or pseudocolloids in natural soil and water systems is a subject of the investigations in the aspect of the transportation not only the toxic substances but also gaining precious minerals or elements, for example Au [97]. The examination of the behavior of Pu in alluvial sediments of Los Alamos depository revealed that Pu loaded on the colloid, translocates few times slower

374 Table 3 Solid phase and solubilites of actinides[90] Oxidation state of An

Solid phase

+3 +4 +5 +6

AnOHC03 AnO2 NaAnO2CO3 AnO2(OH)2*H20

Solubility of An [mol/dm 3] 10 .7 10 -10 10 .5 10 .6

than tritium but more than thousand times faster then it was predicted by two or three phase model [6]. The model investigations of colloid transportation through columns packed with quartz, showed that the migration rate of bigger particles, for example latex, is greater than that of smaller ones [98, 99]. To describe the latex transportation in a column, the model similar to dynamic chromatography was proposed. The transportation in quartz packed column is treated as a particle transportation through capillary. The average rate of particles (v), depends on equivalent radius of capillary Ro, the rate of the liquid (velocity profile fluid) Vr, rate of particle Rp and energy of interaction between particle and capillary (packed quartz) W [99].

Vtr),exE Wr)lrr (v)=

kT

J

(33)

~:~ -RP e x p l - :T(r)]rdr The energy of interaction consists of Van der Waals disperse interaction WVDWand electrostatic interaction of edl WDL W = WD1 + WVDw

(34)

For RpfRo << 1 the interactions between particle and package of the column (quartz grains) may be treated as sphere - plate interaction. Then, the electrostatic interaction will vary with particle potentials ~gl and the plate potential ~2 WD1 (h)= 16~ .Rp.

tanh ewe//.tanh e~2/.exp -Kh 4kT J \ 4kT J

where: h- distance between plate and sphere e- elementary charge ~- dielectric constant K-reciprocal Debaye length K = ~

1000e2N Av -~ ~i ziM 2 i

(35)

375 NAv- Avogadro's number z - valence of the ion M - molar concentration. The above equation couples the influence of electric potentials on the surface of particles and ionic strength of the electrolyte on the electrostatic force interactions. An increase of ionic strength results in drop of WDL by the increase of • parameter. For spherical particle - plate system the Van der Waals force interactions are described by following equation: WVDW (h) = -

+~ + 21n x+l x+l

(36)

where x=h/2R~ A- Hamaker constant. Hamaker constants for pure components of discussed systems are available in many monographies dealing with stability of dispersed systems [100]. For real systems, for example latex-quartz interactions in water, the H a m a k e r constant have to be calculated from pure component data.

Wv w: (A, 11-

X

(37)

where: All, A22 H a m a k e r constants for solids (1- latex and 2 - quartz), A33 Hamaker constant of the liquid medium (water). Velocity of particle migration in the porous systems, calculated in this way, in the opinion of Nagasaka et al., may well predict the behavior of colloids in geological systems[99]. This model contains many simplifications: it assumes a small size of colloid particles in comparison to the packing of the column, does not take into account precipitation on and liberation particles from collectors, heterogeneity of the package, and dynamics of processes and at last variation of chemical and physical conditions. In bibliography, there are known also the more developed models of porous collectors and the particles transportation [101-104]. However, despite many models of the colloid particle - collector interactions, which may be successfully adopted to well-defined systems, their application in such natural systems as soil still does not give the precise description of the transportation process. Comprehensive, excellent review of the state of present knowledge on the colloid transportation in the natural environment was presented by Ryan and Elimelech [6].

REFERENCES 1. W.P. C u n n i n g h a m and B.W. Sato, Enviroment Science, A Blobal Concern, Brown Publishers, Dubuque IA ,1990, 182. 2. B.J. Alloway, Heavy Metals in Soils, Blackie Acad. & Professional, London, 1995.

376 3. K.H. Lieser, Radiochimica Acta, 70 (1995) 355. 4. G.H. Bolt, Soil Chemistry B: Physico-chemical Models, Elsevier Sci. Pub. Comp., New York, 1982. 5. G. Sposito, The Surface Chemistry of Soils, Oxford University Press, New York, 1984. 6. J.N Rayan and M. Elimelech, Colloids Surf., 107 (1966) 1. 7. C.P. Huang, J. Berthelin, J.-M. Bollang, W.B. Mc Gill and A.L.Page (eds.), Environmental Impact of Soil Component Interaction Metals, Other Inorganics and Microbial Activities, CRC, 1995. 8. R.O. James and G.A. Parks, Characterization of Aqueous Colloids by Their Electrical Double-Layer and Intrinsic Surface Chemical Properties, in: Surface and Colloid Sci., Wiley-Interscience, New York, 1982, 12, 119. 9. T. Hiemstra, P. Venema and W.H. Van Riemsdijk, J. Colloid Interface Sci., 184 (I 996) 687. 10.I.D. Brown, Chem. Soc. Rev., 7 (1978) 359. 1 I. G.R. Wiese, R.O. James, D.E. Yates and T.W. Healy, Electrochemistry of the Colloid - Water Interface, in: Electrochemistry of Colloid-Water Interface, Int. Review Sci. Phys. Chem., Series 2, Butterworths, London, 1966, 53. 12. W. Janusz, I. Kobal, A. Sworska and J. Szczypa, J. Colloid Interface Sci., 187 (1997) 381. 13. P. Benes and V. Majer, Trace Chemistry of Aqueous Solutions, Academia, Prague, 1980. 14. N. Haken, B. Fourest, G. Guillaumont and N. Marmier, Radiochim. Acta, 74 (1996) 225. 15.M. Kosmulski, J. Radioanal. Nucl. Chem., Letters, 118 (1987) 209. 16.M. Kosmulski, J. Radioanal. Nucl. Chem., Articles, 129 (1988) 149. 17.M. Kosmulski, A.L. Dawidowicz and J. Szczypa, J. Radioanal. Nucl. Chem., Articles, 131 (I 989) 377. 18.M. Kosmulski, A.L. Dawidowicz, J. Szczypa and A. Waksmundzki, Sep. Sci. Technol., 25 (1990) 953. 19.M. Kosmulski, A.L. Dawidowicz and J. Szczypa, Int. J. Radiat. Appl. Instrum. Part A., 41 (1990) 239. 20.M.A. Shenber and K.J Johanson, Radiochim. Acta, 57 (1992) 165. 21. J. Jablonski, W. Janusz, R. Sprycha and M. Reszka, Physicochem. Problems Miner. Process., 22 (1990) 2 I. 22.A.P. Robertson and J.O. Leckie, J. Colloid Interface Sci., 188 (1997) 444. 23. W. Stumm, Chemistry of the Solid-Water Interface, J. Wiley&Sons, New York, 1992. 24. K.F. Hayes and L.E. Katz, in: Physis and Chemistry of Mineral Surfaces, P.V. Brady (ed)., CRC Press, Boca Raton, 1996. 25.R.O. James and T.W. Healy, J. Colloid Interface Sci., 40 (1972) 42. 26.P.W. Schindler, in: Adsorption of Inorganic at Solid/Liquid Interfaces, Ann Arbor, MI, Ann Arbor Sci, 198 I, I.

377 27.A.L. Herbelin and J. Westall, A Computer program for determination of chemical equilibrium constants from experimental data. Rapport 96-01, Oregon State University, Cornvalis, OR, 1996. 28.L.E. Katz and K.F. Hayes, J. Colloid Interface Sci., 170 (1995) 491. 29. K.Bourikas, H.K. Matrialis, Ch. Kordulis and A. Lycourghiotis, J. Phys. Chem., 100 (1996) 11711. 30. Ch. Ludwig and P.W. Schindler, J. Colloid Interface Sci, 169 (1995) 284. 31.A. Trkov, N. Ogrinc and I. Kobal, Computers Chem., 16 (1992) 341. 32. W. Rudzifiski, R. Charmas, S. Partyka and J.Y. Bottero, Langmuir, 9 (1993) 2641. 33.M. Cernik, M. Borkovec and J.C.Westall, Langmuir, 12 (1996) 6127. 34.P.W. Schindler and W. Stumm, in: Aquatic Surface Chemistry, W. Stumm (ed.), J. Wiley & Sons New York, 1987, 83. 35.D.G. Kinniburgh, in: Adsorption of Inorganic at Solid/Liquid Interfaces, Ann Arbor, MI, Ann Arbor Sci., 1981, 91. 36. C.P. Huang, in: Adsorption of Inorganic at Solid~iquid Interfaces, Ann Arbor, MI, Ann Arbor Sci., 1981,183. 37. D.A. Dzonbak and F.M.M Morel, Surface Complexation Modeling, Hydrous Ferric Oxide, J. Willey& Sons, New York, 1990. 38. Z. Wang and W. Stumm, Neth. J. Agric. Sci., 35 (1987) 231. 39.R.D Harter and R. Naidu, Adv. Agronomy, 55 (1995) 219. 40. C.S. Hurbult Jr and C. Klein, Manual of Mineralogy, J. Wiley & Sons New York, 1977, 391. 41.A. Deville, J. Phys. Chem., 97 (1993) 9703. 42. G. Sposito, Chimia, 43 (1989) 160. 43.J. Chorover and G. Sposito, Geochim. Cosmochim. Acta, 59 (1995) 875. 44.J. Chorover and G. Sposito, Geochim. Cosmochim. Acta, 59 (1995) 3109. 45. G. Sposito, Clays. Clay Minerals, 45(1997) 85. 46. Q. Du, Z. Sun, W. Forsling and H. Tang., J. Colloid Interface Sci., 187 (1997) 221. 47.D.Y.C. Chan, R.M. Pashley and J.P. Quirk, Clay & Clay Minerals, 32 (1984) 131. 48.R.N.J. Comans and D.E Hockley, Geochim. Cosmochim. Acta, 56 (1992) 1157. 49.R.N.J Comans, M. Hailer and P. de Preterm, Geochim. Cosmochim. Acta, 55 (1991) 433. 50. H. Gokturk, C. Eylem, S. Hatipoglu and H.N. Erten, J. Radionalytical. Nuclear Chem., 198 (1995) 449. 51.P.W. Schindler, P. Liechti and J.C. Westall, Neth. J. Agric. Sci., 35 (1987) 219. 52.M.J. Angove, B.B. Johnson and J.D. Wells, Colloids Surf., 126 (1997)137. 53. Ch. Papelis and K. Hayes, Colloids Surf., 107 (1996) 89. 54.A.M. Scheidegger, L. Lamble and D.L. Sparks, J. Colloid Interface Sci., 186(1997) 118. 55.M. Majone, M.P. Papini and E. Rolle, J. Colloid Interface Sci., 179 (1996) 412.

378 56. G.D. Turner, J.M. Zachara, J.P. McKinley and S. C. Smith, Geochim. Cosmochim. Acta, 60 (1996) 3399. 57. C. Degueldre, H.J. Ulrich and H. Silby, Radiochimica Acta, 65 (1994) 173. 58.A. Van Geen, A.P. Robertson and J.O. Leckie, Geochim. Cosmochim. Acta, 58 (1994) 2073. 59.P.Van Cappellen, L. Charrlet, W. Stumm and P. Wersin, Geochim. Cosmochim. Acta, 57 (1993) 3505. 60.P. Wersin, L. Charlet, R. Karthein and W. Stumm, Geochim. Cosmochim. Acta., 53 (1989) 2787. 61.J.A. Davis, Ch. C. Fuller and A. D. Cook, Goechim. Cosmochim. Acta., 51 (1987) 1477. 62.P.Somasundaran and G.E. Agar, J. Colloid Interface Sci., 24 (1967) 433. 63.J.M. Zachara, C.E. Cowan and C.T. Resch, Geochim. Cosmochim. Acta, 55 (1991)1549. 64.J.M. Zachara, J.A Kittrick and J.B. Harsh, Geochim. Cosmochim. Acta, 52 (1988) 2281. 65.H.A.C. McKay, Principles of Radiochemistry, Buttrewoths, London, 1971, 378. 66. H.A. Das and R.D. van Der Weijden, J. Radioanal. Nuclear Chem., 204 (1996) 205. 67.M.F. Benedetti, W.H. Riemsdijk, L.K.Koopal, D.G. Kinniburgh, D.C. Gooddy and Ch. J. Milne, Geochim. Cosmochim. Acta, 60 (1966) 2503. 68. C.J. Milne, D.G. Kinniburgh, J.C.M. de Wit, W.H. van Rjiemsdijk and L.K. Koopal, Geochim. Cosmochim. Acta, 59 (1995) 1101. 69.J.H. Ephraim, C. Patterson, M. Norden and B.Alard, Environ. Sci. Technol., 29 (1995) 622. 70. J.A. Marinsky, M.M. Readdy, J.H. Ephraim and A.S. Mathuthu, Anal. Chim. Acta. 302 (1995) 309. 71.J.G.Hering, Interaction of Organic Matter with Mineral Surfaces, in: Aquatic Chemistry, C.P. Huang, C.R. O'Malia and J.J. Morgan (eds.), Adv. Chem., Ser 244, ACS Washington, 1995, 95. 72.P.A. Arp, Canadian J. Chem., 61 (1983) 1671. 73.M. Cernik, M. Borkovec and J.C.Westall, Environ. Sci. Technol., 29 (1995) 413. 74. C.P. Huang and Y. L. Yang, Water Res., 29 (1995) 2455. 75.J.A. Davis, Geochim. Cosmochim. Acta, 48 (1984) 679. 76.Y. Takahashi, Y. Minai, Y. Meguro, S. Toyoda and T. Tominaga, J. Radioanal. Nucl. Chem., Letters, 186 (1994) 129. 77.A.J. Fairhurst, P.Warwick and S. Richardson, Colloids Surf., 99 (1995) 187. 78.A. Ledin, S. Karlsson, A. Duker and B. Allard, Radiochimica Acta, 66/67 (1994) 213. 79.A.J. Fairhurst, P.Warwick and S. Richardson, Radiochimica Acta, 69 (1995) 103. 80.Y. Takahashi, T. Kimura, Y. Kato, Y. Minai and T. Tominaga, Chem. Commun. 2 (1997) 223.

379 81.M.F. Benedetti, Ch. J. Milne, D.G. Kinninburgh, W.H. Van Riemsdijk and L.K.Koopal, Environ. Sci. Technol., 29 (1995) 446. 82. Ch. J. Milne, D.G. Kinniburgh, J.C.M. de Wit, W.H. Van Riemsdijk and L.K.Koopal, J. Colloid Interface Sci., 175 (1995) 448. 83.A. Ducker, A. Ledin, S. Karlsson and A. Allard, Appl. Geochem., I0 (1995) 197. 84.Y. Takahashi, Y. Minai, T. Ozaki, S. Ambe, M. lwamoto, H. Maeda, F. Ambe and T. Tominaga, J. Radioanal. Nucl. Chem., Article, 205 (1996) 255. 85.Y. Takahashi, Y. Minai, S. Ambe, H. Maeda, F. Ambe and T. Tominaga RIKEN Rev., 13 (1996) 1 I. 86.J.C.M. de Wit, W.H. Van Riemsdijk, M.M. Nedrlof, D.G. Kinniburgh and L.K. Koopal, Anal. Chim. Acta, 232 (1990) 189. 87.M.F. Benedetti, W.H. Van Riemsdijk, L.K. Koopal, D.G. Kinniburgh, D.C Gooddy and Ch. J. Milne, Geochim. Cosmochim. Acta, 60 (1996) 2503. 88.V. Moulin, J. Tits and G. Ouzounnian, Radiochimica Acta, 58/59 (1992) 179. 89. K.H. Lieser, R. Hill, U. Muhlenweg, R.N. Singh, Tu Shu-De and Th. Steinkopff, J. Radioanal. Nuclear Chem., 147 (1991) 117. 90.R.J. Silva and H. Nitsche, Radiochimica Acta, 70/71 (1995) 377. 9 I. J.I. Kim, MRS Bulletin XIX(1994) 47 Forschungszentrum, Karlsruhe. 92.J.I. Kim, in: Handbook on the Physics and Chemistry of the Actinides, A.J. Freeman and C. Keller (eds.), Elsevier Sci. Pub. B.V., Amsterdam 1986, 413. 93. G.R. Chopin and B.E. Stout, Sci. Total Environ., 83 (1989) 203. 94. T.W. Newton and J.C. Suliwan, in: Handbook on the Physics and Chemistry of the Actinides, A.J. Freeman and C. Keller (eds.), Elsevier Sci. Pub. B.V. Amsterdam, 1986, 387. 95.R.P. Larsen, D.M. Nelson, M.H. Bhattacharyya and R.D. Oldham, Health Physics, 44 (1983) 485. 96. S. Tanaka, S. Nagasaki, H. Itagagi and M. Yamawaki, in: High Level Radioactive Waste and Spent Fuel Menagment, S.C. Slate, R. Kohout and A. Suzuki (eds.), Am. Soc. Mech. Engng., 1989, 375. 97.M. Benedetti and J. Boulegue, Geochim. Cosmochim. Acta, 55 (1991) 1539. 98. S. Nagasaki, S. Tanaka and A. Suzuki, J. Nuclear. Sci. Technol., 30 (1993) 1136. 99. S. Nagasaki, S. Tanaka and A. Suzuki, J. Nuclear. Sci. Technol., 31 (1994) 143. i00. P.C. Hiemenz and R. Rajagopalan, Principles of Colloid and Surface Chemistry, Marcel Dekker, New York, 1997. I01. L. Song and M. Elimelech, J. Colloid Interface Sci., 167 (1994) 301. 102. L. Song and M. Elimelech, Colloids Surf., 73 (1993) 49. 103. M. Elimelech, Sep. Technol., 4 (1994) 186. 104. Z. Adamczyk, B. Siwek, M. Zembala and P. Belouschek, Adv. Colloid Interface Sci., 48 (I 994) 15 I.

Adsorption and its Applicationsin Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski(Editor) 9 1998Elsevier Science B.V. All rights reserved.

381

N e w c o m p o s i t e a d s o r b e n t s for t h e r e m o v a l of p o l l u t a n t s f r o m w a s t e waters E.F. Vansant University of Antwerp (UIA), Department of Chemistry, Universiteitsplein 1, B-2610 Wilrijk, Belgium

ABSTRACT Chemical modification techniques were developed to create new composite adsorbents from elutrilithe to improve the removal of neutral, anionic and cationic organic and inorganic pollutants from water. A chemical activation of the elutri!ithe, resulting from reactions with metal salts and gibbsite or boehmite at high temperature 700~ in the absence of oxygen, composite adsorbents were developed for an efficient treatment of waste waters. The activated products were characterised in terms of surface area, and micropore volume and evaluated in terms of affinity and capacity for a number of organic compounds. Relations between the organic compounds and their molecular size, pKa, pH and adsorption temperature were investigated. The presence of gibbsite or boehmite during the modification process gives rise to a composite adsorbent suitable for the removal of anionic and cationic organic or inorganic compounds from waste water. Preliminary tests in pre-pilot installations confirmed the results obtained on lab scale. The preparation of the new composite adsorbent is cheap with an ease of handling. From comparative experiments, it was obvious that the modified elutrilithe adsorbents show a superior affinity and capacity compared to the classical excisting sorbents in the removal of pollutants from waste water.

1.

INTRODUCTION

Waste water, containing traces of organic compounds, causes a great problem in the environmental science. Organic compounds, classified as priority pollutants to remove are phenols [1]. Phenolic compounds are toxic to soil microorganisms at parts-per-million level [2]. The fate of phenols in the environment and their removal from aqueous media is complicated by their low solubility,

382 ability to ionise, low vapour pressure and tendency to undergo oxidation and oxidative polymerisation with humic and fulvic acid-type products [3]. The adsorption technique is one of the alternative treatment processes currently in consideration for achieving the required level in the removal of phenols. It is well known that activated carbon is the most significant adsorbent and widely applied commercial adsorbent in water purification. The removal of trace levels of organic pollutants from extreme large solution volumes is a microseparation process which requires cost-effective adsorbents. In recent years, several alumino-silicates were explored as adsorbents in the treatment of waste water [4-11]. An optimal adsorbent for the removal of organic compounds in waste water should have the following properties: low cost, ease of handling, environmental neutrality, high affinity and high capacity. Among many types of materials, elutrilithe, a natural weathered alumino-silicate containing about 8% of carbon by weight from coal mining heaps, is of potential interest. The property of this type of adsorbent is influenced by a number of factors including the carbon content [12,13] and surface area [14,15]. However, elutrilithe in its natural form has neither a high affinity nor a high capacity for the adsorption of hydrophobic compounds. In spite of its high carbon content, elutrilithe has only a surface area of 10 m2/g, reflecting its low adsorption capacity as reported by Dehon et al. [17]. Only 15-20% of organic materials could be removed from an aqueous solution. The objective of the present work was to explore the possibility of using modified elutrilithe for the removal of organic pollutants, especially phenol compounds. A method to increase the surface area was developed, based on a thermal shock method in order to form perlite. In this way, the surface area of elutrilithe increases from 10 to 32 m2/g [16]. Furthermore, a modification of elutrilithe using inorganic salts at high temperature resulted in a further increase of the surface area, and the affinity for organic compounds. Indeed, these modified elutrilithes could adsorb similar quantities of organics from aqueous solutions compared to activated carbons. Moreover, the modification procedure was adapted to remove efficiently ionic (cationic and anionic) inorganic and organic pollutants from waste waters.

2.

EXPERIMENTAL

2.1. M a t e r i a l s A typical chemical composition of elutrilithe is shown in Table 1.The inorganic salts, CaC12 9 2H20 and ZrC12, were supplied by MERCK Company. All organic compounds used in this study were obtained from ALDRICH. Some important physicochemical properties of the phenol compounds are summarised in Table 2.

383 Table 1 Typical chemical composition of elutrilithe (%wt) [16,18] SiO2

50

Na20

0.3

Fe203

6.5

KO2

4.0

A120~

25

SO~

0.7

MgO

1.9

TiO2

0.8

CaO

0.6

P20~

1.2

MnO

0.1

Carbon

+8.0

+10 m2/g

BET surface area Available form

black powder

Table 2 Physicochemical properties of the phenols Compound

Wavelength of Water solubility* adsorption bend (nm) (g/kg, 20~

pKa*

(20oc)

Phenol

269

77.5

9.98

3 -chlorophenol

274

25.0

9.08

3,5 -dichloro phe no 1

277

4.5

8.15

2,4,6 -trichlorophenol

293

0.8

6.00

*:Freier(1976), Aqueous Solution Data for Inorganic and Organic Compounds, Vol. 1

2.2. P r e p a r a t i o n a n d c h a r a c t e r i s a t i o n of t h e a d s o r b e n t s Modified adsorbents were prepared by treating the crude elutrilithe with CaC12 and ZnC12. 10 grams of the crude sample was blended homogeneously with 0.1 g CaC12 or ZnC12 in a crucible and heated in the absence of air in an oven at 700~ for 3 hours. The modified elutrilithe samples were cooled to room temperature and washed at 70~ several times with de-ionised water in order to remove excess salt. The prepared samples were dried and stored. Similar modifications were carried out with various inorganic salts such as KC1, NaC1, MgC12, ZnC12, etc. An alternative modification, a combination of metal salts with gibbsite and boehmite causes important features for the removal of neutral and ionic compounds from waste waters. The characterisation of natural and modified elutrilithes was carried out by a N2-adsorption technique, using a DigiSorb 2600 (Micromeritics Instrument

384 Corporation), in order to determine the surface area and pore volume. The elutrilithe samples were out-gassed at 150~ for 3 hours, and the nitrogen isotherms were recorded at liquid nitrogen temperature. The surface areas were calculated using the BET equation, and pore volumes were estimated at a relative pressure of 0.99. Surface areas calculated from the Langmuir equation were also reported if the adsorption data correlated better with this equation. External surface areas and pore volumes were estimated using as-plots [19]. The surface area in micropores was calculated by subtracting the external surface area from the BET surface area. 2.3. A d s o r p t i o n p r o c e s s Adsorption isotherms were determined using the batch equilibration technique. Initial concentrations of adsorbates were prepared in the range 0.10 1.00 mmol/1. A series of 100 -ml erlenmeyers containing 0.05 g elutrilithe and 100 ml solution were sealed and shaked at room temperature until equilibrium was obtained. Afterwards, the adsorbents were removed by centrifugation at 4500 rpm for 15 min. Measurements of the adsorbate concentrations were carried out by a direct ultraviolet adsorbance (UVIKON 930 instrument) method, using calibration curves.

3.

R E S U L T S AND D I S C U S S I O N

3.1. C h a r a c t e r i s a t i o n of t h e a d s o r b e n t s Natural elutrilithe has not a large surface area (+10m 2) nor a high affinity for organic compounds, in spite of its high carbon content (Table 1). In order to increase its affinity and adsorption capacity, the surface properties of the elutrilithe were modified. The specific surface area and the carbon content are very important factors influencing the adsorption. To increase the surface area of elutrilithe, as reported by M. De Boodt [16], a thermal shock procedure was used. In the present work, however, the elutrilithe was treated with inorganic salts such as CaC12 and ZnC12 at high temperatures. Indeed, the surface properties of carbonaceous materials can be affected by the presence of the metallic chlorides during the carbonisation and activation [20]. As expected, a significant change in the surface properties and micropores in elutrilithes was observed by these modification processes. Complete adsorption-desorption isotherms were recorded for untreated and treated elutrilithe samples. Figure 1 shows that the nitrogen isotherm a t - 1 9 6 ~ of untreated elutrilithe was a type-II isotherm in the classification of Brunauer, Deming and Teller. The isotherms recorded for the modified samples are different. Both types of the adsorption isotherms and adsorption affinity were influenced by the modification process. The isotherms of the modified products appear to be a combination of the type-II isotherm of the untreated product, and a type-I isotherm resulting from the adsorption in the micropores, introduced by

385 the modification process. To analyse further the adsorption m e a s u r e m e n t s of the modified elutrilithes, external surface areas, surface areas and pore volumes in micropores were e s t i m a t e d using as-plots (Fig. 1). In Table 3, the surface areas and pore volumes are summarised. The surface areas and pore volumes of the modified products are significantly higher compared to those of the initial products. The adsorption data correlate better with the L a n g m u i r equation t h a n with the BET equation. Both BET and L a n g m u i r equations were used to calculate the surface areas of the adsorbents. As a result, the surface effects in modified elutrilithes for organic compounds increase considerably. h~

60 50

Zn-Elut

---.

o~ 4o

30 9

20 9

Ca-Elut

S

10

< 0

0,2

0,4

0,6

0,8

1

Relative pressure (P/Po) Figure 1. Nitrogen adsorption isotherms at- 196~ for crude-Elm and the modified elutrilithes: Ca-Elm and Zn-Elut.

Table 3 Surface area and pore volume of elutrilithes Sample ID

Surface area (m2/g) BET

Pore volume (cm3/g)

Langmuir

External

Micropore

Total

Micropore

Crude-Elut

9.7

--

9.7

--

0.0289

Ca-Elut

59.3

83.0

6.5

52.8

0.0449

0.0248

Zn-Elut

157.3

219.9

4.7

152.6

0.0855

0.0726

386

3.2. Adsorption i s o t h e r m s Figures 2 show adsorption isotherms of phenol (p), 3-chlorophenol (3-CP), 3-5-dichlorophenol (3,5 DCP) and 2,4,6-trichlorophenol (2,4,6 TCP) from water onto unmodified and two modified Ca-Elut and Zn-Elut. The n a t u r a l elutrilithe

0,4 0,35 9

E~ 0,3 0,25 ,.Q 0

0,2 0,15

=

=

0,1

<

0,05

0

Phenol

0

0,2

0,4

0,6

0,8

Equilibrium concentration, mmol/1

0,4 0,35 9

E~ 0,3 0,25 or~

0,2 0,15

o

0,1

<

0,05

3-CP

--

0

~

---

---

0,2

---

0,4

---

~.

.It

_.-

0,6

0,8

1

Equilibrium concentration, mmol/1 Figure 2(A). Adsorption isotherms for phenol and 3-CP on crude-Elm (A), modified elutrilithes Ca-Elut(ll ) and Zn -Elm (+) at room temperature. Equilibrium time = 4 h: Initial concentrations = 0.10-1.00 mM; Adsorbent/solution = 0.05 g/100 ml.

387 0,45 0

0,4

E! 0,35

El

4

t..,

o

o,3 0,25 0,2 0,15

= 0

3,5-DCP

0,1

< 0,05 o 0,2

0,4

0,6

0,8

Equilibrium concentration, mmol/l

bs 0

0,6 0,5 0,4

0

0,3 0,2

0

0,1

< 0,2

0,4

0,6

0,8

1

Equilibrium concentration, mmol/l Figure 2 (B). Adsorption isotherms for 3,5-DCP and 2,4,6-TCP on crude-Elm (A), modified elutrilithes Ca-Elut(ll ) and Zn-Elut(+) at room temperature. Equilibrium time = 4 h; Initial concentrations = 0.10-1.00 mM; Adsorbent/solution = 0.05 g/100 ml.

does not show any detectable adsorption of phenols. However the modified elutrilithes adsorb a considerable amount of phenols and reveal a high affinity. The modification procedure has significantly changed the surface properties from hydrophilicity to organophilicity. Also a significant difference in the adsorption capacity of Ca-Elut and Zn-Elut for various phenols can be observed. The adsorption increases in the order: phenol
388 their water solubilities (P > 3 CP > 3,5 DCP > 2,4,6 TCP) (Table 2). An inverse relationship is observed between the water solubility and the adsorption on the surface of the modified elutrilithes. All the adsorption isotherms on the modified elutrilithes are dominantly a classical type-I curve, obeying the Langmuir equation:

Q~ e qe = ~ l+kCe

(I)

where qe is the amount of solute adsorbed per unit weight of adsorbent (mmol/g), Q~ the solid-phase concentration corresponding to all available sites being filled, i.e., the maximum adsorption capacity, Ce, the liquid-phase concentration at equilibrium, and k, a constant related to the energy of adsorption. Equation (1) can be written in a linear form. 1

1

1

qe

Q0

kQ~ Ce

(2)

In the double reciprocal plots, the amounts of phenols adsorbed by the solid phase versus the amount remaining in solution at equilibrium, give the intercept (1/Q ~ and a slope (1/kQ~ The maximum adsorption capacities and k constants from these Langmuir plots are given in Table 4. The adsorption sequence of the different phenols is reflected by the k and Qo values. Zn-Elut appears to be superior in the phenol adsorption compared to

Table 4 Summary of the Langmuir plot coefficients Adsorbate

Ca-Elut Qo k (mmol/g) (1/mmol

Zn-Elut Qo k (mmol/g) (l/retool)

Phenol

0.157

20.58

0.388

15.73

3-CP

0.173

29.05

0.391

20.63

3,5-DCP

0.190

40.38

0.436

40.76

2,4,6-TCP

0.192

54.37

0.565

41.16

3-CP:3-chlorophenol, 3,5-DCP: 3,5-dichlorophenol; 2,4,6-trichlorophenol

389 Ca-Elut. Comparing the adsorption of the different phenols by Ca-Elut or ZnElut, the adsorption capacity Q~ and the constant k both are in the order: phenol < 3-CP < 3,5-DCP < 2,4,6-TCP, reflecting the hydrophobic properties of these molecules. This behaviour indicates that the surface of the modified elutrilithes has a higher affinity for hydrophobic compounds. The adsorption energy of a molecule is the difference between the free energy of reaction with the surface and the free energy of solvation [21]. The constant k, related to the energy of adsorption, shows that for the most hydrophobic molecule, 2,4,6-TCP, the adsorption on an organophillic Zn-Elut surface is favoured because both components of the adsorption free energy should be highly negative. On the other hand, phenol is a molecule with a small hydrophobic character, the desolvation process results in a positive free energy, lowering the net adsorption free energy. For all four phenols, the net adsorption free energy change favours the adsorption [8]. On the other hand, as expected, the substantially high specific surface area is also a very important factor in the adsorption. The higher adsorption capacities in Zn-Elut reflect the larger surface area compared to Ca-elut (Table 3). The phenol, 3-CP and 3,5 DCP adsorptions on modified clays (cetyl pyridinium cation exchanged montmorillonite and cetyltrimethyl-ammonium-montmorillonite) were studied by Mortland et al. [5]. They reported a type-V isotherms for 3-CP and 3,5 DCP. The adsorption capacities for the 3-CP and 3,5 DCP adsorption, however, were significantly lower compared to the Zn-elutrilithe. In contrast to the amount of phenol adsorbed by modified elutrilithes, phenol was not adsorbed at all from an aqueous solution by the modified clays. This demonstrates that the modified elutrilithes show important affinities and larger adsorption capacities compared to modified clays in the phenols adsorption. 3.3. I n f l u e n c e of t h e pH Several experiments were carried out to study the influence of the pH on the adsorption of phenols by Zn-Elut. The acidity constants (pKa values) for the different phenols are shown in Table 2. Experiments, with constant initial adsorbate and adsorbent concentrations, were set up. The dependence of the adsorption of phenols on the pH of the solution is presented in Figure 3. The adsorption increases slowly with increasing the pH value in the pH < pKa range. Maximum adsorption of all four phenols was observed at pH close to the pKa values, but decreases sharply at pH values greater than the pKa of the adsorbate. This suggests that the dissociation of the adsorbate influences the adsorption capacities. The variation in adsorption with the degree of dissociation is not totally unexpected since the dissociated form is more soluble in an aqueous solution [20,22]. The dissociated compounds will be less adsorbed because of the stronger adsorbate-solvent bonds affecting the affinity for the adsorbate. In general, the adsorption capacities were larger for the neutral phenol than for the phenolate form.

390 0,4 0,35 0

E~ 0,3 -~ 0,25 0 r.~

0,2 0,15

o

0,1

< 0,05

0

5

10

15

pH value

0,5 0,45 0,4 0

E~ 0,35 0,3 0 r.~

0,25 0,2

= 0,15

0

<

0,1 0,05

0

2

4

6

8

10

12

pH value Figure 3. Variation in the adsorption of phenols on Zn-Elut with pH at room temperature. +:phenol, I-! "3-CP, A'3,5-DCP, II :2,4,6-TCP. Initial concentrations = 0.50 mM; Adsorbent/solution ratio = 0.05 g/100 ml; Equilibrium time = 4 h.

391 3.4.

Influence

of temperature

Adsorption isotherms were determined for 3,5-DCP on Zn-Elut at temperature of 5, 21 and 30~ The results are shown in Figure 4 and illustrate that the adsorption of phenols is an exothermic process. The adsorption of phenols decreases with increasing temperature. However the temperature influence is small. From the variation in the Langmuir constant k with temperature, the enthalpy of adsorption (AH) can be calculated by the equation, k - ko exp(AH/RT) In k2 _ AH(T2 T1) kl RT2T1

(3)

with kl and k2 the Langmuir constants at adsorption temperatures T1 and T2. The resulting value of AH (--1.3 kcal/mol) is in the order of magnitude expected for a relatively weak physical adsorption on modified elutrilithe. In contrast, this value is much larger t h a n the values associated with chemisorption processes ( < - 1 0 kcal/mol).

0,45 9

0,4

9 m

9 m

A

m

m

mm

mm

9 m

T

T

T

7 ------'+

8

10

0,35 9

~

o,3

a~ 0,25 9 r.~

0,2 0,15

9

<

o,1 0,05

0

2

4

6

Equilibrium concentration, mmol/1 Figure 4. Influence of temperature on the 3,5-DCP adsorption by modified elutrilithe Zn-Elut. A:5~ m. 21~ +: 30~ Initial concentrations = 0.10- 1.00mM; Adsorbent/solution ratio = 0.05 g/100 ml; Equilibrium time = 4 h.

392

3.5. Adsorption-desorption isotherms Adsorption and desorption isotherms were determined for 3,5-DCP on Zn-Elut. The adsorption isotherm was carried out in the usual fashion, while the desorption isotherm was obtained by decanting the s u p e r n a t a n t from each vessel at the end of the adsorption part of the experiment, and then reinserting 100 ml of deionized water into each flask. The reaction flask were then resealed and allowed to re-equilibrate for a period of seven days. At the end of this period, the concentration of adsorbate was determined in each flask. These equilibrium concentrations in conjunction with corresponding adsorption phase data were used to compute the desorption isotherm. The results of this experiment are presented in Figure 5. A significant hysteresis is observed. This demonstrates that the 3,5-DCP adsorption on Zn-Elut is a reversible process with important diffusion effects.

0,45 0,4

.

b.0

'~9 E~

035 , 0,3

a~ 0,25 t~

o (D

0,2 0,15

o

<~

o,1 i 0,05

0

0,2

0,4

0,6

0,8

1

Equilibrium concentration, mmol/1 - - ~ - - A d s o r p t i o n ----!1~ Desorption

Figure 5. Adsorption-desorption isotherms for 3,5-DCP on Zn-Elut at room temperature. Initial 3,5-DCP concentrations = 0.10 - 1.00 mM; Adsorbent/solution ratio = 0.05 g/100 ml; Adsorption equilibrium time = 4 h; Desorption equilibrium time = 7 h.

393

3.6. A d s o r p t i o n of other organic c o m p o u n d s The adsorptive properties of modified elutrilithes for other organic compounds are presented in Figure 6. It is obvious from this figure t h a t the Zn-Elut, exhibits a significant selective adsorption. F u r t h e r m o r e Zn-Elut has a very high affinity for aromatic compounds, chlorobenzenes, toluene en chlorotoluene.

4,5 4 E~ 3,5 3 9

2,5

2 9 =

1,5

<

1 0,5 1

i]

4

3

2

5

7

6

8

9

10

Component number 5 4,5 b~

4

3,5 3 O

2,5 9 Elutdlithe

2 =

[] Zn-Elut

1,5

0

E~ <

1 0,5 0 11

12

13

14

15

16

17

18

19

Component number Figure 6. Adsorption of organic compotmds by unmodified and modified Zn-Elut adsorbents. Dosage = 0.5 g/100 ml, temperature = 24 ~ C, equilibrium time = 4 hr each adsorbate initial concentration = 50 ppm. Organic components: 1. diethylether, 2.1-propanol, 3.2-butanol, 4. chloroform, 5. dichloroethane, 6. benzene, 7.3-pentanol, 8. propylacetate, 9.3,3-dimethyl2-butanol, 10. toluene, 11.2,2-dimethyl-3-pentanol, 12.2-chlorobenzene, 13.1,1,2,2-tetrachloroethane, 14.2-chlorotoluene, 15.2-chlorophenol, 16.1,2-dichlorobenzene, 17.4-methylbenzaldehyde, 18.2,6-dimethylaniline, 19.1,2,4-trichlorobenzene.

394 There are many factors which influence both affinity and capacity of adsorbents from an aqueous solution. The molecular structure, or nature of the adsorbate, is in particular important in the adsorption process. In general, molecules with a low polarity, solubility and ionisability, tend to adsorb preferentially, in these modified elutrilithe.

3.7. Other important properties by adapting the modification procedure If the modification of the crude elutrilithe is carried out in combination with gibbsite or boehmite the new modified composite adsorbent reveals a significant ionic exchange. Indeed, crude elutrilithe shows only a low cationic exchange capacity (10 meq/100 gr) . After this modification procedure followed by a basic treatment total cation exchange capacities of 100 meq/100 gr can be observed to remove cationic pollutants (NH4, heavy metals, etc.). On the other hand, an acidic treatment create total anionic exchange capacities up to 120 meg/100 gr, useful for the removal of anionic pollutants, such as NO~, PO34, C I , etc. The total ionic exchange capacity (cationic or anionic) of this modified elutrilithe can be controlled by the quantity of gibbsite or boehmite and type of metal salts during the modification process. Efficient removals of NO~, PO34, SO~ and CI from waste waters were proven. Also, the origin of the elutrilithe reveals no significant differences in the removal of pollutants from waste water. All modified composite samples obtaining from the UK, Poland, Russia, South-Africa and Canada show a simular removal efficiency of neutral, anionic and cationic organic and inorganic pollutants from waste waters.

3.8. Water purification tests in pre-pilot installations The removal of organic and inorganic pollutants from waste water using the Zn-elut adsorbent in a pre-pilot installation with downflow fixed-bed columns or in a batch suspension method, confirmed the results obtained on labscale. The average capacities of the pre-pilot installations were 0.5 m 3 per hour.

3.9.

Regeneration

Because of economic and solid waste disposal considerations, it is feasible to regenerate spent modified elutrilithe for subsequent re-use than to dispose of it. In general, experiments have shown that with modified elutrilithe (Ca-Elut, ZnElut) and classical activated carbons, identical regeneration processes can be used to remove the previously adsorbed materials and to re-institute its ability to adsorb impurities. A thermal regeneration process at 700~ in N2 with limited quantities of oxidising gases, using a rotary-tube furnace was sufficient. However, a new basic or acidic treatment was necessary re-install the cationic or anionic exchange capacities.

395 4.

CONCLUSIONS

A new composite adsorbent was developed for the removal of neutral, anionic and cationic organic or inorganic pollutants from waste waters. From comparative experiments, it was obvious that the modified elutrilithe adsorbents show a superior affinity and capacity compared to the classical sorbents in the removal of pollutants from waste waters. Furthermore, important anionic and cationic exchange behaviour could be generated to remove efficiently charged organic or inorganic compounds from water. Moreover, the preparation of the new composite adsorbent is inexpensive with an ease of handling and with in improved performance in the waste water treatment compared to existent adsorbants.

REFERENCES 1. M. Kuwahara, N. Shindo and K. Munakata, J. Agric. Chem. Soc. Jpn., 44 (I 979) 169. 2. P.M. Chapman, G.P. Romberg and G.A. Vigers, J. Water Pollution Control Federation, 54, 292. 3. T.D. Thompson and W.F. Moll, Clays and Clay Minerals, 21 (1973) 337. 4. W. Karickhoff, S. Brown and A. Scott, Water Research, 13 (1979) 241. 5. M. Mortland, S. Sun and A. Boyd, Clays and Clay Minerals, 34, 5 (1986) 581. 6. C. Means, G. Wood, J. Hassett and L. Banwart, Environ. Sci. & Tech., 14, 12 (1989) 1524. 7. T. Nolan, R. Srinivasan and H. Fogler, Clays and Clay Minerals, 37, 5 (1989) 487. 8. Keeran R. Srinivasan and H. Scott Fogler, Clays and Clay Minerals, 38, 3 (1990) 287. 9. C. Zielke and J. Pinnavaia, Clays and Clay Minerals, 36, 5 (1988) 403. i0. C. Voice and J. Weber Jr., Water Research, 17, I0 (1983) 1433. 1 I. J. Weber Jr. C. Voice, M. Pirbazari, E. Hunt and M. Ulanoff, Water Research, 17, I0 (1983) 1443. 12. W.C. Steen, D.F. Paris and G.L. Baughman, Water Research, 12 (1978) 665. 13. M.C. Lee, R.A. Miller and E.S.K. Ghian, J. Envir. Sci. Hlth., A14(5) (1979) 415. 14. R. Haque, D.W. Schmedding and V. Freed, Envir. Sci. Technol., 8 (1974) 139. 15. Y. Hiraizumi, M. Takahashi and H. Nishimura, Envir. Sci. Technol., 13 (1979) 580. 16. M. De Boodt, Report: Elutrilithe, Etude des PropriJtJs Physiques (Laboratoire de physique du sol, R.U.G. Gent, Belgium). 17. C. Dehon, P. Bouviez, D. Demeulder et al., Report: PropriJtJs PhysicoChimique de l'Elutrilithe, ObservJes sur des Lisiers, des Eaux UsJes, des Gaz (Institut provincial d'hygiene et de bacteriologie, Mons, Belgium) (1985).

396 18. RYAN EUROPE, Report: Characteristiques moyennes de l'elutrilithe, Charleroi, Belgium (1988). 19. S.J. Gregg and K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, London, (1982) 90. 20. J.W. Hassler, Activated Carbon, Chemical Publishing Co., New York (1963). 21. N.A. Klimenko, A.A. Permilovskaya and A.M. Koganovskii, Kolloid Zh., 36 (1974) 788. 22. J.S. Mattson and Jr. H.B. Mark, Activated Carbon Marcel Dekker, Inc., New York (1971).

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

Carbon materials

as adsorbents

397

for v a p o u r p o l l u t a n t s

F. J. L6pez-Garz6n, I. Fernandez-Morales, C. Moreno-Castilla and M. Domingo-Garcia Grupo de Investigaci6n en Carbones, Dpto. de Quimica Inorg~nica, Fac. de Ciencias, Universidad de Granada, 18071 Granada, Spain. 1. I N T R O D U C T I O N Adsorption on porous materials has been proved to be an efficient technique for use in a wide range of very i m p o r t a n t applications ranging from the elimination of e n v i r o n m e n t a l pollutants to the concentration or storage of products of industrial interest. Within the large n u m b e r of possible adsorbents, carbon materials have proved to be the best for these applications because of their high capacity for adsorption. This is directly related to their textural and chemical properties (i.e. surface porosity and chemical groups on the surface) and because of the ease of changing these properties by t h e r m a l or chemical t r e a t m e n t s [1-3]. Another r e m a r k a b l e aspect is the variety of raw materials from which these adsorbents can be obtained. The yield and the selectivity of a particular adsorption process is determined by the n a t u r e of both the adsorbent and the adsorbate. Therefore the kind of interaction t h a t takes place between the molecules of the substance to be adsorbed and the surface of the adsorbent is very i m p o r t a n t [4-5]. There are mainly two types of interactions to be considered: the first one are those produced by London dispersion forces (non-specific interactions) and the second by electrostatic forces through dipole-dipole interactions, hydrogen bonding...etc (specific interactions). Therefore, in the adsorption on porous materials two main factors m u s t be t a k e n into account: the porosity and the chemical functionalities of the adsorbate and adsorbent [6]. The porosity can determine the adsorption depending on the shape and size of the molecules in such a way t h a t two molecules with only small differences can be discriminated, i.e. the adsorbent can acts as molecular sieve. Therefore the adsorption process can be designed to eliminate a pollutant or to selectively recover a substance [7-10]. Porous carbon materials usually contain polymodal distribution of pores the sizes of which vary from molecular dimension to several h u n d r e d nanometers, but they can be tailored with high capacity of adsorption for a particular adsorbate by an appropriate selection of raw materials, p r e p a r a t i o n process and specific t r e a t m e n t s [11-13].

398 Lignocelulosic materials have long been used to obtained activated carbons with high adsorption capacity. In the present chapter activated carbons obtained from agricultural by-products (olive stones, H carbons, and almond shells, C carbons) are analysed in relation to the possible application as eliminator of pollutants. To gain a more complete knowledge of the parameters that can influence on the surface nature of these porous adsorbents, the raw materials were submitted to different methods of carbonization and activation periods [14,15]. Some of the activated carbons used as adsorbents were obtained by carbonization and simultaneous activation, in a flow of CO2, of almond shells (Cseries) and olive stones (H-series) at 1123 K and for different time periods ranging from 2 to 30 hours. Other series were obtained by carbonization in N2 and further activation with CO2 (HA-series) or steam (HW-series). When activated carbons with homogeneous microporosity are required the use of organic copolymers as starting material gives good results. Depending on the experimental conditions and the type of organic polymer used it is possible to obtain carbon materials that range from those with high porosity but very narrow pore size distribution, showing molecular sieve character, to those with very low porosity. For this purpose organic copolymer can be carbonized at different temperatures in an inert atmosphere [16,17] or they can be prepared using several recipes for copolymerization before the thermal treatment [18-21]. The copolymer of vinylidene and vinyl chloride, commercially known as Saran, was carbonized in nitrogen at temperatures ranging from 973 to 1573 K to obtain adsorbents with different average micropore dimensions depending on the temperature. Molecular sieve character is often found for samples prepared at highest temperatures [10]. The other type of carbon material studied prepared by copolymerization of organic molecules are the glassy carbons. These were produced by polymerization of furfuryl alcohol, using different recipes [13], and further carbonization. Alternatively, the porous network of the adsorbents can be modified by chemical treatments in such a way that the introduction of chemical groups in their structure can change the capacity and selectivity for adsorption. In this sense activated carbons obtained from Saran were treated with HNO3 and CS2 to introduce oxygen and sulphur chemical groups. Some of these carbons showed a molecular sieve character for specific adsorbates as a consequence of the fixation of the chemical groups at the entrance of the micropores [22]. As mentioned earlier, when the molecules to be adsorbed have chemical functionalities the specific interactions become very important in the adsorption process. In this sense the introduction of chemical groups on the surface of the adsorbents can also have a very important roll in the adsorption process because of electrostatic or chemical interactions [23-25]. Besides the treatment with HNO3 and CS2 already mentioned, some of the carbons were also treated with H202 and KI to test the elimination of pollutants with high dipole moments or with chemical functions. To better understand the importance of both specific and non-specific interactions in a determined adsorption process, adsorption on

399 non-porous carbons materials has also been tested: a graphitized black (V3G) and three graphites (Pyrolitic, Acheson and Degussa) [26,27]. The adsorption of a wide number of adsorbates of very different kind has been studied. Among the non-polar molecules without chemical functionalities a series of hydrocarbons of different size and shape (linear, cyclic and branched) were chosen: n-alkanes (from n-C4 to n-Cg), benzene, cyclohexane and 2,2-dimethyl butane (2,2-DMB). With these molecules the importance of the average dimension and morphology of the micropores for the adsorption of a particular pollutant is discussed. Substances with different functionalities were used to know the extent in which the specific interactions can improve the adsorption. These substances are: acetone, diethyl ether, tetrahydrofurane (THF), carbon tetrachloride, chloroform, dichloromethane, methyl iodide and n-alcohols [28]. Finally, the study of the adsorption processes was carried out in two different regimes. On the one hand in static conditions, using a gravimetric adsorption system, which gives very valuable information on the adsorption properties of the samples in the range of relative pressures between 10 .4 and 1. On the other hand in dynamic conditions using Inverse Gas Solid Chromatography (IGSC) [29]. From the practical point of view one should take into account that in most real cases the amount of substance in the gas phase to be eliminated or concentrated is very low and frequently the temperature is above room temperature. In addition these molecules are in flue gases, such that the adsorption has to be carried out under dynamic conditions. These experimental requirements are easily achieved by IGSC technique.

2. GRAVIMETRIC M E A S U R E M E N T S 2.1. A c t i v a t e d c a r b o n s from l i g n o c e l u l o s i c origin The data reported in this section deal with activated carbons prepared by carbonization and simultaneous activation (in a flow of CO2) of almond shells (C-series) and olive stones (H-series) at a temperature of 1123 K with different treatment periods [14]. The adsorption of benzene, n-hexane, cyclohexane and 2.2-dimethyl butane (2,2-DMB) was studied on these carbons at 303 K. This includes the kinetics and the static adsorption isotherms at the same temperature and P/Po = 0.6. Diffusion parameters were obtained under unsteady state, for small values of time t using the equation: V -V 0 V e --V 0

6

/Dt/1/2

~1/2 ~ r2 )

(1)

Where v is the amount adsorbed at time t, Vo is the amount adsorbed at time t = 0, Ve is the amount adsorbed at equilibrium, D is the diffusion coefficient and ro is the length of diffusion path. From the kinetics measurements, adsorption rates, RL, and diffusion parameters, D1/2/ro were obtained [30] and these are compiled in Table 1.

400

Table 1 Apparent adsorption rate, RL, and hydrocarbons on activated carbons

diffusion

parameters

Dl/2/ro of the

D 1/2/ro (rain-1/2)

RL (cm 3 min "1/2) Adsorbates

H-2

H-13

H24

H30

H-2

H-13

H-24

H-30

n-hexane

0.24

0.41

0.65

0.86

0.25

0.21

0.30

0.43

benzene

0.24

0.54

0.66

0.57

0.23

0.34

0.30

0.31

cyclohexane

0.27

0.54

0.59

0.19

0.27

0.34

2,2-DMB

0.19

0.59

0.70

0.19

0.27

0.36

C-2

C-13

C-24

C-30

C-2

C-13

C-24

C-30

n-hexane

0.46

0.79

1.00

1.00

0.30

0.56

0.54

0.49

benzene

0.45

0.57

0.66

0.62

0.31

0.36

0.33

0.32

0.44

0.62

0.85

0.37

0.34

0.48

0.23

0.47

0.45

cyclohexane

2,2-DMB 0.10 0.20 0.79 0.74 Reprinted from: M. Domingo-Garcia et al. [14].

0.21

An increase in the diffusion parameters, D1/2/ro, and the adsorption rate, RE, with the time of treatment was apparent for all the adsorbates, which means less kinetic restrictions for the adsorption on the most activated samples. Adsorption of 2,2-DMB and cyclohexane on H-2 and C-2 samples were so small that the kinetic parameters are meaningless. This suggests a discrimination capacity (or molecular sieve behaviour) for molecules with a mean size of 0.56 nm. Adsorption isotherms of benzene and 2,2-DMB on both series of samples (C and H) are shown in Figures 1 and 2. An increase in the adsorption capacity with activation time of carbons is observed, although in general, beyond 24 hours of activation treatment this increase is quite small. Otherwise, except for 2,2-DMB and cyclohexane on H-2 and C-2, and 2,2-DMB on C-13 the rest of the isotherms can be assigned to type I in the BDDT classification which corresponds to the adsorption in a predominant microporous adsorbent, typical among the majority of activated carbons. The exceptions mentioned above give isotherms indicating a low adsorption capacity for these systems and, from their shapes, they could be assigned to a type IV character. This could mean that the little adsorption of these two hydrocarbons took place chiefly on the mesopores or external surface of these adsorbents [1,31]. This difference in the adsorption capacity of the several adsorptives shown by the less activated carbons supports the molecular sieve character above mentioned.

401 0,8 a

,~ 0,6 eLO

g >

0,8

C-30

C-24 .- ~

-

C-24 0,4

,~ 0,6

C-13

:

"

C-30

0,4

C-2 >

0,2 0,0 0

i

i

i

i

i

0,2

0,4

0,6

0,8

l

C-13

0,0

0

0,2

,

,

,

0,4

0,6

0,8

P/Po

P/Po

Figure 1. Adsorption isotherms of (a) benzene and (b) 2,2-DMB on C-series.

0,8

a

H-30

O,8 ---- H-24

0,6 ~

~

~3

H-13

"~ 0,4 > 0,2

b

H-30

0,6 ~

H

-

2

4

0,4 H-13

H-2

0,2 H-2

0

!

i

i

i

i

0,2

0,4

0,6

0,8

1

P/Po

0

ll'--ll"--r

0

t

0,2

~

I ~ -

0,4

I

I

0,6

0,8

P/Po

Figure 2. Adsorption isotherms of (a) benzene and (b) 2,2-DMB on H-series.

The analysis of the isotherms gives information not only on the adsorption capacities of the carbons but also on the porous structure which is responsible for the behaviour of the adsorbents. Among the several approaches that can be used to analyse the adsorption isotherms, the Dubinin-Radushkevich [32,33] (DR) theory of volume filling of micropores (equation 2), was chosen and the results were compared with those obtained by applying the linear form of the Langmuir isotherm [34]. The DR equation reads:

V=V 0exp-

1 P/ 21

E0~X P 0

(2)

Where Vo is the total volume of micropores, Eo is the characteristic energy of adsorption and ~ is the "similarity coefficient" which depends on the adsorptive. Plots of the DR equation are represented in Figures 3 and 4 for all the hydrocarbons on H-13 and H-30. In general, the linearity of these representations extends up to P/Po = 0.25 except for 2,2-DMB and cyclohexane on H-2 and C-2,

402

-0,1

-0,2 -0,3

-

9

C6H14 C6H6

> -o,4

~

2,2-DMB

-0,5

C6HI2

C6H6

-o,2 ~"

~ ~

a

~

-"

-

~

-0,3

C6H14 ~ " ~

2,2-DMB

"~"---n-----~~..

C6Hi2

-0,6 0

, 1

Ig2 P/Po

-0,4

, 2

0

,

,

1 lg2 P/Po

2

Figure 4. DR plots for H-30.

Figure 3. DR plots for H-13.

which display a very pronounced u p w a r d t u r n before this relative p r e s s u r e was attained. The total volume adsorbed of each adsorbate, Vo, is obtained from e x t r a p o l a t i o n of the l i n e a r r a n g e of these plots. Table 2 shows the Vo values calculated a n d the ratio of the a m o u n t adsorbed at a relative p r e s s u r e of 0.1, V0.1, to t h a t at the s a t u r a t i o n relative pressure, Vs, V01/Vs . It is a p p a r e n t that, with the exception above mentioned, the adsorption process is almost completed at low relative

Table 2 Volumes of adsorption obtained from DR r e p r e s e n t a t i o n s Vo (cm3/g)

Vo.JVs x l 0 0

adsorbates

H-2

H-13

H-24

H-30

H-2

H-13

H-24

H-30

n-hexane

0.30

0.48

0.70

0.66

83

93

93

95

benzene

0.32

0.47

0.63

0.70

78

95

91

90

cyclohexane

0.06

0.43

0.58

0.56

40

93

93

91

2,2-DMB

0.02

0.37

0.57

0.62

20

73

88

92

C-2

C-13

C-24

C-30

C-2

C-13

C-24

C-30

n-hexane

0.40

0.43

0.61

0.56

82

95

95

96

benzene

0.37

0.42

0.52

0.60

85

93

89

91

cyclohexane

0.08

0.37

0.52

0.56

59

94

90

92

2,2-DMB

0.09

0.17

0.62

0.59

43

79

89

96

Reprinted from M. Domingo-Garcia et al. [14].

403 pressures, which indicates t h a t the adsorption is not allowed to develop to many multilayers at higher relative pressures [35] and t h a t the adsorbents have reasonably narrow pore systems. The Vo/V'o ratios for n-hexane/2,2-DMB and benzene/2,2-DMB couples (which have the largest size difference) are 15 and 16 in H-2 sample, and are very close to unity in the rest of the series. For C-series the ratios for C-2 are smaller and also very close to unity in the rest of the series. This means t h a t the discrimination capacity of these carbon samples for these molecules is low, although as explained above, there is some molecular sieve behaviour at low degree of t r e a t m e n t (H-2 and C-2). U p w a r d deviations from linearity of DR plots are generally interpreted as an "additional" adsorption capacity of the adsorbates on supermicropores and small mesopores or multilayer formation on non porous surfaces [35,36]. DR plots for the adsorption of 2,2-DMB on the H-series are represented in Figure 5.

0,0 -

C-30

-0,4 -

C-24

> -0,8

," o--_..__.__~_.~____.__._____.

- 1,2

o~~___....._.~_...,~

-1,6

, 1

0

C-13

-~ C-2 , 2

" ", 3

lg? P/Po Figure 5. DR plots for the adsorption of 2,2-DMB on C-series.

It is a p p a r e n t t h a t with the increased activation time, not only a dramatic rise in the adsorption capacity is produced but also the range of linearity becomes greater and the u p w a r d deviation at high relative pressures is clearly reduced. This behaviour supports the hypothesis t h a t the adsorption of the largest molecules on carbons H-2 and C-2 mainly takes place on the external surface with multilayer formation at high relative pressures. This means that the micropore volume of these two carbons is not accessible to molecules with dimensions of cyclohexane or larger. The small molecular sieve behaviour of these two carbons and the disappearance of this property with the increase of activation time are better demonstrated by plotting the micropore volume, Vo, obtained for each carbon sample with every molecule against the m e a n molecular size of these adsorbates (Figure 6) [8,14]. From this representation it is also possible to conclude t h a t after 24 hours of activation t r e a t m e n t there is no clear improvement in the adsorption capacity on either the H or C carbons. Characteristic adsorption energies for every adsorbent, Eo, derived from the slope of the DR representations of the benzene isotherms (which is generally

404 0,8

0,8 C-24

0,6 []

m

,~0,6

H-24

fi 0,4 o

or-

~

,

,

0,45

0,5

C-13

~0,2

, ~ ~ 0,55

C-30

O

0,2

0,4

2-

E0,4

H-13

0 0,6

d (nm)

0,4

0,45

0,5 d (nm)

0,55

0,6

Figure 6. Volume adsorbed V0, obtained from DR equation, versus the mean molecular size of the adsorbates with six carbon atoms (see Table 16 for molecular dimensions).

taken as s t a n d a r d adsorbate), and the characteristic dimension of the micropores Lo [37] are compiled in Table 3.

Table 3 Characteristic adsorption energy, Eo, and characteristic dimension of the micropores, Lo Adsorbents

Eo (kJ/mol)

Lo (nm)

Adsorbents

Eo (kJ/mol)

Lo (nm)

H-2

16.60

0.78

C-2

21.68

0.60

H-13

21.40

0.62

C-13

22.06

0.58

H-24

21.71

0.59

C-24

18.22

0.71

C-30

22.80

0.57

H-30 22.44 0.58 Reprinted from: M. Domingo-Garcia et al. [ 14].

It is, therefore, apparent that the increase in activation time virtually does not modify the micropore size of the activated carbons because the values of the adsorption energy, Eo, are very similar when the activation time progresses. This could mean that the only effect of activation was to favour the access to smaller micropores. The molecular sieve behaviour shown by the less activated carbons is therefore attributed to constrictions at the entrance of the micropores caused by chemical functionalities linked to carbon atoms at the edges of the entrance [4]. After certain activation time (around 13 hours) the constrictions and, consequently, the discriminative behaviour was lost.

405 Application of the Langmuir model (P/V vs. P) to the type I isotherms gives linear plots in a wide range of pressures. VL/Vs ratios (VL obtained from the slope of the linear plots and Vs obtained from the plateau of the isotherms), are in most cases close to unity (Table 4) except for 2,2-DMB and cyclohexane on H-2 and C-2. Such results suggest that neither multilayer formation or capillary condensation occur on those systems and that the adsorbate-adsorbent interactions are far greater than the adsorbate-adsorbate ones.

Table 4 VL/Vs ratios. VL obtained from Langmuir equation. Vs obtained from DR equation VL/Vs

VL/V~

H-2

H-13

H-24

H-30

C-2

C-13

C-24

C-30

n-hexane

1.00

1.02

0.96

1.00

0.86

1.02

1.03

1.01

benzene

1.00

1.02

0.99

1.00

1.02

1.04

1.03

1.00

cyclohexane

0.56

1.02

0.98

1.00

0.67

1.00

1.03

1.00

2,2-DMB

0.42

0.97

0.98

1.04

0.57

0.90

0.95

1.00

Reprinted from: M. Domingo-Garcia et al. [ 14].

The VL values coincide very closely with the volumes, Vo, calculated by the DR equation (Table 2). It is, therefore, possible to conclude that the micropore size distribution of these carbons is in the range of the molecular size of the organic vapours adsorbed and little, if any, adsorption took place on the external surface of the adsorbents working at relative pressures, P/Po, below 0.8. Surface areas of the carbon materials were calculated using the VL values. These are considerably smaller than those obtained by CO2 measurements at 273 K for the less activated samples, H-2 and C-2, and for these carbons the area increases with a decrease in the mean molecular dimension of the adsorbate. Nevertheless, as the activation increases the areas for all the vapours and that of CO2 become quite similar. This supports the existence of constrictions at the entrance of the pores, earlier commented, hindering the access of the adsorbates. As shown previously agricultural by-products, such as olive stones and almond shells, are good raw materials to obtain carbonaceous adsorbents of organic vapours after the appropriate activation process with quite high adsorption capacities (around 0.6-0.7cm3/g), which is important for their potential application as pollutant removal adsorbents. For these samples, prepared by carbonization and simultaneous activation in C02, the process acts mainly by opening small micropores and eliminating chemical functionalities hindering the access of the adsorbates.

406 2.2.

Carbon

materials

from polyfurfuryl

alcohol

(Glassy

carbons)

The preparation of these carbon materials basically consisted of the polymerization and slow carbonization of furfuryl alcohol [13,21,38,39]. Some t e x t u r a l characteristics like pore volumes and a p p a r e n t surface areas are s u m m a r i z e d in Table 5. I m p o r t a n t differences in the pore size distribution in the region of meso, V2, and macropores, V3, are found for the samples. Moreover, SN2<<Sco2, which indicates t h a t the microporosity is very narrow or t h a t there are constrictions at the entrance of the micropores [40,41].

Table 5 Textural characteristics of glassy carbon Sample

SN2 (m2/g)

Sco2(m2/g)

V3 (cm3/g)

V2 (cm3/g)

Vl (cm3/g)

P1

475

408

0.214

0.033

0.119

P2

24

257

0.000

0.041

0.093

P3

398

344

0.024

0.235

0.163

P4

458

468

0.241

0.000

0.105

Reprinted from: M. Domingo-Garcia et al, [41 ].

The molecules used as adsorbates were n-hexane, benzene, cyclohexane and 2,2-DMB, as well as two polar adsorptives, methyl iodide and chloroform. The kinetics adsorption curves for benzene and methyl iodide on three of these carbons are given in Figure 7 (The behaviour of carbon P4 was very similar to P1). Values of adsorption rates, RL, and diffusion parameter, D1/2/ro, are given in Table 6.

0,20

0,20 P3

a

9 P3

0,15

0,15

o,,o

o,,o

> 0,05

0,05

~

b

~ P2

0,00 . 0

.

.

. . 0,00 T 200 300 400 0,8 1,2 1,6 0 100 t (min) t (min) Figure 7. Kinetics adsorption curves on glassy carbons: (a) benzene and (b) methyl iodide. 0,4

i

l

i

!

500

407 Table 6 Apparent adsorption rate, RL, and diffusion parameters, adsorption of methyl iodide and benzene

Dl/2/ro, for the

RL (cm 3 g" min 1/2) P1

P2

P3

Benzene

0.023

0.002

0.043

methyl iodide

0.049

0.015

0.068

D~/ro (min -~)

Benzene

P1

P2

P3

0.178

0.020

0.419

0.097

0.580

methyl iodide 0.572 Reprinted from: M. Domingo-Garcia et al. [12].

These data indicate considerable differences for the adsorption process of these hydrocarbons on each sample, particularly on sample P2 for which these parameters are more than 10 times lower than for P1 and 20 times lower than for P3 when benzene is adsorbed. It is possible, therefore, to assume important differences in the porous system of the three carbon samples due to the different preparation recipes. Actually, from data in Table 5 one can observe that carbon P2 has no macroporosity at all, whereas P1 presents a large one. It is known that macroporosity is very important for the transport and quick penetration of the adsorbate molecules to the micropores. However, the kinetics parameters are lower for carbon P1 t h a n for carbon P3, although the latter shows a less extensive macroporosity. This fact, together with data in Table 5 and the restricted N2 adsorption for carbon P2 due to pore constrictions, suggests that there are also important differences in their microporous system. On the other hand, the adsorption rate and the diffusion parameters of methyl iodide on all the samples are much higher t h a n those of benzene and this is more marked for P1 and P2 than for carbon P3. This means that at the same temperature and relative pressure methyl iodide is kinetically better adsorbed than benzene. The adsorption-desorption isotherms of benzene, n-hexane and 2,2-DMB are depicted in Figure 8. Most of them can be assigned to BDDT type I, i.e. corresponding to predominantly microporous adsorbents; nevertheless, differences in their shape and in the desorption process also indicate a variety in the characteristics of the porous system. Isotherms of benzene are depicted in Figure 8(a). For carbons P1 and P3 the isotherms obtained with 2 hours of adsorption time are coincident with those obtained after 12 hours; therefore, only the isotherms obtained with long adsorption time are depicted in Figure 8. For carbon P2, however, the adsorption-desorption isotherm obtained after 2 hours of adsorption time is very different to that obtained after 12 hours and both have

408

0,3

P3

0,25 P3

a

0,20 0,2

-~-: ~ . _ _. ~ ...... - . .o ~. - -.~ ,.~ - ~ o ~~

Pl

P1

-~ 0,15

E gO,1

"

_,-

A-

> 0,10

P2 (12 h)

P2 I

0,05 zl~*! ** *,

*

0,0 0

0,2

*,

,

0,4

0,6

-! m - -ID-

P*2 (2 h) ,

0,00

0,8

0,0

HI

-I--

!

I

0,2

0,4

I

0,6

0,8

1,0

P/Po

P/Po

0,2 P3

~0,1 E > 0,1

0,0 0,0

i

i

i

i

0,2

0,4

0,6

0,8

i

1,0

P/Po

Figure 8. Adsorption isotherms on the glassy carbons: (a) benzene, (b) n-hexane and (c) 2,2-DMB.

been depicted in this figure. The shape of all these isotherms suggests that sample P1 appears to have predominant adsorption on micropores with little participation on the external surface. However, samples P2 and P3 exhibit an upward deviation in the region of high relative pressure, which is normally interpreted either as capillary condensation in supermicropores and small mesopores or adsorption on the external surface of the adsorbent [35,36]. These two effects follow a different adsorption mechanism to that of adsorption on micropores; capillary condensation often produces an hysteresis loop at high relative pressure when the desorption process is carried out. Desorption isotherms represented in Figure 8(a) show a very small deviation from the adsorption branches on samples P1 and P3, which extends to the lowest relative pressure region. The value of these deviations in volume of adsorbate is less than 0.01 cmS/g. Nevertheless, a very pronounced hysteresis loop in the complete range of relative pressure studied, is found for the two isotherms on sample P2. Of the several hypothesis given [1,42] to explain the appearance of low pressures hysteresis loop, taking into account the high rigidity of glassy carbons, it has to be considered that activated passage of molecules through pre-existing constrictions into wider pores appears to be the most appropriate for these

409 samples. This interpretation is consistent with the restricted adsorption of N2 already mentioned found for carbon P2 (SN2<< Sco2). The adsorption isotherms of n-hexane and 2,2-DMB (obtained after 2 hours of adsorption time) are depicted in Figure 8(b, and c). Adsorption of n-hexane gives the same kind of isotherms as those obtained for benzene on every carbon. Cyclohexane, however, is not adsorbed on P2 and a very small amount appears on P1 and P4 while a considerably greater adsorption is found on P3. Moreover 2,2-DMB is only adsorbed on P3. All this means that carbons P1, P2 and P4 present a high level of discrimination (or molecular sieve behaviour) for molecules with a minimum critical size such as cyclohexane or larger, while P3 presents the most accessible microporosity. In this sense the Vo/V'o ratio for n-hexane/2,2-DMB, benzene/2,2-DMB, n-hexane/cyclohexane and benzene/cyclohexane couples can not be calculated because of the negligible value of adsorption of 2,2-DMB and cyclohexane on P1, P2 and P4. However, for P3 the value for the benzene/cyclohexane ratio is between 1 and 2. Adsorption-desorption isotherms of methyl iodide and chloroform (Figure 9) show a behaviour comparable to that of benzene. Nevertheless for methyl iodide on sample P2 at relative pressures less than 0.4 the deviation between the adsorption and desorption branches becomes much smaller than that for benzene with a similar hysteresis loop to that expected for capillary condensation on supermicropores or small mesopores. This result, together with the kinetics parameters, indicates that methyl iodide passes more easily through the pore constrictions than benzene does in spite of their similar minimum critical dimension. With respect to chloroform, which has a minimum critical dimension slightly higher than that of n-hexane, the adsorption isotherms on the four samples were comparable to that of benzene and different from that of cyclohexane. Thus the above mentioned molecular sieve effect shown by carbons P1, P2 and P4 appears between chloroform and cyclohexane.

0,25

..~

a

0,20

-

0,15

o~~-.---#

-

A-.....

- ~/>-

_~

-o

0,25 ]

P3

P3

0,20

P1

"~ P 2

0,15

E

~" 0,10

g

> 0,10

0,05 ~ ' ~ - ~ '

=J

=

~

~

-&

M -A-

~

~

:

.

.

0,4

0,6

i

P2

0,05

0,00

,

0

b

0,2

0,4

0,6 P/Po

0,8

1

0,00

!

0

0,2

0,8

1

P/Po

Figure 9. Adsorption (close symbols) desorption (open symbols) isotherms on glassy carbons of (a) methyl iodide and (b) chloroform.

410

The behaviour of the adsorbates in the adsorption of these molecules has to be related with the micropore network. In order to obtain information of these networks, the DR equation has been applied to the adsorption isotherms of all the adsorbents. Characteristics curves obtained from the application of DR equation are represented in Figure 10. The micropore volume accessible to each organic molecule calculated from the linear region of these curves are compiled in Table 7 together with the adsorption volume at relative pressure of 0.9.

0,00

-0,5

b

a *

>

%

~ ~"~_

-1,00 ~ t ' ~ ~ l ~ ~ ~ ~

n-hexane "*

.

-1,5 0

~ benzene" ~ .

.

3

de

>o)

=

--2,00

chloroform .

3,ool

.

6

9

methyl iodide

12

0

Ig2 (P/Po)/IT~

chloroform 5

,

,

10

15

Ig2(e/Po)/~ 2

-0,50 ] c

methyl iodide

~-1,00" ~ ~ n z e n e __

\ I -2,00 . 0

o,',,oro,orrn

9

2,2-DMB . .

.

5

10

15

Ig2 (P/Po)/132

Figure 10. Dubinin-Radushkevich plots of benzene, n-hexane, 2,2-DMB, methyl iodide and chloroform on (a) P 1, (b) P2 and (c) P3.

Plots for carbon P1 (Figure 10 a) seem to indicate that the adsorption of benzene, n-hexane, methyl iodide and chloroform takes place on the same type of micropores. However, the downward deviation from linearity observed for benzene, n-hexane and chloroform at low relative pressures can indicate that in this region equilibrium is not attained for the adsorption of these molecules. This type of deviation is often explained in terms of restricted diffusion into the narrowest pores or because the molecular sieve effect can lead to difficulties in adsorption at low relative pressures[2,43]. It should be noted that for n-hexane this deviation extends up to much higher values of relative pressure than for chlorofom, although the minimum critical size of the latter is somewhat

411 greater.This fact could be due to packing restrictions in the micropores [2] because of the great length of n-hexane. On the contrary, for methyl iodide an upward deviation appears in this region of low relative pressures, suggesting that this molecule has good accessibility into the narrowest pores.

Table 7 Micropore volume, Vo (cm3/g), from DR equation and volume at P/Po =0.9, Vs P1 P2 P3 Vo

V~

Vo

V~

Vo

V~

Benzene

0.173

0.21

0.068

0.14

0.156

0.23

n-hexane

0.169

0.19

0.030

0.06

0.113

0.22

0.092

0.13

2,2-DMB methyl iodide

0.176

0.20

0.093

0.17

0.114

0.23

chloroform

0.177

0.21

0.072

0.13

0.113

0.32

Reprinted from: M. Domingo-Garcia et al. [12].

For carbon P2 the DR curves of the four adsorbates are depicted in Figure 10(b). On this adsorbent the curves of benzene, methyl iodide and chloroform corresponding to isotherms with the same adsorption time are not at all coincident. As previously indicated by the kinetics adsorption measurements, benzene and methyl iodide present very low diffusion parameters on this carbon, although for the latter this is almost five times higher than for the former. Actually, the volume Vs of benzene adsorbed increased from 0.07 cma/g after 2 hours of adsorption time up to 0.14 cma/g after 12 hours, and up to 0.17 cm3/g after 5 days. For methyl iodide the same volume of 0.17 cm3/g is attained after 12 hours. This fact could explain the very different slopes for the corresponding characteristic curves. Therefore, although adsorption on carbon P2 appears to be kinetically restricted for the three adsorptives, this is much smaller for methyl iodide than for benzene and the highest restriction is found for chloroform. As already suggested this restriction seems to be due to constrictions in the entrance to the pores [41]. Another fact to be considered is the high dipole moment of methyl iodide, which can produce specific interactions with the oxygen functional groups, which are on the carbon surface. This contribution accumulates with the dispersion forces leading to a considerable increase of the adsorption energy in the initial region of the adsorption process, i.e. at very low relative pressure. This specific contribution decreases following micropore filling of the adsorbent [6,23]. Chloroform is the less easily adsorbed of the three adsorptives as its critical

412 dimension of 0.43 nm is the greatest. The fact t h a t chloroform, which is also a polar molecule, suffers stronger restriction t h a n benzene to access micropores of P2 indicates t h a t the main p a r a m e t e r s controlling the passage through the pores constrictions is the critical dimension and the shape of the molecule. The molecular sieve effect between benzene and cyclohexane confirms the network of these glassy carbons to be made up by slit-shaped micropores. As pointed out earlier, the micropore system of carbon P3 also appears to have i m p o r t a n t differences with respect to those of carbon P1, P2 and P4. This becomes more obvious when the characteristic curves of the different adsorbates are analysed. The most i m p o r t a n t difference is t h a t the adsorption capacity of cyclohexane and 2,2-DMB on sample P3, although smaller t h a n t h a t of the others adsorbates, turns out to be quite important. Nevertheless, comparing the curves for the different hydrocarbons, it can be observed t h a t the slope of the linear region of the curves obtained with cyclohexane and 2,2-DMB is much steeper t h a n t h a t of the other curves, which usually suggests adsorption on wider pores [33]. Apart from this, the curves of benzene and methyl iodide are coincident as well as those of n-hexane and chloroform. The adsorption of n-hexane seems to suffer some restrictions at low relative pressures while for chloroform only a linear region t h a t extends over a wide range of relative pressures, (P/Po from 10 .3 to 1.5 10 -1) is found. Moreover, at very low relative pressures an upward deviation appears not only for the characteristic curve of methyl iodide but also for t h a t of benzene. On the other hand, observing the Vs values given in Table 7 it is evident t h a t certain molecular sieve behaviour between chloroform and cyclohexane is also shown by carbon P3. The steep slopes of the characteristic curves of cyclohexane and 2,2-DMB suggest t h a t adsorption probably takes place in the supermicropores and small mesopores. It is interesting to note t h a t all plots in Figure 10 show an upward deviation from linearity near the s a t u r a t i o n relative pressures which can be attributed to the filling of supermicropores or small mesopores by a co-operative m e c h a n i s m [44] which involves little, if any, e n h a n c e m e n t of the adsorption energy. Characteristic energies of adsorption, Eo, calculated from the linear region of the characteristic curves of benzene, methyl iodide and 2,2-DMB on carbons P1 and P3 are listed in Table 8. Moreover the corresponding average micropores width, Lo, and the range of validity of r e l a t i v e pressures from which these p a r a m e t e r s were obtained, are also included. The average micropore width, Lo, obtained for carbon P1 with benzene is quite coincident with t h a t calculated with methyl iodide. However, for carbon P3 a slight difference of this p a r a m e t e r is found between these two adsorbates. The very low characteristic adsorption energy, Eo= 6.1 kJ/mol, corresponding to 2,2-DMB on this carbon is typical of adsorption on supermicropores or even small mesopores as pointed out earlier. Therefore one can conclude t h a t these data do not correspond to a specially narrow microporous system, which is really surprising taking into account the high discrimination effect already described.

413 Table 8 Characteristic adsorption energy, Eo, and average micropore width, Lo

P1

benzene

methyl iodide

Eo (kJ/mol)

20.9

21.7

Lo (nm)

1.17

1.10

2,2-DMB

Range of validity for P/Po 4.5 10-3-1.7 10 -1 1.3 102-3.5 10 -1 P3

Eo (kJ/mol)

17.1

19.8

Lo (nm)

1.5

1.26

6.1

Range of validity for P/Po 2.7 10-2-1.6 10 -1 3.0 10-2-3.5 10 -1 4.0 10-2-3.0 10 -1 Reprinted from: M. Domingo-Garcia et al. [ 12].

An u p w a r d deviation from the linearity in the region of low relative pressures (such as the adsorption of methyl iodide on carbons P1 and P3 or benzene on carbon P3) in DR plots could be the consequence of the superposition of two extreme ranges of microporosity [33]. These two ranges can be well approximated by a binomial equation, known as the Dubinin-Isotova (DI) equation:

E/1

V=V01exp-

EOl[3 x RT In

+ Vo2 exp -

E0213

(3)

V01 is the micropores volume obtained from the linear region at very low relative pressures and according to Stoeckli et al. [45], could be related with the micropore filling process which gives a characteristic adsorption energy Eol. V02, and E02 have been associated with the beginning of the secondary micropore filling process, i.e. the volume and energy of adsorption on the walls of relatively large micropores. Therefore the volume V02 could provide information on the monolayer capacity of the largest micropores walls. Table 9 contains all these p a r a m e t e r s for the adsorption of methyl iodide on samples P1 and P3 and of benzene on P3. From these new data one can assume t h a t more t h a n 70% of the total adsorption volume Vs (Table 7) on carbon P1 takes place on micropores with an average dimension, L01, of 0.74 nm. Nevertheless this new value is still not small enough to explain either the drastic molecular sieve behaviour found for cyclohexane or the restricted diffusion of benzene at low values of relative pressures shown earlier. This fact supports the existence of some kind of constrictions at the entrance of the pores. For carbon P3, V01 and E01 data for both benzene and methyl iodide are quite close and indicate t h a t 56% of the total adsorption volume takes place on micropores of about 1 nm or less. This means t h a t 44% of the total micropores volume (i.e., 0.102 cma/g) corresponds to supermicropores or small mesopores which filled up by the secondary mechanism

414 with very low adsorption energy Eo2. This a m o u n t is very similar to the Vo value found for the adsorption of 2,2-DMB on P3, which means t h a t this molecular probe is only adsorbed on the wider pores by a secondary mechanism. Table 9 P a r a m e t e r s obtained from the Dubinin-Isotova equation Adsorbate

Vol (cma/g) Eol(kJ/mol)

Lol(nm) Range of validity for P/Po

P1

methyl iodide

0.146

27.7

0.74

3.0 10-3-3.0 10 .2

P3

benzene

0.128

21.4

1.10

3.0 10-3-3.0 10 .2

methyl iodide

0.120

23.2

1.01

3.0 10-3-3.5 10 .2

V02 (cma/g) Eo2(kJ/mol)

Range of validity for P/Po

P1

methyl iodide

0.040

7.7

3.0 10-2-5.0 10 -1

P3

benzene

0.049

5.0

3.0 10-2-6.0 10 -1

methyl iodide

0.047

6.4

4.0 102-5.0 10 -1

Reprinted from: M. Domingo-Garcia et al. [12].

From the analysis of these four glassy carbons one can conclude t h a t carbons P1, P2 and P4 have a complete discrimination (molecular sieve behaviour) for the adsorption of chloroform, with a critical dimension of around 0.43 nm, and cyclohexane, with a critical dimension of 0.56 nm. Nevertheless, when the isotherms of benzene or methyl iodide on sample P1, P2 and P4 are analysed by the application of DR and DI models the average micropore sizes calculated do not justified theirs molecular sieve properties. This fact, along with the kinetics results, indicates t h a t constrictions or some narrowness at the entrance to the pores are responsible for the molecular sieve character found for these glassy carbons. Carbon P3, however, shows a different behaviour for the adsorption of these molecules, and although it acts as a molecular sieve for cyclohexane and 2,2-DMB at low relative pressure, these two hydrocarbons were adsorbed on supermicropores or small mesopores. As a general conclusion, one can point out t h a t both kind of adsorbents here discussed, activated carbons of lignocelulosic origin and glassy carbons, can be tailored to adsorb selectively organic molecules. For the former carbons this behaviour essentially depends on activation time, while for the latter it depends on the preparation formula. In addition, this discriminative behaviour is much higher on glassy carbons t h a n on those obtained from agricultural by-products. Nonetheless, the adsorption capacities of activated carbons of lignocelulosic origin are always much higher t h a n those of the glassy carbons.

415

2.3. A d s o r p t i o n of CO2 f r o m d i l u t e d e n v i r o n m e n t s A particular case of interest is the adsorption of CO2. This process is currently used as an almost routine measurement in order to determine the surface area of adsorbents. For this purpose the adsorption is carried out from low to high relative CO2 concentration. The data to be reported deal with the adsorption kinetics of CO2 from diluted CO2-N2 mixtures, conditions prevailing in flue gases produced in fossil-fuel-based power plants. Activated carbons prepared by carbonization (H0) and further activation by CO2 (H14, H25 and H35) and by steam (HW20 and HW44) have been used as adsorbents. Additional experimental and theoretical data are given elsewhere [46]. The textural characteristics and the micropore size distributions of these samples are shown in Table 10 and Figure 11.

Table 10 Textural characteristics of the active carbons SN2(m2/g)

Sco2(m2/g)

Vo (cm3/g)

Eo (kJ/mol)

H0

246

813

0.314

22.23

HI4

725

974

0.376

20.92

H25

910

1012

0.391

20.68

H35

1190

1220

0.471

19.01

821

0.317

20.58

HW40 1475 1517 Reprinted from: M. A. Salas-Peregrin et al. [46].

0.586

15.45

HW20

812

3 H25

E2

] |

.-.

b HW20

H35

E

.._... _J

}Q1 0

0,8

"EI

|

|

!

|

|

1

1,2

1,4

1,6

1,8

0

0,8

L (nm)

Figure 11. Micropore size distribution of activated carbons.

,

,

i

1,3

1,8

2,3

L (nm)

416 Diffusion parameters were obtained using equation 1. The values of Ve, Vo and D~/ro at 298 K obtained with a CO2-N2 mixture containing 13.5 % of CO2 by volume for all these activated carbons are compiled in Table 11. These results can be explained on the basis of the micropore characteristics of the adsorbents. Therefore, the carbonized sample, H0, has the smallest value of Dl/2/ro because it possesses the narrowest micropores as can be deduced from its highest adsorption energy, Eo (Table 10) and from the distribution of micropores (Figure 11). The negative value of Vo indicates that there is a retardation in the CO2 adsorption process.

Table 11 Values of Ve, Vo and D~/ro obtained with a CO2-N2 mixture containing 13.5% of CO_9 by volume Sample

Ve

(cma/g)

Vo (cma/g)

D~/ro (s -~)

H0

16.38

-0.09

0.0083

HI4

18.30

5.17

0.0174

H25

18.31

5.96

0.0174

H35

17.11

2.98

0.0174

HW20

17.86

4.67

0.0177

HW44

14.22

3.72

0.0204

Reprinted from: M. A. Salas-Peregrin et al. [46]. When sample H0 is activated in either CO2 (H-series) or steam (HW-series), there is an increase of both Ve and D1/2/ro, and Vo reaches a positive value so that there is no retardation in the CO2 adsorption process. The increase in Ve and D1/2/ro is produced as a consequence of the raised micropore volume and because of an opening of the microporosity produced by the activation process, which can be deduced from the decrease in Eo (Table 10) and from Figure 11. Samples H14 and H25 have coincident Ve and D1/2/ro values because these two samples have almost the same micropore size distribution and similar Eo values. When activation increases, in samples H35 and HW44, there is a decrease in the Ve value, and this is more marked in the case of the most activated sample, HW44. These results, at first, are surprising because H35 and HW44 samples have the highest micropore volume from their respective series. Thus, these results indicate that for a low CO2 concentration in the gas phase (13.5% of CO2 in N2) the higher the degree of activation of the carbon the lower CO2 adsorption capacity at equilibrium. The same trend, as shown in Table 10, is found for other lower C02 concentrations in the C02-N2 mixture. Similar results have been reported [47] for the adsorption of Volatile Organic Compounds (VOC) at trace level onto activated carbon fibres.

417 These results can be explained as being due to a decrease in the number of the smallest micropores available in the most activated samples, which can be deduced from Figure 11 for H35 and HW44, respectively. This occurs because the wider micropores do not benefit from the overlapping adsorption potential of opposite pore walls, and thus, do no experience the enhanced adsorption observed for the narrower micropores. The results found are quite important, because they show that when using activated carbons to remove or store CO2 from low concentration environments, their surface area or pore volume alone is not an adequate design parameter, but micropore size distribution is the controlling factor. Therefore, these results indicate that activated carbon H14 (a sample with a low degree of activation) performs best under these experimental conditions. Moreover, these results together with those reported [47] permit the above conclusions to be applied not only for the adsorption of CO2 but also to the adsorption of VOC at very low concentrations. Consequently, these ideas should be borne in mind when analyzing results in the following sections. 3. A D S O R P T I O N F R O M V E R Y D I L U T E D A T M O S P H E R E S

In the previous sections the adsorption of several molecules at relatively high concentrations measured in gravimetric systems has been considered. In this section the adsorption of VOC under dynamic conditions, at relatively high temperatures and at very low vapour concentration is considered. Among the VOC, several hydrocarbons (linear, cyclic and branched) and some organic compounds with different functionalities as acetone, diethyl ether, tetrahydrofurane, carbon tetrachloryde, chloroform, dichloromethane, methyl iodide and n-alcohols have been chosen. Adsorption has been studied using IGCS. Details of the experimental conditions are give elsewhere [13,22,26,27,29,41,48]. 3.1. A d s o r p t i o n o f V O C o n n o n - p o r o u s c a r b o n m a t e r i a l s

When adsorption is produced on non-porous surfaces the process is controlled by displacement of the electronic density of the molecules produced by action of the electrostatic field of the adsorbent. As a consequence of this the degree of adsorption of linear hydrocarbons increases with the number of carbon atoms of the molecule. Therefore, the specific retention volume Vs (a p a r a m e t e r which measures the degree of adsorption) of, for example, n-nonane is higher than that of n-hexane. Moreover, the degree of adsorption decreases as the temperature increases. These kind of behaviours are shown in Table 12 for the adsorption of linear hydrocarbons at two temperatures on three non-porous carbon materials: a graphitized carbon black (V3G) and two graphites (Pyrolitic and Acheson). For the same reason as that explained above, the relationship at different temperatures between in Vs versus the polarizability of the n-alkanes is linear. Moreover, the differential heats of adsorption are, in absolute values, low and close to the liquefaction heats of the adsorbates [29]. This means that the adsorbate-adsorbent interactions are similar to a d s o r b a t e - a d s o r b a t e ones, i.e.

418 Table 12 Specific retention volumes for the adsorption of n-alkanes on non-porous graphites

Vs (cma/m 2) 343 K

363 K

Adsorbate

V3G

Pyrolitic

Acheson

V3G

Pyrolitic

Acheson

n-C6

4.6

2.6

2.2

2.6

1.5

1.5

n-C7

19.4

7.4

8.4

9.2

3.4

4.3

n-Cs

81.7

28.8

34.0

34.9

12.5

12.9

n-C9

331.6

121.1

109.6

125.5

47.6

44.7

the adsorbate-adsorbent interaction is produced through London dispersion forces. W h a t it is noteworthy is t h a t this kind of behaviour remains in the adsorption on non-porous carbon materials of hydrocarbons which have non-linear shapes. This is the case of benzene, cyclohexane and isooctane. The interaction of these adsorbates on non-porous graphite (Degussa) is similar to t h a t of n-alkanes. Therefore, one can conclude t h a t the adsorption is produced as a consequence of a displacement of the charges of the molecules under the action of the surface of the adsorbent (non-specific interaction). This is shown in Figure 12.

7 363 K

S

if}

>

3

Isooctane Benzene

J

.t,

9 n-alkanes

,~~" Cyclohexane

8

i

i

i

12

16

20

Ct(A3) Figure 12. Variation of In Vs versus the polarizability of the adsorbates for the adsorption on a non-porous graphite.

However, this behaviour is different if the adsorbate has the same shape as t h a t of the n-alkanes but different chemical functionalities. This is the case with the data in Table 13, in which the Vs values of n-alkanes and n-alcohols on graphite (Degussa) are compiled. The degree of adsorption is larger for an

419 Table 13 Specific retention volumes and differential heats of adsorption of n-alcohols and n-alkanes on graphite (Degussa) Vs (cma/m 2)

.AHOA

Adsorbate

333 K

343 K

353 K

363 K

(kJ/mol)

l-propanol

3.6

2.7

2.0

1.5

29.7

l-butanol

7.4

5.6

3.9

2.7

33.1

l-pentanol

13.3

9.1

6.3

4.6

35.1

l-hexanol

60.3

37.7

26.6

16.1

42.4

n-pentane

2.2

1.8

1.3

1.0

28.2

n-hexane

11.4

8.0

5.6

4.0

34.3

n-heptane

42.1

27.7

18.7

12.6

39.6

n-octane

134.3

76.7

50.9

34.5

44.2

n-nonane

487.9

267.7

175.9

104.6

49.6

1465.6

837.2

441.4

55.0

n-decane

n-alcohol t h a n for an n-alkane with the same n u m b e r of carbon atoms. This is because adsorption of n-alcohols depends on their deformation polarizability and on the orientation polarizability (specific interaction) which depends on the dipolar moment, p. Of these two factors, the orientation polarizability is far larger t h a n the deformation polarizability. Consequently, the behaviour of polar molecules u n d e r these conditions is different from the non-polar molecules and depends mainly on the orientation polarization. Moreover, for this reason the differential h e a t of adsorption is also higher, in absolute values, for an n-alcohol t h a n for an n-alkane with the same n u m b e r of carbon atoms. It could be desirable to improve the behaviour of these kind of materials in order to increase their capacity of adsorption. For this purpose the t r e a t m e n t of V3G with oxygen at different degrees of burn-off has been carried out. The results obtained are basically similar to these obtained with the original carbon black [26]: a linear relationship between In Vs and polarizability of the hydrocarbons, such t h a t adsorption increases with the polarizability of the chain length of the n-alkane as well as with the lowering in the adsorption t e m p e r a t u r e . However, the differential h e a t of adsorption for each n-alkane increases in absolute value as the percentage of burn-off increases, as a consequence of the irregularities produced on the surface of the sample [26]. From a practical point of view these data together indicate that the elimination by adsorption on non-porous carbons of hydrocarbons is favoured by the polarizability of the adsorbate such t h a t the adsorption increases with the

420 polarizability of the molecule. Moreover, the adsorption of n-alcohols in the same conditions is more favoured t h a n that of the n-alkanes due to the dipolar moment of the former molecules. 3.2. A d s o r p t i o n o f V O C o n p o r o u s c a r b o n m a t e r i a l s It has already been mentioned that the adsorption on these kind of materials has to be considered taking into account two main factors: the porosity and the chemical functionalities of the adsorbate and adsorbent. The adsorption of linear molecules on activated carbons follows the same trend found for non-porous carbons, i.e. the plot of in Vs versus the n u m b e r of carbon atoms (polarizability) of the adsorbates is a straight line. Therefore, the mechanism of adsorption can be considered as non-specific. Nevertheless, those non-linear hydrocarbons do not follow the same trend as that found for non-porous carbons. This behaviour is shown in Figure 13 in which the adsorption of linear (n-alkanes), branched (2,2-DMB) and cyclic hydrocarbons (benzene and cyclohexane) on two activated carbon series, C and H, is plotted.

H-13 H-2

CeH

9

"J

n-alkanes

0

Y

~ 1

> _c

> --CO

-2

~

06H12 ~,

-1

,

,

,

10

12

14

oc (A 3)

9 06H12

~ t1"

A 2,2-DMB 8

/)P . / n-alkanes

C6H6m

2

-2

8

9 2,2-DMB ,

,

,

,

,

,

9

10

11

12

13

14

o~ (A 3)

Figure 13. In Vs versus the polarizability for the adsorption on activated carbons.

The fact that these non-linear molecules do not follow the behaviour found for n-alkanes suggests that the shape and size of these molecules is an important factor in the adsorption on porous materials. This can be seen in Figure 14 in which the specific retention volumes are plotted versus the mean molecular dimensions of the adsorbates (see Table 16). It is interesting to emphasize that all the adsorbates have different shapes and sizes and the same n u m b e r of carbon atoms (six). The general trend found is that the adsorption decreases as the size of the molecules increases. Moreover, at low degree of t r e a t m e n t of the raw material (C-2 and H-2 carbons) the specific volumes of adsorption are clearly lower t h a n at higher periods of treatments. In all cases the lower volume of adsorption is for 2,2-DMB which is the largest molecule (0.60 nm). In addition the adsorption of this molecule, which is almost negligible at low degrees of treatments (C-2, C-13 and H-2) increases at higher degrees of t r e a t m e n t of the raw materials. The separation ratios, (Vs/Vs'), on H-2 for the n-hexane/2,2-DMB and benzene/2,2-DMB couples are 87.5 and 101.8.

421 20

C-24

20 ~ E

C-13

E

E E ~ {/1

H-13

tar)

>

>

10

10

0 0,40

H-3

0 0,45

0,50 d (nm)

0,55

0,60

0,40

,

,

0,45

0,50 d (rim)

, 0,55

I, 0,60

Figure 14. Specific retention volumes versus the molecular size of hydrocarbons with six carbon atoms (see Table 16 for molecular sizes).

However, for samples obtained with long periods of t r e a t m e n t s , H-30, these ratios are 2.7 and 1.5 respectively. These separation ratios are in all cases higher t h a n those obtained in the adsorption at high relatives pressures above commented. This is because the n u m b e r of molecules to be adsorbed in this case is very low. Consequently, the possibility of finding a n u m b e r of pores with a dimension close to the molecular dimension is high. In contrast to this at high relative pressures the n u m b e r of molecules to be adsorbed is very high and consequently this possibility is clearly lower. On the other hand, the separation ratios found at low concentration support the hypothesis t h a t progressive t r e a t m e n t s of the raw m a t e r i a l open the microporosity of the activated carbons. For this reason it can be concluded that, in general, the activation times higher t h a n 13 hours do not improve the capacity of adsorption of these pollutants from very diluted concentrations. Consequently, on using almond shells or olive stones to prepare activated carbons for removing these organic molecules one can diminish the period of t r e a t m e n t of the raw material. Similar results to those obtained with these series are also reported [15] for activated carbons prepared by activation of chars (HA-series), although the Vs values are slightly lower. The second factor commented above, to be considered in the adsorption of VOC concerns the chemical functionalities of both adsorbate and adsorbent. A case of practical interest is the adsorption of methyl iodide. Iodine-131 is one of the dangerous substances produced in the fission process, but it mainly survives in the atmosphere as methyl iodide. The elimination by adsorption of this substance has been studied using adsorbent materials such as zeolites, silica gel and activated carbons, but the latter seem to show the best results [48-50]. Methyl iodide has an i m p o r t a n t dipolar m o m e n t ~ = 1.62 D. As a result it can be expected to make an i m p o r t a n t contribution of specific interactions in the adsorption of this molecule. One could, therefore, expect an increase in the adsorption capacity in cases in which specific interactions are favoured.

422 To study this effect several activated carbons treated with H202 [41] have been used for the adsorption of methyl iodide. In Table 14 some selected results are recorded. Samples with the symbol O appended at the end of the name have been obtained by treatment the original samples with H202. So, the comparison of data in Table 14 deals with the effect of the oxygen functionalities, on the surface of the active carbons, on the capacity for methyl iodide adsorption. An increase in the adsorption capacity of samples treated with H202 is apparent from these data. Surprisingly, this increase does not mean a more exothermic process as could be expected from a specific interaction. In fact, these data show no energetic differences between the interaction in the original samples and in those obtained after H202 treatment, having a higher differential heat of adsorption, in absolute value, than the liquefaction heat of methyl iodide, -AHL = 27.6 kJ/mol [48].

Table 14 Specific retention volumes, Vs, at 473 K, and differential heats of adsorption, AH~ of methyl iodide on activated carbons Sample

Vs(cma/m 2)

-AHOA(kJ/mol)

Sample

Vs(cma/m 2)

-AH~ (kJ/mol)

C-2

1.2

56.1

C-2-O

1.7

53.6

H-2

0.6

64.8

H-2-O

1.2

52.2

HA2

0.6

49.2

HA2-O

0.9

58.1

HA4

1.0

52.2

HA4-O

0.9

52.5

Reprinted from: F. Carrasco-Marin et al. [48].

Another possible way to increase the capacity of activated carbons for methyl iodide adsorption is by impregnation with KI [41,51]. The data in Table 15 show an increase in the adsorption capacity of the samples treated with KI. In addition they show a decrease in the differential heats of adsorption, although they are also higher, in absolute value, than the methyl iodide liquefaction heat. Hence, the results are similar for the samples treated with H202 and with KI. This could be because the interactions in both cases are produced by a mixed mechanism (specific + non-specific) such that the interaction of the adsorbate with the pores can be more exothermic than with oxygen or with KI. This aspect will be considered later on discussing the adsorption on glassy carbons. These results show that it is possible to increase the adsorption capacity for methyl iodide of activated carbons if different chemical factors capable of specific interactions are introduced. However, these interactions do not appear to be more exothermic than those inside the pores.

423 Table 15 Specific retention volumes at 443 K, Vs, and differential heats of adsorption, AH~ of methyl iodide Vs(cm3/m 2) Sample

-AH~ (kJ/mol)

Original

With KI

Original

With KI

C-13

3.5

3.9

57.3

48.7

C-30

2.8

3.0

53.3

48.3

H-13

3.0

3.2

53.0

51.0

H-24

2.2

3.1

49.8

47.0

Merck

0.3

1.2

57.2

59.9

In relation to carbon materials with a tailored porosity for adsorption of VOC, several adsorption data obtained on carbons prepared from Saran copolymer and from polyfurfuryl alcohol have been reported [10,13,41,52,53]. The behaviour of carbon materials obtained by carbonization of Saran in the adsorption of linear hydrocarbons is the same as that described above, i.e. a linear relationship between In Vs and the number of carbon atoms. Moreover, the Vs values are higher than those obtained with other carbon materials and with a maximum adsorption for the temperature of 1373 K. However, Vs is also affected by the shape and size of the non-linear molecules such that the Vs/Vs' ratio for benzene/cyclohexane couple is close to unity, for treatment temperatures ranging from 973 to 1173 K, while it raises to 210 for the sample obtained at 1573 K. In addition, the adsorption of 2,2-DMB is negligible in all the samples. Therefore, these carbon materials have a large capacity for VOC adsorption and even with high capacity of discrimination of adsorbates depending on the shape and size and on the preparation temperature of the samples. This behaviour of the sample prepared at 1573 K is explained by shrinkage of the carbonaceous structure which permits the adsorption of benzene and inhibits that of molecules larger than cyclohexane. The behaviour of these carbon materials is so versatile that the porous structure can be modified by different treatments. The structure of the porous network of the sample obtained at 1573 K can be progressively opened by mildly gasification at 1 and 6% of burnoff, and the new samples so obtained loose the capacity of discrimination for adsorption of the benzene/cyclohexane couple. The introduction of oxygen chemical groups in these samples by treatment with HNO3 diminishes the capacity of adsorption of hydrocarbons. This is a consequence of fixation of chemical groups at the entrance of the pores [10,54] This decrease is the largest for 2,2-DMB for which the adsorption is almost negligible due to their larger molecular size (0.60 nm).

424 30 180

E

20

.....,

-'~'140

E

o v r

>

E

10

n-pentane

>

100

n-butane 0

973

r"

'

'

'

1173

1373

1573

60

973

T(K)

,

,

,

1173

1373

1573

T (K)

Figure 15. The specyfic retention volumes of linear hydrocarbons versus the carbonization temperature of Saran.

Moreover, the effect produced by the oxygen functionalities can be enhanced if the chemical t r e a t m e n t is carried out with CS2, because in this case the chemical functionalities fixed at the entrance of the pores are sulphur compounds which are larger in size t h a n the oxygen compounds. This effect of the sulphur functionalities is shown in Figure 16. In this figure the separation ratios for the n-hexane/benzene, benzene/cyclohexane and n-hexane/cyclohexane couples are plotted versus the percentage of sulphur content in a sample obtained by carbonization of S a r a n at 1173 K and further t r e a t m e n t s with CS2 . A very high discrimination capacity is shown for 4.3% sulphur content for benzene/cyclohexane and n-hexane/cyclohexane couples, as a consequence of cyclohexane having a larger molecular size t h a n benzene and n-hexane. In addition, the n-hexane/benzene ratios are almost the same in all cases. This suggests t h a t the sulphur functionalities have partially closed the access to the interior of pores [22].

1800 1400

c c

1000 >

benzene/ cyclohexane

600 n-hexane/ benzene

200 -200 0

1

2

3

4

5

s(%)

Figure 16. Effect of the sulphur content on the discrimination capacity of sample S 1173.

425 Concerning the carbon materials obtained by polymerization and further carbonization of furfuryl alcohol the samples used were those whose adsorption behaviour at high relative pressures has been commented above. The following VOC were adsorbed at very low vapour concentration [55]: carbon tetrachloride (CCI4), chloroform (CHCla), dichloromethane (CH2C12), acetone (C3H60), diethyl ether (CnHloO), t e t r a h y d r o f u r a n e (C4H80, THF), n-alkanes (CnH2n+2, from C4 to C7), benzene (C6H6), cyclohexane (C6H12) and 2,2-DMB (C6H14). The change of In Vs for the adsorption of linear hydrocarbons with the n u m b e r of carbon atoms follows, as is usual, a straight line. Again, when the shape and size of the adsorbates are different, the porosity of the samples is a very i m p o r t a n t factor in determining the degree of the adsorption. This is shown in Figure 17 in which Vs values obtained at 473 K for the four carbons are plotted against the critical dimension of the adsorbates with the same n u m b e r of carbon atoms, but with different shapes and sizes. The plots obtained for the Vs values at the other t e m p e r a t u r e s are similar to these.

P1 ~E 4 m

o

P3 0

E o

P4

-A

r~ 2 ;>

P2 0

?

L~

0,4

.- ~

.....

9 -:43---

---4

0,45

0,5 0,55 0,6 d (nm) Figure 17. Specific retention volumes versus the molecular size, for the adsorption of organic molecules with six carbons atoms on glassy carbons (see Table 16 for molecular dimensions).

Sample P1 (Figure 17) can be seen to discriminate between the critical dimension of benzene and cyclohexane. This behaviour is the same as t h a t described above for the adsorption of these adsorbates m e a s u r e d at high relative pressures. One can conclude t h a t sample P1 has such a narrow distribution of micropores t h a t it behaves similarly for adsorption in a wide range of experimental conditions, i.e. from a very low relative pressure (at zero surface coverage) up to P/Po = 1, and from 303 K up to 533 K. The trend of Vs values for P4 is very similar to t h a t for P1. Moreover, P4 has the same discrimination, or molecular sieve effect, for the m i n i m u m critical size between benzene and cyclohexane. These data indicate t h a t P4 also has a very narrow distribution of micropores similar to t h a t of P1. P3 has a molecular sieve effect, but for a m i n i m u m critical size larger t h a n t h a t of P1 and P4. In this case this behaviour appears for the cyclohexane/2,2-DMB couple. Therefore, like for samples P1 and

426 P4, sample P3 has a similar molecular sieve behaviour as that observed for samples P1 and P4, but for larger molecules. The comparison of the adsorption of hydrocarbons to that of organic molecules capable of specific interactions can give useful information to determine the driving forces of the adsorption in each case. Once these are known it would be possible to design the properties of the carbon materials to adsorb these substances. The specific retention volumes for these molecules are compiled in Table 16.

Table 16 Specific retention volumes, Vs (cma/m 2) on glassy carbons at 473 K Vs(cm3/m 2) P4

Mean molecular size (nm)

P1

P2

P3

CC14

0.06

0.052

0.46

CHCla

0.77

0.074

1.39

0.44

0.59

CH2C12

0.66

0.112

1.13

0.51

0.57

CHaCOCH3

1.39

0.113

1.03

1.05

0.33

CH3CH2OCH2CH3

2.12

0.094

1.13

1.26

0.41

THF

0.92

0.I00

2.32

0.52

0.56

C6H14

3.29

0.302

3.84

0.84

0.405

C6H6

4.83

0.131

3.31

3.21

0.52

C6H12

0.07

0.054

2.65

0.05

0.56

2,2-DMB

0.05

0.052

0.20

0.04

0.60

0.64

Reprinted from: M. Domingo-Garcia et al. [55].

For P1 and P4 the polar molecules have Vs values higher t h a n that of cyclohexane although in many cases their minimum critical dimensions are larger or similar to that of cyclohexane. It is, therefore, likely that these molecules are adsorbed in part by chemical surface groups, i.e. by specific interactions. For P2 the values of Vs are much lower than for the other samples. This could be related to the pore constrictions in the microporosity and to the lack of macroporosity and the almost negligible mesoporosity of this sample, above discussed [55]. For P3 the trend of Vs is different to that of P1 and P4. Moreover for P3 a plot of Vs versus the mean critical size of all these molecules shows (Figure 18), with the exception of acetone and diethyl ether, a monotonical decrease of Vs as the molecular dimension increases.

427

o

C6HI4 .. C6H6 ~

t.,,,i

E

~

= THF H12

E

o > v

C3H60 9

"IxHCCI3

CaHIoO 9

CH2C129 ,

0,3

0,4

0,5

2 , 2 - D M B ~ 9 CC14

0,6

0,7

d (nm) Figure 18. Variation of Vs with the mean molecular size of the adsorbates on P3.

It is, thus, difficult to determine whether adsorption of the polar molecules is produced as a consequence of specific interactions or of the wider micropore system of this sample which allows adsorption to occur inside the micropores by a non-specific interaction. It could even result from a combined mechanism in the micropores and on the chemical surface groups. It is also a p p a r e n t t h a t the same molecular sieve behaviour found at high relative pressures for the chloroform/cyclohexane couple appears at very low vapour concentrations. The s t a n d a r d enthalpies of adsorption and the liquefaction heats of these molecules are compiled in Table 17. Comparing the -AHOA values of the hydrocarbons with their -AHL (liquefaction heats), it is noteworthy t h a t for those molecules which, according to the values of V~, can reach the microporosity (n-alkanes and benzene on P1 and P4, and n-alkanes, benzene and cyclohexane on P3) the s t a n d a r d enthalpy of adsorption is much higher in absolute value t h a n the liquefaction heat. The -AHOA values are even more t h a n twofold the -AHL value in m a n y cases. Similar results are always obtained for the adsorption of hydrocarbons from diluted atmospheres on porous carbon materials [9,10,13,15,56-66]. Thus, in all the above described cases the absolute values of -AHOA are much higher t h a n the liquefaction heats. The only cases in which the values of-AH~ and -AHL are very close is when the adsorption occurred in nonporous carbons (V3G and Pyrolitic, Acheson and Degussa graphites) or in the external surface of the porous carbons. If neither n-alkanes nor cyclohexane can be considered to be capable of specific interactions, these high values of-AH~ could be produced as a consequence of a very good fit of the molecules inside the pores [13,52,66], such t h a t the closer the size of the molecule and pore dimension the higher the absolute values of-AH~ This is a consequence of the, so called,

428 surface curvature effect (SCE) and of the proximity of the pore wall increasing the adsorption potential [55,67]. In the case of benzene it is generally accepted t h a t high absolute values of-AH~ are produced either because adsorption occurs in the slit-shaped pores with dimensions similar to the molecular size or possibly specific interactions could be taking place due to the existence of ~ electrons in this molecule [9,67,68-70]. In one of the carbon materials used (P2) the value of -AHOA for benzene is close to -AHL which means t h a t the specific contribution is probably very small.

Table 17 S t a n d a r d enthalpies of adsorption on glassy carbons -AH~176 P1

-AHL

P2

P3

2.9

3.6

36.9

CHC13

38.9

17.4

47.4

43.1

31.4

CH2C12

34.3

25.4

40.5

44.9

31.7

CH3COCH3

49.1

33.5

50.6

53.5

32.0

CH3CH2OCH2CH3

49.9

42.3

51.1

52.6

29.1

THF

44.3

23.4

59.5

49.1

29.8

n-C4H10

29.8

14.0

37.0

27.0

24.3

n-CsH12

45.3

28.5

50.7

52.4

27.6

n-C6H 14

65.2

42.1

65.7

75.9

31.7

n-CTH16

88.1

64.9

82.1

97.8

37.1

C6H6

53.5

36.3

58.8

64.1

34.1

C6H12

6.0

8.9

53.9

16.1

32.8

20.6

4.2

30.4

CC14

2,2-DMB 3.7 3.2 Reprinted from: M. Domingo-Garcia et al. [55].

P4

(kJ/mol) 31.9

With regards the polar molecules and the P1, P3 and P4 samples the -AH% values have a higher absolute value t h a n the liquefaction heats in all cases. This suggests t h a t for P1 and P4, in which Vs is not related to their molecular sizes, the adsorption of these molecules is mainly controlled by specific interactions. In P3, although the interaction could probably also be specific, an a p p a r e n t relationship between Vs and the molecular size has also been shown (Figure 18). Therefore, the high absolute values of-AH~ should be produced as a consequence of the good fit of the molecules in the pores similar to the behaviour observed for

429 hydrocarbons, although some specific contribution, i.e. a combined mechanism of two contributions: specific + non-specific, can not be excluded. W h a t is noteworthy from the comparison of-AH~ values of the hydrocarbons and of the other molecules is t h a t some of the former can be adsorbed more exothermically t h a n polar molecules capable of specific interactions. To u n d e r s t a n d these data one should bear in mind t h a t the classification of these two types of interactions was based [71] on materials which can be considered as basically non-porous. Consequently, the -AHOA values for the non-specific interactions were clearly lower in absolute value t h a n the specific ones and very close to the liquefaction heats because the former were produced on flat surfaces. However, the situation is different in the case of porous materials because adsorption can be produced on the external surface, on the chemical groups or inside the micropores of similar size to the adsorbate. The most plausible of the three possibilities is the latter since this is t h e r m o d y n a m i c a l l y favoured. It is worth noting t h a t the n u m b e r of molecules to be adsorbed is very low (zero surface coverage) hence the probability of finding pores of similar size to the molecule is very high. However, for polar molecules the interaction produced by the dipolar m o m e n t can not be excluded and consequently a combined mechanism of specific + non-specific adsorption should be considered. The data reported in Table 17 clearly indicate t h a t in porous m a t e r i a l s it is unsafe to deduce the type of adsorbate-adsorbent interaction exclusively on the basis of the -AH~ values. In other words, although there can be different driving forces in the adsorption, the non-specific interactions (see for instance -AH~ for C6H14) are not necessarily less energetic t h a n the specific ones, at least when adsorption is produced inside the micropores and at very low coverage. In the case of a combined (specific + non-specific interactions) mechanism, the specific component of the surface free energy can be determined [57,59,60,72,73]. The results are collected in Table 18 for all these molecules except for benzene and carbon tetrachloride because neither carbon tetrachloride nor benzene

Table 18 Specific component of the surface free energy -AGsp(kJ/mol) P1

P2

P3

P4

CHC13

1.4

1.3

2.3

2.5

CH2C12

6.2

5.4

6.3

6.8

14.0

8.0

10.3

13.2

CH3CH2OCH2CH3

8.0

3.5

3.8

8.4

THF

7.2

5.0

7.9

6.8

CH3COCH~

Reprinted from: M. Domingo-Garcia et al. [55].

430 produce specific interactions. This is u n d e r s t a n d a b l e for carbon tetrachloride which has no dipolar m o m e n t and can be considered a spherical molecule. Nevertheless, this behaviour is more unexpected for benzene for which specific interactions are normally expected, because of the unlocalized ~ electrons. Since this behaviour of benzene is repeated for the four samples, it can be concluded t h a t its adsorption is non-specific in all cases and the high absolute values of -AH~ (Table 17) can be explained because adsorption is produced in slit-shaped pores of a similar size as the molecule. This finding again supports the previous suggestion t h a t -AHOA is not a very useful criterion to establish w h e t h e r an interaction is specific or non-specific. Very low values of the specific component of the surface free energy can be observed for CHC13. This suggests t h a t although the adsorption m e c h a n i s m of this molecule is considered a combined mechanism, the porosity is much more i m p o r t a n t t h a n the chemical surface groups, because the specific contribution is never higher t h a n 15 % of the total value of the s t a n d a r d free energy [55]. On the other h a n d the specific contributions for the adsorption of acetone and diethyl ether r e p r e s e n t more t h a n 50% of the total value of the s t a n d a r d free energy for the former and around 40% for the latter. Thus, in the adsorption of these molecules the interactions with the chemical surface groups seem to be as i m p o r t a n t as the porosity of the samples which could explain why these molecules do not follow the general trend shown in Figure 18. The values for the other molecules are lower t h a n those of acetone and in most cases they represent less t h a n 30% of the total value. From a practical point of view, these findings are very i m p o r t a n t because they indicate t h a t the following points should be t a k e n into consideration when one a t t e m p t s to increase the capacity of adsorption of these molecules from diluted atmospheres: i) Slit-shaped pores are very convenient for the adsorption of benzene. They can be produced in carbon materials. ii) Porosity is more i m p o r t a n t for the adsorption of CHC13 t h a n the oxygen functionalities of the adsorbent. iii) The importance of the porosity and of the chemical functionalities is very similar for the adsorption of acetone and diethyl ether. From these data one can conclude t h a t these glassy carbons have very narrow micropores distributions which permit t h e m to behave similarly from zero surface coverage (very low vapour concentration) to high surface coverage. Moreover the criterion frequently used to discriminate between specific and non-specific interactions on the basis of the s t a n d a r d enthalpy of adsorption is not useful when microporous materials are used as adsorbent and when adsorption is carried out at zero surface coverage.

431 ACKNOWLEDGEMENTS

This work has been supported by the DGYCIT under project PB94-0754.

REFERENCES

1. S.J. Gregg and K. S. W. Sing, Adsorption Surface Area and Porosity, Academic Press, London 1987. 2. R.M. Bradley, Carbon, 29 (1991) 893. 3. T. Wigmans, Carbon, 27 (1989) 13. 4. B. McEnaney, Carbon, 26 (1988) 267. 5. K.S.W. Sing, Carbon, 27 (1989) 5. 6. J. Rozwadowski, J. Siedlewski and R. Wojsz, Carbon, 17 (1979) 411. 7. J. Koresh and A. Softer, J. C. S., Faraday I, 76 (1980) 2457. 8. A.N. Ainscough and D. Dollimore, Langmuir, 3 (1987) 798. 9. M. Domingo-Garcia, I. Fernfindez-Morales, F. J. L6pez-Garz6n, C. Moreno-Castilla and M. J. Prados-Ramirez, J. Colloid and Interface Sci., 136 (1990) 160. 10. I. Fernfindez-Morales, A. Guerrero-Ruiz, F. J. L6pez-Garz6n, I. Rodriguez-Ramos and C. Moreno-Castilla, Carbon, 22 (1984) 301. 11. T.G. Lamond and H. Marsh, Carbon, 1 (1963) 293. 12. M. Domingo-Garcia, I. Fernfindez-Morales, F. J. L6pez-Garz6n and C. Moreno-Castilla, Langmuir, 13 (1997) 1218. 13. M. Domingo-Garcla, I. Fernfindez-Morales, F. J. L6pez-Garz6n, C. Moreno-Castilla and M. Pyda, J. Colloid and Interface Sci., 176 (1995) 128. 14. M. Domingo-Garcia, I. Fernfindez-Morales, F. J. L6pez-Garz6n and C. Moreno-Castilla, Langmuir, 7 (1991) 339. 15. F.J. L6pez-Garz6n, C. Moreno-Castilla, A. Guerrero-Ruiz, F. Rodriguez-Reinoso and J. de D. Ldpez-Gonzalez, Ads. Sci. Technol., 1 (1984) 103. 16. T. G. Lamond, J. E. Metcalfe and P. L. Walker, Jr., Carbon, 3 (1965) 59. 17. H. Kitagawa and N. Yuki, Carbon, 19 (1981) 470. 18. E. Fitzer, K. Mueller and W. Schaeffer, Chemistry and Physics of Carbon, P. L. Walker, Jr., (ed.), vol. 7, Marcel Dekker, New York, 1971. 19. J. L. Schmitt and P. L. Walker, Jr., Carbon, 9 (1971) 791. 20. J. L. Schmitt and P. L. Walker, Jr., Carbon, I0 (1971) 87. 21. C. Moreno-Castilla, O. P. Mahajan, P. L. Walker, Jr., H. J. Jung and M. A. Vannice, Carbon, 18 (1980) 271. 22. C. Moreno-Castilla, I. Fern{mdez-Morales, M. Domingo-Garcia and F. J. L6pez-Garz6n, Chromatographia, 20 (1985) 709. 23. F. Rodriguez-Reinoso, M. Molina-Sabio and M. A. Mufiecas, J. Phys. Chem., 96 (1992) 2707. 24. Y. Matsumura, K. Yamabe and H. Takahshi, Carbon, 23 (1985) 263.

432 25. P. Davini, Carbon, 28 (1990) 565. 26. F. J. LSpez-GarzSn and M. Domingo-Garcia, Chromatographia, 21 (1986) 447. 27. F. J. LSpez-GarzSn, M. Domingo-Garcia and I. Fernandez-Morales, Chromatographia, 23 (1987) 97. 28. M. Rozwadowski, K. E. Wisniewski and R. Wojsz, Carbon, 22 (1984) 273. 29. M. Domingo-Garcia, I. Fernandez-Morales, F. J. LSpez-GarzSn and C. Moreno-Castilla, J. of Chromatography, 294 (1984) 41. 30. S. P. Nandi and P. L. Walker, Jr., Fuel, 43 (1964) 385. 31. H. Marsh, E. A. Heintz and F. Rodriguez-Reinoso (eds.), Introduction to Carbon Technologies, University of Alicante, (Spain), 1997. 32. M. M. Dubinin, J. Colloid and Interface Sci., 23 (1967) 487. 33. M. M. Dubinin, Chemistry and Physics of Carbon, P. L. Walker, Jr., (ed.), vol. 2, Marcel Dekker, New York, 1966. 34. J. Koresh, J. Colloid and Interface Sci., 88 (1982) 398. 35. A. N. Ainscough, D. Dollimore and G. R. Heal, Carbon, 11 (1973) 189. 36. H. Marsh, Carbon, 25 (1987) 49. 37. M. M. Dubinin and H. F. Stoeckli, J. Colloid and Interface Sci., 75 (1980) 34. 38. E. E, Hucke, US Patent No. 3 859 421 (1975). 39. K. Kinoshita, Carbon Electrochemical and Physicochemical Properties, Wiley, New York, 1988. 40. F. Rodriguez-Reinoso and A. Linares-Solano., Chemistry and Physics of Carbon, P. A. Thrower, (ed.), Marcel Dekker, New York, vol. 21, 1981. 41. M. Domingo-Garcia, I. Fernandez-Morales and F. J. LSpez-GarzSn, Carbon, 31 (1993) 75. 42. A. Bailey, D. A. Codenhead, D. Davies, D. H. Everett and A. Miles, Trans. Faraday Soc., 67 (1971) 23. 43. H. F. Stoeckli, F. Kraehenbuehl, A. Lavanchy and U. Huber, J. Chem. Phys., 81 (1984) 785. 44. P. J. M. Carrot, R. A. Roberts and K. S. W. Sing, Carbon, 25 (1987) 56. 45. H. F. Stoeckli, D. Hugnenin and A. Greppi, J. Chem. Soc., Faraday Trans., 89 (1993) 2055. 46. M. A. Salas-Peregrin, F. Carrasco-Marin, F. J. LSpez-GarzSn and C. Moreno-Castilla, Energy & Fuels, 8 (1994) 239. 47. K. L. Foster, R. G. Fuerman, J. Economy, S. M. Larson and R. Road, J. Chem. Mat., 4 (1992) 1068. 48. F. Carrasco-Marin, M. Domingo-Garcia, I. Fernandez-Morales and F. J. LSpez-GarzSn, Carbon, 29 (1991) 629. 49. M. Kikuchi, K. Funabashi, H. Yusa and Y. Takashima, Radiochem. Radioanal. Letters, 45 (1980) 279. 50. M.M. Kopecni, J. J. Comor, M. Todorovic, L. Vujisic and D. L. J. Vuckovic, J. Colloid and Interface Sci., 134 (1990) 376. 5 I. M. Domingo-Garcia, I. Fernandez-Morales, F. J. LSpez-GarzSn and M. Pyda, Chromatogaphia, 34 (1992) 568.

433 52. F. J. LSpez-GarzSn, M. Pyda and M. Domingo-Garcia, Langmuir, 9 (1993) 531. 53. M. Domingo-Garcia, I. Fernandez-Morales, F. J. LSpez-Garz6n and C. Moreno-Castilla, Current Topics in Colloid and Interface Sci., in press. 54. A. A. Lizzio, S. N. Vyas and M. Rostane-Abadi, Abstracts of Carbon'94, Granada (Spain), 1994, 530. 55. M. Domingo-Garcia, F. J. L6pez-Garzdn, C. Moreno-Castilla and M. Pyda, J. Phys. Chem., 101 (1997)8191. 56. E. Papirer, J. Li, H. Balard and J. Jagiello, Carbon, 29 (1991) 1135. 57. J-B. Donnet and S-J. Park, Carbon, 29 (1991) 955. 58. H. Grajek, S. Neffe and Z. Witkiewicz, J. Chromatogr., 600 (1992) 67. 59. J. Jagiello, T. J. Bandosz and J. Schwarz, Carbon, 30 (1992) 63. 60. J. Jagiello, T. J. Bandosz and J. Schwarz, J. Colloid and Interface Sci., 151 (1992) 433. 61. S. K. Milonjic, L. Djuricic and M. J. Polovina, Chromatographia, 29 (1990) 228. 62. M. C. Baquero, P. Granados, P. Rodriguez and J. M. Rincon, Coal Science, J. A. Pajares and J.M.D. Tascon (eds.), vol. 1, Elsevier, Amsterdam, 1995. 63. R. Y. Qin and J-B. Donnet, Carbon, 32 (1994) 65. 64. M. A. Montes-Mor~n, Ph. D. Thesis, Universidad de Oviedo (Spain), 1997. 65. J-B. Donnet, R. Ch. Bansal and M-J. Wang (eds.), Carbon Black, 2na edition, Marcel Dekker, New York, 1993. 66. E. G. Derouane, J-M. Andr6 and A. A. Lucas, J. Catal., 110 (1988) 58. 67. F. J. LSpez-Garz6n and M. Domingo-Garcia, Adsorption on New and Modified Inorganic Sorbents, A. Dabrowski and V. A. Tethykh (eds.), Elsevier, 1996. 68. J. H. Burness and J. G. Dillard, Langmuir, 7 (1991) 1713. 69. X. L. Cao, B. A. Colenutt and K. S. W. Sing, J. Chromatogr., 183 (1991) 555. 70. S. K. Milonjic, T. S. Ceranic and M. Dj. Petkovic, Chromatographia, 27 (1989) 306. 71. A. V. Kiselev and Y. I. Yashin, Gas Adsorption Chromatography, Plenum Press, New York, 1969. 72. C. Saint-Flour, E. Papirer, Ind. Eng. Prod. Res. Dev., 21 (1982) 666. 73. S. Dong, M. Brondle and J-B. Donnet, Chromatographia, 28 (1989) 666.

Adsorption and its Applications in Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski(Editor) 9 1998Elsevier Science B.V. All rights reserved.

435

S e l e c t i v e a n d r e v e r s i b l e a d s o r b e n t s for n i t r i c o x i d e f r o m h o t combustion gases R. Long and R.T. Yang* Department of Chemical Engineering, The University of Michigan Ann Arbor, Michigan 48109-2136, USA

This report provides an updated review and discussion of all selective, reversible sorbents for adsorption of NOx from combustion gases. The sorbents must selectively adsorb NOx over other gas molecules that are also contained in combustion gases: SO2, H20, CO2, 09 and N2.

1. I N T R O D U C T I O N Removal of NOx from exhaust gases is a challenging problem which has been extensively studied worldwide in recent years. The NOx emission is a major cause for the formation of acid rain and for other environmental problems. Selective Catalytic Reduction (SCR) of NOx has been the most effective means for NOx abatement. For large power plants, V2OJTiO2 has been the main commercial catalyst for SCR with NH3 for stationary sources [1]. However, for relatively small scale combustors, such as diesel-fueled and gasoline-fueled engines in vehicles, the use of NH3- based SCR technologies is not practical because of the high cost and NH3 slip. The three way catalyst (Pt-Rh-Pd) is an effective catalyst for SCR (mainly by CO) used in automobiles under rich-burn conditions [2,3], but it suffers from severe loss of activity for NO reduction in the presence of excess oxygen, which is the prevalent condition for diesel or lean-burn gasoline engines. SCR of NOx with hydrocarbons under excess oxygen conditions has been actively studied by many groups most recently [4-8]. A large number of catalysts have been found to be active for these reactions, such as Cu, Fe, Co, Ce, Ga, and H exchanged zeolites, noble metals supported on 7-alumina, metal oxides, pillared clay, and so on. A summary of these catalysts has been published recently by Amiridis et al. [5]. Among them, Cu- and Co-ZSM-5 are the most intensively investigated; however, they are deactivated rapidly by moisture and S02 [6-8]. Noble metal based catalysts appear to be free from deactivation by H20 and SO2, but additional problems, such as narrow window of operation * Address all correspondenceto R.T.Yang

436 temperature, high selectivity for N20 formation and oxidation of SO2 to SO3, inhibit their application in industry [5]. So successful catalyst development is necessary before this technology becomes applicable to diesel or lean-burn gasoline vehicles. A promising alternative approach for the removal of NOx is NOx trapping, or adsorption/absorption of NOx. Adsorption is divided into physical and chemical adsorption. Physical adsorption (surface adsorption and micropore filling) is rapid and reversible, but is less selective for specific gas species. The amount of physical adsorption can greatly exceed a monolayer capacity. It usually occurs as a result of intermolecular forces, such as van der Waals forces and capillary condensation. The normal boiling points of NO and NO2 are 121 K and 294 K, respectively. Therefore, NO2 can be easily condensable on microporous solids by pore filling around room temperature, whereas NO is relatively more difficult. In the presence of oxygen, NO trap by pore filling can be facilitated by the formation of NO2. Chemisorption results from the interaction between the adsorbate molecule and the adsorption site, which is selective to specific gases. The amount of chemisorption is less than the monolayer capacity. Chemisorption can occur at low or high temperatures. In order to remove NOx efficiently from exhaust gases, a very specific sorbent is needed. The sorbent must be able to selectively adsorb NOx from oxygen-rich combustion gases which contain NOx, 02, H20, SO2, CO2 and N2. The desired temperature range for NOx trapping is 300-400~ although temperatures outside this range may be prevalent depending on the specific application. The sorption rates must be high, e.g., suitable for applications at space velocities > 3,0001/h. The sorption must be reversible either by increasing temperature or decreasing pressure, so a desorption stream concentrated in NOx can be obtained [9]. The concentrated stream can be recycled to the combustion zone for NO decomposition into N2. Alternatively, desorption/decomposition can be accomplished by injecting a reducing gas. Still another alternative, applicable to lean-burn engines, is to dope noble metals in the sorbent and to run the engine with pulses of rich-burn conditions, during which time the adsorbed NOx is decomposed into Ne [10,11]. There has been a long search for such a sorbent for NOx as reviewed recently [12,13]. The more promising sorbents have been supported transition metal oxides [14-18], ZSM-5 or MFI zeolites exchanged by Cu 2§ and other cations [19, 20], FeeO~ dispersed on activated carbon fibers (AFC) [21, 22], zeolite [23], Y-Ba-Cu-O [24, 25], mixed metal oxides [10, 26-28] and carbon [31]. The two most promising sorbents, in terms of both NOx capacity and rate of uptake, appear to be Mn-Zr (1:1 molar ratio) mixed oxides [27] and Ce-doped CuO/TiOe [28], reported recently. In this paper, the available literature on adsorption of nitric oxides is reviewed, and the development of NOx removal techniques through sorption on solid materials is discussed.

437 2. NOx A D S O R P T I O N AT NEAR AMBIENT T E M P E R A T U R E

Iwamoto and coworkers [19,20] studied the adsorption of NO on various metal ion-exchanged zeolites with a fixed bed adsorption apparatus. In the adsorption experiment, 1,000 - 2,000 ppm of NO in He was introduced in a stainless steel column containing the adsorbent. After each adsorption run, pure He was introduced into the column to desorb NO from the adsorbent. The amount of reversible adsorption (Qrev) and irreversible adsorption (Qirr) of NO measured at 273K on various cationexchanged MFI zeolites are summarized in Table 1.

Table 1 NO adsorption properties of various cation-exchanged MFI zeolites Amount of NO adsorbed/(cm3g -1) Adsorbent Content of cation/(wt%) Na-MFI(23.3)- 100 Ca-MFI(23.3)-54 Sr-MFI(23.3)- 105 Ba-MFI(23.3)-80 Mg-MFI(23.3)-46 Cu-MFI(23.3)-157 Ag-MFI(23.3)-90 Co-MFI(23.3)-90 Mn-MFI(23.3)- 127 Ni-MFI(23.3)-68 Zn-MFI(23.3)-96 Fe-MFI(23.3)-62 Cr-MFI(23.3)-41 Ce-MFI(23.3)-8 La-MFI(23.3)-7 H-MFI(23.3)-100

2.81 1.32 5.45 6.44 0.69 5.90 10.85 3.06 4.20 2.41 3.79 2.12 0.87 0.43 0.40 0.13

Qrev reversible 0.16(0.006) c 1.81(0.246) 2.71(0.195) 1.50(0.143) 0.69(0.109) 4.28(0.206) 3.38(0.150) 1.52(0.131) 1.19(0.069) 1.03(0.112) 1.01(0.078) 0.52(0.061) 0.38(0.101) 0.34(0.496) 0.25(0.388) 0.12(0.004)

Qirr irreversible 00.(0.000) c 1.56(0.212) 0.20(0.014) 1.44(0.137) 0.22(0.035) 14.90(0.716) 0.54(0.024) 19.69(1.693) 5.81(0.339) 6.64(0.727) 0.50(0.039) 3.08(0.362) 1.16(0.308) 0.34(0.496) 0.24(0.372) 0.32(0.011)

Adsorption time, 45 min; desorption time, 60 min; concentration of NO, 997 ppm; adsorption temperature, 273K; adsorbent weight, 0.5 g; flow rate, 100cm3 min 1. bConcentration of NO, 1,910ppm. cUnit, (NO molecules)'(cation) -1. Reprinted from: Zhang et al. [ 19]. a

438 The Qrev and Qirr changed significantly with the metal ion. For transition metal ion-exchanged zeolites, the values of Qirr were larger t h a n those of Qrev except for Zn-MFI and Ag-MFI. In contrast, Qrev was greater t h a n Qirr on alkaline earth metal ion-exchanged zeolites. The amount of reversible adsorption per cation decreased in the order Ca 2+ > S r 2+ > B a 2+ > M g 2+ The order of Qrev was: Transition Metal Ion - Alkaline Earth Metal Ion > Rare Earth Metal Ion - Alkali Metal Ion - H § Among these zeolites Cu-MFI and Co-MFI showed the largest Qrev and Qirr, respectively. In CuMFI, Qrev and Qirr were found to be proportional to the exchange amount of copper ion, but the Qrev and Qirr per copper ion were constant. IR spectroscopy indicated that most of the reversibly adsorbed NO was NO + adsorbed on Cu 2§ and that the irreversibly adsorbed NO was in the forms of NO +, nitrate (NO~), nitrite (NO~), and NO~. The amounts of reversible and irreversible adsorption of NO were also dependent on the zeolite structure. Table 2 and Table 3 show the results of copper ion-exchanged zeolites and silver ion-exchanged zeolites, respectively.

Table 2 Effect of zeolite structure on NO adsorbability of copper ion-exchanged zeolites

Adsorbent

Content of cation/(wt%)

Amount of adsorption of NO/(cm3.g-1) Reversible

Irreversible

Cu-MFI(23.3)-68

2.63

2.29(0.247) b

7.46(0.805) b

Cu-OFF/ERI(7.7)-81

5.45

2.28(0.146)

5.55(0.270)

Cu-MOR( 10.5)-76

5.26

2.11(0.114)

6.69(0.361 )

Cu-LTL(6.0)-34

3.22

1.23(0.108)

2.38(0.210)

Cu-FER(12.3)-66

3.89

1.42(0.104)

4.82(0.353)

Cu-FAU(2.6)-60

9.27

1.15(0.035)

0.62(0.019)

Cu-FAU(5.6)-83

7.99

0.86(0.031)

1.52(0.055)

Adsorption time 45 min; desorption time, 60 min; concentration of NO, 1,910 ppm; adsorption temperature, 273 K; adsorbent weight, 0.5 g; flow rate, 100 cm 3 min-1. bUnit, (NO molecules) (cation) -1. Reprinted from: Zhang et al. [ 19].

a

439 Table 3 NO adsorption properties of various silver ion-exchanged zeolites amount of adsorption Adsorbent Content of of NO (cma'g -1) ca tio n/(wt%)

Re versib 1e

Irreversible

Ag-MFI(23.3)-104 b

12.38

6.16(0.240 c)

4.11(0.160 c)

Ag-MFI(23.3)- 104

12.38

5.14(0.200)

4.99(0.194)

Ag-MOR(15.0)-112 b

17.09

5.76(0.162)

2.18(0.061)

Ag-MOR( 15.0-112

17.09

5.02(0.141 )

2.39(0.067)

Ag-FER( 12.3)-76

13.66

4.71(0.166)

1.51(0.053)

Ag-OFF/ERI(7.7)

16.71

1.12(0.032)

0.28(0.008)

Ag-LTL(6.0)-37

13.43

0.28(0.010)

0.16(0.006)

Ag-FAU(5.6)- 101

25.76

0.40(0.007)

0.56(0.010)

Ag-FAU(2.6)-98

37.10

4.12(0.053)

3.99(0.052)

Ag-LTA(2.0)- 103

41.22

3.38(0.039)

0.57(0.007)

5.90

4.28(0.206)

14.90(0.716)

Cu-MFI(23.3)-157 b

a Adsorption time, 45 min; desorption time, 60 min; adsorption temperature 273 K. bAdsorption time 60 min; desorption time 120 min. cUnit NO-molecule (Ag ion) -1. Reprinted from: Zhang et al. [20].

The amount of reversible adsorption and irreversible adsorption of NO per copper ion were found to decrease in the following order: MFI > OFF/ERI > MOR > LTL > FER > FAU This result is consistent with the increase in the aluminum content in the zeolites. From Table 3, one can see that Qrev also changed significantly with the zeolite structure among the silver ionexchanged zeolites. Ag-MFI and Ag-MOR showed the highest Qrev. Similar to the Cu-MFI sorbent, Qrev of Ag-MFI and Ag-MOR increased with the ion exchange level, and Qrev/Ag § of Ag-MFI was constant at different exchange levels of silver ion. The real exhaust gases also contain various gases such as NO2, 02, CO2, SO2, CO and H20. Iwamoto et al. [19] further studies their influence on NO adsorption properties in Cu-MFI zeolite. The effects of each gas is shown in Table 4.

440 Table 4 Effect of preadsorbed gases on the adsorption properties of Cu-MFI(23.3) - 147 a Amount of adsorption Preadsorbed gas b

of NO/(cm3-g -1) Reversible

Irreversible

NO2(4,680 ppm)/He

7.14

2.21

02(99.5%)

4.26

14.38

CO2(20%)/He

4.25

12.19

SO2(2,170 ppm)He

3.92

7.86

CO(1,890 ppm)/He

1.39

4.15

H20(3%)/He

0.22

0.45

None

4.35

17.83

Adsorption time, 60 min; desorption time, 120 min; concentration of NO, 1,000 ppm; adsorption temperature, 273 K; adsorbent weight, 0.5 g; flow rate, 100 cm 3 min -~. bThe adsorbent was heated at 773 K for 5 h under helium stream (50 cm3"min-~) before preadsorption treatment. After the preadsorption the sample was purged with helium at room temperature. Reprinted from: Zhang et al. [19]. a

The preadsorption of NO2 resulted in an enhancement of Qrev for NO, which was probably due to the result that the irreversibly adsorbed NO2 provided new sites for NO molecules to produce N203 [19]. When 02, CO2, or SO2 were preadsorbed, almost no change in Qrev for NO was found. CO and H20 poisoned the adsorbability of NO in the Cu-MFI zeolite. The Qirr for NO was also described by the adsorption of CO and H20. Kaneko et al. [21,22,32] prepared cz-FeOOH and Fe203 highly dispersed on ACFs (active carbon fibers), which have very high adsorption capacity for NO near room temperature. For instance, the amount of NO adsorption was about 160 mg/g on cz-FeOOH dispersed ACF at 303 K and 80 kPa NO pressure. This type of NO adsorption seems to have both chemisorption and physical adsorption characteristics. NO was mainly adsorbed by a micropore-filling mechanism, which was inferred by the measurement of pore volume using N2 adsorption on ACF-5 after exposure to NO. The highly dispersed cz-FeOOH particles on the ACF assisted the micropore filling of NO through their chemisorption action. A 13 X molecular sieve was also reported to have a high capacity for NO adsorption in the presence of 02 by micropore filling, the NO capacity reached 75 mg/g at 296 K [23,36]. However, HeO, SO2, COz, all have strong inhibiting effects on this sorbent.

441 Several other research groups also investigated the adsorption of NO on metal oxides. The results are summarized in Table 5.

Table 5 Amounts of NO adsorbed on various adsorbents near room t e m p e r a t u r e Adsorbent

Amount of NO adsorbed a (rag/g)

Ref.

SnO2

5

29

CeO2

5

30

NiO

1

15

Co304

6

15

CuO/7 A1203

36

15

NiO/7 A1203

36

15

Co304/~ A1203

7

15

Fe2OJ7 A1203

45

15

Fe2OJSiO2

5

16

Fe-Y zeolite

18

17

Fe304

22

18

Fe304

24

14

a- Fe203

9

14

Jaosites

3-10

33

10-20

34

4-6

35

13

12

a-FeOOH [3-, 7- FeOOH r

a At 13 kPa and room temperature. Reprinted from: Kaneko and Inouye [ 12].

It is noted t h a t Fe304 and (z-FeOOH showed the highest adsorption capacity for NO at room t e m p e r a t u r e among pure oxides. When oxides such as CuO, NiO, Fe203 were supported on 7-A1203, the adsorption of NO was high. This was due to the high surface area [12].

442

Q

NO A D S O R P T I O N AT HIGHER T E M P E R A T U R E S FROM COMBUSTION GASES

Some sorbents exhibited excellent adsorption capacities for NO at high temperatures [10,11,27,28]. These sorbents are attractive candidates for application to the automobile industry as well as the power industry. The main advantage is that the adsorbed NO can be directly converted to N2 on a catalyst (e.g., the three-way catalyst) under a "rich-burn" condition at the same room temperature. This can be accomplished by a cyclic operation [12]. Another advantage is that the sorbents have the high sorption rate at high temperature, which is suitable for application for the removal of NO from exhaust gases at a high space velocity.

3.1. S u p e r c o n d u c t i n g b a r i u m and y i t t r i u m - c o n t a i n i n g s o r b e n t s Misono et. al. [24] reported that NO and CO could be rapidly adsorbed into superconducting YBa2Cu30~. After pre-evacuation at 300~ the sample adsorbed approximately 2 mol/mol oxide for NO at the same temperature. The adsorbed NO molecules were almost completely desorbed when the temperature was increased to 400~ For these Y-containing oxides, Yamashita and coworkers found that the NO adsorptivity decreased according to the order [37]: YSr2Co3Ox > YBa4CosOx > YSr2Mn.3Ox > YSr2V3Ox. TPD and IR results showed that the adsorbed NO molecules were oxidized to NO~ by lattice oxygen. The adsorbed NO was desorbed as a mixture of NO/O2. Ba-Cu-O mixed oxides have also been reported by Arai et al. [38] to have a high adsorption capacity for NO/N02 at 200~ This adsorption reaction was accelerated by the presence of oxygen. XRD results indicated the formation of Ba(NO3)2/CuO. In the presence of 02, a large amount of NOx was liberated from the sample at temperatures above 500~ However, the NO adsorption capacity for this sorbent was completely vanished by the presence of 8% CO2 because of the formation of surface BaCO3.

3.2. Mixed m e t a l o x i d e s Since the sorbents containing Ba are easily deactivated by CO2, Arai and coworkers developed materials for NOx adsorption which did not contain rare earth and alkaline earth metals. Several mixed-oxide sorbents containing Mn and/or Zr are shown in Table 6 [27]. The uptake of NO in the presence of 02 or absence of 02 was measured at 200~ in a tubular reactor. The presence of 02 promoted the adsorption of NO on these oxides. The Mn-Fe, Mn-Zr, and Mn-Cu systems exhibited a high uptake of NO. The Mn-Zr oxide showed the highest uptake of NO both with and without 02. NO was hardly detected in the effluent gas from the fixed bed adsorption during the initial 60 minutes. After that, the concentration of NO at the outlet

443 gradually increased with time. The Mn-Zr ( 1:1 ) oxide did not show NO removal after 6 hours on-stream. The total a m o u n t of NO removal in 6 hours of operation was 0.133 mol-NO/mol-Zr. In order to compare the removal capacity m e a s u r e d at a fixed gas-phase NO concentration, a gravimetric analysis was also performed for the NO uptake. The a m o u n t of NO adsorption in Mn-Zr oxide (Mn/Zr = 1) after s a t u r a t i o n was 1.43 wt % of the original oxide, which corresponded to 0.047 mol-NO/mol-Zr. The NO adsorption a m o u n t m e a s u r e d by the gravimetric analysis was s o m e w h a t smaller t h a n t h a t from the t u b u l a r reactor by gas phase analysis. The difference from these two m e a s u r e m e n t s was likely due to the diffusion resistance. The gas was forced to pass through the oxide particles in the t u b u l a r reactor, which enhance the diffusion rates.

Table 6 NO removal by mixed oxides containing Mn and/or Zr NO removala(%) Oxide

0% 02

10% 02

MnOx'AlzOa

9.9

14.8

MnOx'Cr202

0

2.2

MnOx'CuO

0

10.2

MnOx-FezO3

11.9

37.4

MnOx'Mo03

5.9

15.2

MnOx-TiOz

0

25.0

MnOx ZrOz

100.0

100.0

ZrO2"AlzO3

0

8.7

ZrO2"CrzO3

0

14.3

ZrO2"CuO

2.7

27.0

ZrOz'FezO3

0.4

8.7

ZrOz'MoO3

0.3

2.4

ZrO2"TiOz

7.6

17.4

Note. Calcination temperature 450~

temperature 200~ W/F =1 g's.cm 3 aNO removal after 30 rain of use. Reprinted from: Eguchi et al. [27].

0.1 vol.% NO, 0 or 10% 02, He balance. Reaction

444 Arai et al. [27] also studied the sorption capacities of NO in the Mn-Zr oxides with different Mn/Zr molar ratios. The results are shown in Table 7.

Table 7 Capacity of Mn-Zr Oxides for NO Removal Sample

Capacity for NO removal mol/mol-Zr

Mn-Zr oxide (Mn/Zr = 5)

0.105

Mn-Zr oxide (Mn/Zr =1)

0.133

Mn-Zr oxide (Mn/Zr =1/5)

0.034

Mn-Zr oxide (Mn/Zr =1/9)

0.029

1 wt% Pt/Mn-Zr oxide (Mn/Zr =1)

0.058

1 wt% Rh/Mn-Zr oxide (Mn/Zr =1)

0.035

1 wt% Ru/Mn-Zr oxide (Mn/Zr =1)

0.042

1 wt% Pd/Mn-Zr oxide (Mn-Zr =1)

0.073

900 ppm NO, 10% 02, He balance. Reaction W/F -1 g s cm-3. Reprinted from: Eguchi et al. [27].

Note. Calcination temperature 450~

temperature 200~

The NO removal was 100% for every Mn-Zr oxide at the start of NO supply and then gradually decreased after 10 to 60 minutes of operation. The adsorption capacity of NO was the largest in the Mn-Zr oxide with Mn/Zr = 1. When decreasing or increasing the Mn/Zr ratio, the amount of NO adsorption decreased. Addition of noble metals also decreased the adsorption capacity of NO, as shown in Table 7. The XRD data indicated that the amorphous phase in the Mn-Zr oxides (Mn/Zr = 1), which had a large surface area, was especially active for NO removal. The adsorbed NO in Mn-Zr oxides was almost completely desorbed under an NO/O2/He atmosphere when the temperature was increased to 400~ indicating that the sorption and desorption was almost reversible. The amount of NO uptake in the Mn-Zr oxides was hardly affected by the presence of CO2 (10%). The NO removal in the first 130 minutes was only slightly affected by H20 and the total amount of uptake was enhanced with H20. The reason of the promoting effect of H20 is not understood [27]. 3.3. CuO - based

sorbents

More recently, the sorbents CuO/Ti02 and Ce-CuO/Ti02, which showed higher adsorption capacities than Mn-Zr oxides, have been prepared in our laboratory [28]. The CuO/Ti02 sample was prepared by using incipient wetness

445 impregnation with aqueous solution of Cu(NO3)2 on TiO2. The Ce dopant was added also by the incipient wetness procedure using aqueous cerium nitrate solution on the CuO/TiO2 sample. The adsorption/desorption experiments of NO were performed in a thermogravimetric analyzer, equipped with a p r o g r a m m e d t e m p e r a t u r e control unit. The results of NO2 uptake at 300~ on CuO/TiO2 and Ce-CuO/TiO2 sorbents are shown in Figure 1 and Figure 2, respectively.

...

o

o

< m

9

f

9 r~

[NO] = 2000 ppm =4%

o

.,..~

9 Adsorption at 300~ 9 Desorptlon at 450 C

9

9

< z •

2:

O

0~,,,,,,,,,,,,,,,,,,,,,,,I,,,,,,,I 0

20

40

60

80

100

L

120

140

160

Time (minute) Figure 1. Adsorption and desorption of NO2 in 5% CuO/TiO2. Desorption was achieved by heating to 450~ in 2 min in the same gas flow. Reprinted from: Li et al. [28]. It can be seen t h a t the m a x i m u m NO adsorption on CuO/TiO2 was 6 mg/g under the conditions NO = 2,000 ppm, 02 = 4%, balance = He. When CeO2 was added to CuO/TiO2, both the NO2 capacity and the uptake rate increased. The sorbent capacity was increased from approximately 6 to 7.7 mg/g, or approximately a 30% increase. The initial sorption rate was increased from 3.4 to 5.0 mg/g, both at t = 10 min, or a 50% increase. The chemisorption rate was clearly the controlling step in the uptake, and the CeO2 dopant substantially increased the chemisorption rate. Since the oxidation of NO is involved in the chemisorption of NO in the presence of 02, the significant increase in both chemisorption rate and sorption a m o u n t as a result of adding Ce to CuO/TiO2 was a t t r i b u t e d to the unique oxygen storage property as well as the redox

446 property of Ce. Results of desorption of NOx at 450~ over the two sorbents, in the same gas flow, are also shown in Figure 1 and Figure 2. Heating from 300~ to 450~ took approximately 2 minutes, during which time a small amount of NOx was desorbed. The results showed that rapid desorption was accomplished at 450~ The working capacity of the sorbent, i.e., the reversible amount, depends on the time of desorption. It is clear from Figure 1 and Figure 2 that well over 95% of the amount adsorbed was desorbed rapidly. 8

6

=z

4

<

o• Z

0 0

20

40

60

80

100

Time (minute) Figure 2. Adsorption and desorption of NOx on Ce-CuO/TiO2 (2% Ce, 5% CuO by wt). Desorption was achieved by heating to 450~ in 2 min in the same gas flow. Reprinted from: Li et al. [28].

To make a direct comparison with MnOx/ZrO2 (1:1 molar ratio) sorbent studied by Eguchi et al. [27], the Ce-CuO/TiO2 was subjected to the same sorption conditions, i.e. 1,000 ppm of NO and 200~ A lower 02 concentration of 4% was used, however. The lower 02 concentration would only lower the NOx sorption rate and capacity [27], hence, the comparison was a conservative one for our sorbent. The results with the two different sorbents are compared in Figure 3. The comparison showed that both sorption rate and capacity were significantly higher for the Ce-CuO/TiO2 sorbent than the Mn/Zr oxides. The initial uptake rate was more than doubled with our sorbent, while the final capacity was higher by approximately 15%. The final NOx capacity for the Ce-CuO/TiO2 sorbent corresponded to approximately 12 A2/NO2 or NO~.

447

18 ,.Q o rar

16-1412-

O

<

10-

9 o,..~ 8O

[NO] = 1000 ppm

6-

II

.~

4- [ /

or~ < o• Z

2-

4%

9 Adsorption at 200 C 9 Desorptlon at 450 C 9 K. Eguchi at al's Data o

10-, 0

,

,

,

i

20

,

,

i

i

40

i

,

,l

,

60

,,

l'ii

80

,'

i

,'i

100

,

i

120

,'

,'

i'

~' i"

140

,

160

Time (minute) Figure 3. Adsorption and desorption (in the same gas flow) on Ce-CuO/TiO2 (2% Ce, 5% CuO by wt) compared with the MnO/ZrO2 (1 : 1 molar ratio) sorbent of Eguchi et al. [27] under the same conditions except 02 = 10% in their work (which yielded larger adsorption than 4% 02). Reprinted from: Li et al. [28]. The effects of CO2, H20 and SO2 in NOx adsorption were also investigated for the CeCuO/TiO2 sorbent at 200~ The sorbent was first exposed to a gas flow containing 13% CO2 and 4% 02 (in He). No uptake or weight change was observed. However, when 13% CO2 was added to the gas flow containing NO (1,000 ppm) and 02 (4%), changes in both NOx sorption rate and capacity were observed. The NOx sorption was changed by CO2 in two ways. First, a small, but clear, decrease in the initial sorption rate was observed. Second, the NOx capacity increased by over 25% due to the presence of CO2. Since the uptake rates were measured only by weight gain, the surface species were not known. No changes were observed in the desorption due to CO2. The adsorption of H20 was first measured with 2.7% H20 in He at 200~ the process was performed after 90 minutes, then a flow gas containing 2.7% H20, 1,000 ppm of NO, and 4% 02 was introduced. The presorbed H20 did cause a reduction in the NOx capacity. The final amount of NOx adsorbed on the H20 presorbed sample was approximately 72% of t h a t without H20 presorption. However, the rapid uptake of NOx on the H20 -presaturated sorbent was similar to that without H20 presorption. The effect of H20 was also investigated by reversing the sequence of adsorption. In

448 this experiment, the Ce-CuO/TiO2 sorbent was first exposed to 1,000 ppm of NO and 4% 02, followed by the addition of 2.7% H20 in the mixture. The result showed that NOx and H20 adsorbed nearly independently; that is, they adsorbed on different sites. When the sample coadsorbed NOx and H20 was heated rapidly to 450~ in the same gas flow that contained H20, NO, and 02, both H20 and NOx desorbed rapidly. Therefore, the Ce-CuO/TiO2 sorbent coadsorbed H20 and NOx both reversibly. When the Ce-CuO/TiO2 sorbent was first exposed to SO2 (1,500 p p m ) + 02, an uptake of approximately 2 mg/g was observed in 90 minutes. Then NO was introduced to the gas flow, a rapid uptake of NOx was observed. However, the NO adsorption was reduced by the adsorbed SO2. The NOx capacity was reduced by approximately 20%. Desorption was accomplished by rapid heating (in 2 min) to 450~ in NO/SO2/O2/He. A significant amount of adsorbate was not desorbed at 450~ This amount was 2 mg/g, equal to that of the preadsorbed S02/02. It is known that the Ti02 surface is sulfated by SO2 + 02 in the temperature range of this study and that the surface sulfate can not be desorbed at 450~ [39]. It seems likely that the irreversible uptake was due to surface sulfate on Ti02, whereas the NOx was bonded to Cu 2§ sites. An intriguing NOx abatement process is adsorption/desorption and recycle of high concentration NOx into the combustion chamber. An existing process designed for the cleanup of coal and oil combustion flue gas containing both SOx and NOx has been developed as the NOxSO process [40-42]. In this process, a CuO/A1203 adsorbent is circulated between a fluidized bed adsorber where it adsorbs and absorbs SOx and NOx, a fluidized bed heater where NOx desorbs and is catalytically decomposed into N2, a regenerator where strongly bound SOx is removed by the reaction with H2 or CH4, and a final fluidized bed cooler before returning to the adsorber. In the NOxSO process, the NOx recycle stream contains about 3,000 ppm NOx and is returned to the primary combustion stage of the combustor. About 70% of NOx can be converted into N2 and 02. The process can be performed because this concentration is higher than the thermodynamic limit of NOx formation at the temperature of the combustion (about 2,000 ppm) and the reaction between NOx and fuel can take place in the fuel rich primary combustion. The adsorption/recycle scheme may be applied to the removal of NOx in the exhaust gas of vehicles. However, the corrosion caused by the relatively high accumulation of NOx in the presence of water vapor must be overcome before this application is possible. An interesting process involving cyclic adsorption/reduction has been investigated for automotive applications [10]. It is known that the three-way catalyst is very effective to remove NO from the exhaust gas from vehicles which operate with a stoichiometric mixture of air and fuel. However, it is substantially less effective in reducing the NOx emissions in lean-burn exhaust gases. Therefore, the technology may be feasible by first storing NOx in the solid sorbent in the oxidized form during the crusing of an automobile in lean-burn condition, then the stored NOx is released and reduced by the three-way catalyst when the atmosphere is changed intermittently to the stoichiometric air/fuel ratio. Arai et

449 al. [27] reported that the adsorbed/absorbed NOx was more easily desorbed in the reducing atmosphere than in the oxidizing atmosphere for the MnOx/ZrO2 (1:1 molar ratio) oxides. Brogan and coworkers [11] reported that, with a monolith substrate absorber combining with a three-way catalyst, a cycle of 30 s lean-burn and 30 s burn gave an average 94% NOx conversion on a prototype lean-burn engine.

3.4. Cu 2+- e x c h a n g e d pillared clays Pillared interlayered clays (PILCs) and their ion-exchanged forms have been studied extensively in our laboratory both as gas sorbents and as catalysts for the SCR reaction. At the room temperature, PILC adsorbs NO much more strongly than CO2, and adsorbs H20 only weakly [43]. For the SCR reaction, we have found that PILCs have higher activities (as well as poison resistance) than the commercial SCR catalysts by NH3 [44-46] and by hydrocarbon [47]. In particular, Cu(II) ion-exchanged pillared clay is approximately five (5) times more active than the Cu(II) ZSM-5 catalyst for C2H4 - SCR of NO at 300~ [47]. More importantly, the Cu(II)-PILC is not deactivated by H20 and SO2 (while Cu(II)ZSM-5 is severely deactivated by H20). Results of TPD experiments for NO adsorbed on Cu(II)-PILC showed that significant amounts of NOx remained on the Cu(II)-PILC surfaces at 400~ Also, in comparing different pillared clays (all ion-exchanged with Cu(II)), we have found that the HC-SCR activity is approximately proportional to the amount of chemisorbed NOx. The most active PILC (and the PILC with the highest NOx chemisorption capacity) is Cu(II) exchanged A1203 pillared clay (while ZrO2-PILC has the least activity). Moreover, chemisorption of NO on Cu(II)-PILC is assisted by O2, apparently because NO2 is the adsorbed species. Our results indicate that the Cu(II) ion-exchanged pillared clays should have high NOx chemisorption capacities at 300~176 Therefore, these sorbents are also promising sorbents for future applications. 3.5. S u l f a t e d m e t a l o x i d e s Since combustion gases usually contain various amounts of SO2, the surfaces of metal oxides are sulfated by SO2 and 02. Hence it is important to consider the sulfated forms of metal oxides for their NOx sorption. The CuO/A1203 for simultaneous SOJNOx removal is already discussed in the foregoing. The sulfated oxides can have high acidities. Much research has been done to characterize these sulfates, including "superacids". An interesting sulfated oxide is TiO2. In our work on V2OJTiO2 for NH3 SCR, it was seen that NO, at 1,000 ppm or below, does not chemisorb on V20~ at temperatures above 300~ It does, however, chemisorb on TiO2 at substantially higher temperatures, e.g., above 400~ We have also found that the sulfated TiOe surface can also chemisorb NOx at high temperatures, e.g., above 400~ [48]. The sulfated TiOz was formed by contacting TiO2 to SOJO2 (as seen in flue gas) at 300 ~ 400~ where the TiO2 surface was sulfated. Thus, the surface of TiO2 will always remain sulfated in the combustion gas.

450

F i g u r e 4 s h o w s the N O x u p t a k e a n d d e s o r p t i o n by TiOz (from d e n s i f i c a t i o n of D e g u s s a P -25 TiO2 powder, B E T surface a r e a - 50 m2/g. F i g u r e 5 s h o w s the NOx u p t a k e a n d d e s o r p t i o n by s u l f a t e d TiO2. E(D

12

g.4

O

10=

8

O

< =

6

,.o 9

o C~

4

E~

2

[~ 41/~ 9 Adsorpt!on at

9 r,~

< d z

5

9 Desorption at 450~

0

T

0

'

'

'

I

60

'

'

'

I

120

'

'

'

I

'

180

'

I

'

'

'

240

300

Time (minute) Figure 4. Adsorption/Desorption of NOx on Degussa titania (without sulfation treatment), both in the same gas atmosphere. E

8-

..Q O

6Z 0

< =

4-

O r~

C3 "~ O o,...,

9

2-

t 9 Adsorption[02] =4% at 300~ 9 Desorption at 450~

<

oZ•

0~,,,i,,,i,,,i,,,i,,, 0 30 60 90

'''1

'''

I ' ' '

I

120 150 180 210 240

Time (minute) Figure 5. Adsorption/Desorption of NOx on sulfated Degussa titania, both in the same gas atmosphere.

451 It is seen that the unsulfated form adsorbed more NOx; however, most of the NOx was irreversibly adsorbed. The opposite result was seen after the TiO2 was sulfated, where most of the NOx was reversibly adsorbed, which is a desirable property. IR results [49] showed that the NOx adsorbed in the form of bidendate NO3 ions possibly bonded to Ti sites. Thus the TiO2 surface was modified by SO~- ions possibly by electron transfer which weakened the bond between the neighboring Ti site and NO~.

3.6. Heteropoly compounds and polymeric sorbents It has been found in our laboratory that heteropoly compounds, e.g., H3PW12040" 6H20 have the unique selectivity for absorption of NOx from a simulated combustion gas [50,51]. Upon rapid heating, a large fraction of the absorbed NO evolves as N2. In our work, the NO absorption temperature was limited to approximately 230~ In our search for a sorbent for SO2 removal from combustion gases, it was found that styrenic polymeric sorbents (commercially available, with surface areas exceeding 1,000 m2/g) had high selectivities for SO2 over CO2 and H20 [52]. We have also found that the polymeric sorbents had even higher selectivities toward NOx t h a n SO2 and are extremely hydrophobic. For example, over 100 mg/g of NO was chemisorbed at the room temperature from a simulated flue gas. The chemisorbed NO forms monomers or dimers on the benzene rings on the surface of the polymer, possibly by ~-complex bonding with the ~-electrons. The bonding is strong enough to w a r r a n t an investigation of the polymeric sorbents for high t e m p e r a t u r e NOx trapping. However, the thermal stability temperatures for the styrenic polymers are only slightly above 200~ which will be the temperature limits. High surface area polymeric sorbents other than styrenic types are also available, such as the acrylic types. These sorbents also have large amounts of ~-electrons on the surface, and some of them have higher thermal stabilities. These polymers have a high potential as selective sorbents for NOx from combustion gases. However, oxidation should be considered before the polymers can be used.

ACKNOWLEDGEMENTS This work was supported by NSF CTS-9520328 and DOE DE-FG-22-96PC96206.

REFERENCES 1. H. Bosch and F. Janssen, Catal. Today, 2 (1988) 369. 2. K.C. Taylor, Catal. Rev.-Sci. Eng., 35 (1993) 457.

452 3. 4. 5. 6. 7. 8. 9.

M.M. Zwinkels, S.G. J~ir~i and G. Menon, Catal. Rev.-Sci. Eng., 35 (1993) 319. M. Shelef, Chem. Rev., 95 (1995) 209. M.D. Amiridis, T. Zhang and R.J. Farrauto, Appl. Catal. B.,10 (1996) 203. Y. Li, P.J. Battavio and J.N. Armor, J. Catal.,142 (1993) 561. M. Iwamoto and N. Mizuno, Proc. Inst. Mech. Eng., Part D, 207 (1993) 23. Y. Li and J.N. Armor, Appl. Catal. B., 5 (1995) L257. R.T., Yang, Gas Separation by Adsorption Processes, Butterworth: Boston, Chapter 5, 1987. 10.M.S. Brogan, R.J. Brisley, A.P. Walker, D.E. Webster, W. Boegner, N.P. Fekete, M. Kramer, B. Krutzsch and D. Woigtlander, SAE Tech. Pap., No. 952490,1995. 11. W. Bogner, M. Kramer, B. Krutzsh Pischingner, D. Voigtlander, G. Wenninger, F. Wirbeleit, M.S. Brogan, R.J. Brisley and D.E. Webster, Appl. Catal B., 7 (1995) 153. 12.K. Kaneko and K. Inouye, Adsorp. Sci. Technol., 5 (1988) 239. 13. H. Arai and M. Machida, Catal. Today, 22 (1994) 239. 14.K. Otto and M. Shelef, J. Catal., 18 (1970) 184. 15.H.C. Yao and M. Shelef, in: Catalytic Chemistry of Nitrogen Oxides, R.L. Klimisch and J.G. Larson (eds.), Plenum, New York (1978) 45. 16. K. Segawa, Y. Chen, J.E. Kubsh, W.N. Delgass, J.A. Dumesic and W.K. Hall, J. Catal., 76 (1982) 112. 17. S. Yuen, Y. Chen, J.E. Kubsh, J.A. Dumesic, N. Topsoe, H. Topsoe, J. Phys. Chem., 86 (1982) 3022. 18. C.R.F. Lund, J.J., Schorfheide and J.A. Dumesic, J. Catal., 57 (1979) 105. 19. W.X. Zhang, H. Yahiro, N. Mizuno, J. Izumi and M. Iwamoto, Langmuir, 9 (1993) 2337. 20. W.X. Zhang, H. Yahiro, N. Mizuno, J. Izumi, M. Iwamoto and J. Izumi, J. Mater. Sci. Lett., 12 (1993) 1197. 21.K. Kaneko, Langmuir, 3 (1987) 357. 22.K. Kaneko, Colloid Surf., 37 (1989) 879. 23. W. Joithe, A.T. Bell and S. Lynn, Ind. Eng. Chem. Process Res. Dev., 11 (1972) 434. 24.K. Tabata, H. Fukuda, S. Kohiki and M. Misono, Chem. Lett., (1988) 799. 25. N. Mizuno, M. Yamato and M. Misono, J. Chem. Soc., Chem. Commun., (1988) 887. 26.M. Machida, K. Yasuoko, K. Eguchi and H.O. Arai, J. Chem. Soc., Chem. Commun., (1990) 1165. 27.K. Eguchi, M. Watabe, S. Ogata and H. Arai, J. Catal.,158 (1996) 420. 28. W.B. Li, R.T.Yang, K. Krist and J.R. Regalbuto, Energy & Fuels, l l (1997) 428. 29.F.Solimosi, J. Kiss, J. Catal., 41 (1976) 202. 30.M. Niwa, Y. Furukawa and Y. Murakami, J. Colloid Interface Sci., 86 (1982) 260. 31.A.M. Rubel and J.M. Stencil, Energy Fuels,10 (1996) 704.

453 32.K. Kaneko, N. Fukuzaki and S. Ozeki, J. Chem. Phys., 87 (1987) 776. 33.K. Inouye, I. Nagumo, K. Kaneko and T. Ishikawa, Z. Phys. Chem.,131 (1982) 199. 34.K. Kaneko and K. Inouye, J. Chem. Tech. Biotech., 37 (1987) 11. 35. T. Hattori, K. Kaneko, T. Ishikawa and K. Inouye, Nippon Kagaku Kaishi, (I 979) 423. 36. D.P. Valenzuela and A.L. Myers, Adsorption Equilibrium Handbook, Prentice Hall, Englewood Cliffs, NJ (1989). 37.M. Kishida, T. Tachi, A. Kato, H. Yamashita and H. Miyadera, Shokubai, 33 (1991) 163. 38.M. Machida, S. Ogata, K. Yasuoka, K. Eguchi and H. Arai, in: Proc. of the I0 th Int. Congress on Catalysis, L. Guczi, F. Solymosi, P. T~t~nyi (eds.), Akad~miai KiadS, Budapest, (1992) 2645. 39.K.S. Yoo, S.D. Kim and B.B. Park, Ind. Eng. Chem. Res., 33 (1994) 1786. 40. Y.T. Yeh, R.J. Demski, J.P. Strakey and J.I. Joubert, Environ. Prog., 4 (1985) 223. 41.J.L. Haslbeck, W.T. Ma and L.G. Neal, NOxSO report for DOE contract No. DE-FC22-85PC81503, June 1988. 42. G. Centi and S. Perathoner, Ind. Eng. Chem. Res., 36 (1997) 2945. 43.R.T.Yang and M.S.A. Baksh, AICHE J., 37 (1991) 679. 44.R.T. Yang and J.E. Cichanowicz, US Patent No. 5 415 850 (1995). 45a. R.T. Yang, J.P. Chen, E.S. Kikkinides, L.S. Cheng and J.E. Cichanowicz, Ind. Eng. Chem. Res., 31 (1992) 1440. 45b. J.P. Chen, M.C. Hausladen and R.T. Yang, J. Catal., 15 (1995) 135. 46. L.S. Cheng, R.T. Yang and N. Chen, J. Catal., 164 (1996) 70. 47. R.T. Yang and W. Li, J. Catalysis, 155 (1995) 414. 48. J.P. Chen and R.T. Yang, J. Catal., 139 (1993) 277. 49. R.T. Yang and W. Li, Appl. Catal., in press (1997). 50. R.T. Yang and N. Chen, Ind. Eng. Chem. Res., 33 (1994) 825. 51. N. Chen and R.T. Yang, J. Catal., 157 (1995) 76. 52. E.S. Kikkinides and R.T. Yang, Ind. Eng. Chem. Res., 32 (1993) 2365.

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998Elsevier Science B.V. All rights reserved.

455

Adsorption processes in spacecraft environmental control and life support systems L. A. D a l l B a u m a n a and J. E. Finn b a NASA Johnson Space Center, Houston TX*, USA b NASA Ames Research Center, Moffett Field CA, USA

The environmental control and life support system on a spacecraft maintains a safe and comfortable environment in which the crew can live and work by supplying oxygen and w a t e r and by removing carbon dioxide, w a t e r vapor, and trace contaminants from cabin air. Although open-loop systems have been used successfully in the past for short-duration missions, the economics of current and future long-duration missions in space will make nearly complete recycling of air and water imperative. A variety of operations will be necessary to achieve the goal of nearly complete recycling. These include separation and reduction of carbon dioxide, removal of trace gas-phase contaminants, recovery and purification of humidity condensate, purification and polishing of w a s t e w a t e r streams, and others. Several of these can be performed totally or in part by adsorption processes. These processes are good candidates to perform separations and purifications in space due to their gravity independence, high reliability, relatively high energy efficiency, design flexibility, technological maturity, and regenerative nature. For these reasons, adsorption has historically played a key role in life support on U.S. and Russian piloted spacecraft. Among the life support applications t h a t can be achieved through use of adsorption technology are removal of trace contaminants and carbon dioxide from cabin air and recovery of potable water from waste streams. In each of these cases, adsorption technology has been selected for use onboard the I n t e r n a t i o n a l Space Station. The requirements, science, and h a r d w a r e for these applications are discussed. H u m a n space exploration may eventually lead to construction of planetary habitats. These h a b i t a t s may provide additional opportunities for use of adsorption This manuscriptwas prepared while L. DallBaumanwas employedat Johnson Space Center. She has sincejoined AlliedSignal Inc., 50 E. AlgonquinRd., Des Plaines IL 60017-5016USA.

456 processes, such as control of greenhouse gas composition, and may have different resources available to them, such as gases present in the planetary atmosphere. Separation and purification processes based on adsorption can be expected to continue to fulfill environmental control and life support needs on future missions. 1.

INTRODUCTION

A spacecraft's environmental control and life support system (ECLSS) provides a breathable atmosphere by supplying 02 and by controlling CO2, water vapor, and trace contaminant levels in cabin air. The system must also supply the crew with water for drinking, food preparation, and hygiene use. Although operating in space offers unique challenges, the ECLSS is essentially a network of unit operations in which chemical reactions, separation processes, and heat transfer play important roles. An understanding of chemical engineering principles and operations is therefore essential to the design process. Technologies commonly used in the chemical process industries, such as distillation and adsorption, have been tailored for use in space. A variety of factors drive ECLSS design. First and foremost, equipment must be safe, reliable and effective in reduced gravity. Certain physical phenomena are strongly gravity-dependent; for example, separation of vapor and liquid phases is considerably more difficult in space than on earth. This increases the complexity associated with any operation in which two-phase flow, condensation, or boiling occurs. Due to the tight constraints placed on spacecraft and system mass and volume, hardware must be lightweight and compact and the use of consumables (e.g., reagents) and expendables (e.g., non-regenerable sorbents) must be minimized. Limited power is available, so equipment must be energy-efficient. In addition, process equipment should not require frequent maintenance or a large stock of spare parts. Determining the best system for a given mission is essentially an optimization problem. The objective is to provide the crew with a breathable atmosphere and potable water while minimizing mass, volume, and power. The relative weights of the constraints are determined by specific mission characteristics; for example, power is the dominant driver for the International Space Station (ISS). For very short missions, an "open-loop" system is generally the best available solution. In such a system, all water and 02 needed for the mission are stowed onboard prior to launch while CO2 and wastewater are discarded or stored. No effort is made to recover useful substances from waste products or to regenerate materials. Power requirements are relatively low and mass and volume requirements can be reduced if waste products are discarded rather than being stored.

457 As mission length and/or crew size increases, use of an open-loop scheme becomes less practical. The problem of supplying water for the crew can be used to illustrate this point. For advanced missions (i.e., missions beyond ISS) the National Aeronautics and Space Administration (NASA) has allocated approximately 23 kg (50 lb) of water per person per day to meet all needs: drinking, food preparation, personal hygiene, and laundry. This is a meager allowance by terrestrial standards; a typical household uses 265 kg (580 lb) per person per day, with up to 95 kg (200 lb) being used for a single shower [1]. However, at a launch cost of thousands of dollars per kilogram, supplying all water at launch quickly becomes prohibitively expensive. In addition, the storage volume available on a spacecraft is limited and does not allow transportation of large quantities of water. On long missions, the mass and volume associated with the process equipment in a regenerative ECLSS can be less t h a n the mass and volume of the 02 and water required by an open-loop configuration. The first step toward closing the ECLSS loop has historically been to reduce reliance on expendable materials. For example, CO2 was originally removed from the atmosphere on all NASA space shuttles by irreversible absorption on LiOH. The two shuttles designated as extended duration orbiters have been retrofitted with a cyclic CO2 removal subsystem in which a solid amine material is used to reversibly sorb CO2. At the end of each half-cycle, the C02 is desorbed to space vacuum so that the amine can be reused. Mass of the LiOHbased subsystem increases linearly with crew size and mission duration, but the mass of the solid amine subsystem is fixed for a crew of up to seven and is essentially independent of mission duration. For a seven-person crew on an eighteen-day mission, replacing LiOH with solid amine and making the necessary h a r d w a r e changes reduces the effective mass of the CO2 removal subsystem by 189 kg (416 lb) [2]. System closure can be dramatically increased by recycling waste streams or recovering useful substances from them. In a fully closed system, the air revitalization system (ARS) retains the CO2 removed from the habitat atmosphere instead of discarding it and ultimately recovers 02 from the CO2. A closed water recovery system (WRS) recycles wastewater streams including humidity condensate, spent hygiene water, and urine to produce water suitable for ingestion and hygiene use. The added mass of equipment required to recover 02 and to purify wastewater is offset by the reduction of mass needed for storage. The point at which an openloop system's mass and volume become equal to a regenerative system's mass and volume is determined by mission p a r a m e t e r s including crew size and mission duration. The source of spacecraft power is another important parameter. The U.S. space shuttle fleet relies on fuel cells to generate electricity. Since clean water is generated as a byproduct, there is no potential benefit in adding a WRS to the space shuttle ECLSS. Because ISS will use solar panels instead of fuel cells, it will not

458 have a built-in water supply and a WRS with a high degree of closure will be necessary. 2.

SPACECRAFT APPLICATIONS

On earth or in space, adsorption processes can be useful in both air and water purification. Three spacecraft life support operations that can employ adsorption are trace contaminant control, CO2 removal, and potable water recovery. The following sections provide historical perspectives for each application, together with descriptions of hardware to be used onboard ISS and h a r d w a r e currently in use onboard the Russian space station Mir. Possible future applications are also described. 2.1. T r a c e c o n t a m i n a n t c o n t r o l In the closed environment of a spacecraft or planetary base, trace quantities of potentially harmful substances can be more deleterious to h u m a n health t h a n similar levels would be in a less confined space. Crew exposure is constant r a t h e r t h a n intermittent and in the future will last for weeks or even months. Atmospheric trace contaminants may be organic or inorganic and may have biological or nonbiological origins. Some of the compounds found in space shuttle air samples [3], such as methane, can be identified as being products of crew metabolism. Others, such as dichloromethane, are likely produced by equipment off-gassing. Still others, such as 2-propanol, arise from evaporation of cleaning solvents. Adsorption has played a role in trace contaminant control throughout the U.S. h u m a n space flight program. Activated charcoal was used to remove contaminants onboard Mercury, Gemini, and Apollo spacecraft as well as on Skylab [4]. In these programs, the primary goal was to control odors rather t h a n to remove specific contaminants. The early Soviet space program also relied on activated charcoal for contaminant control in its Soyuz and Salyut spacecraft [5]. More recently, a list of spacecraft maximum allowable concentrations (SMACs) for some two hundred contaminants has been compiled [6]. The list includes a wide variety of aromatic and aliphatic hydrocarbons, halocarbons, and inorganics. The compilation of this list has driven a search for methods of removing or destroying particular compounds and classes of compounds. For example, the air quality requirements imposed on the space shuttle and Spacelab programs have resulted in the addition of an ambient temperature catalytic oxidizer intended primarily for destruction of CO [7]. Efforts to develop a trace contaminant control subsystem (TCCS) to remove specific compounds as well as odors expected on a space station have been underway since the early 1970s [8]. A contaminant load model was originally derived from

459 Apollo equipment off-gassing data and h u m a n metabolic studies. Based upon that model, a multi-step contaminant removal process was proposed in which an expendable activated charcoal bed would remove irreversibly adsorbed compounds, a regenerable activated charcoal bed would remove more weakly adsorbed compounds, and a catalytic oxidizer would convert reactive compounds to CO2, N2, and water vapor. The contaminant load model was used to optimize the relative sizes and flow rates for the charcoal beds, with the resulting design calling for a flow rate of 129 m3/hr (76 cfm) through the 14.5 kg (32 lb) expendable bed and a flow rate of 5.1 m3/hr (3 cfm) through the 2.2 kg (4.9 lb) regenerable bed. As more complete and accurate information became available, the original contaminant load model was refined and the TCCS design evolved [9-11]. The design ultimately selected for use on ISS is represented schematically in Figure 1 [12]. It includes an activated charcoal bed, a catalytic oxidizer, and a LiOH bed. The charcoal bed removes high molecular weight compounds that cannot be readily desorbed and is therefore an expendable item. It also removes ammonia, which is a potential catalyst poison. The bed contains 22.7 kg (50 lb) charcoal impregnated with 10% (weight) phosphoric acid to enhance its ammonia removal capability.

catalytic

,eater

oxidizer '~1 recuperative heatexchanger

~l activated - - - I ~ Q charcoal bed

from cabi

''

I ~1 LiOHbed I

blower ~

l

return

bypass Figure 1. ISS trace contaminant control subsystem schematic. The activated charcoal bed removes high molecular weight organics and potential catalyst poisons. The catalytic oxidizer destroys low molecular weight compounds that are not captured by the charcoal bed. The LiOH bed removes acid gases produced by oxidation of halocarbons.

460 Cabin air flows through the bed at 15.3 m3/hr (9 cfm). After leaving the bed, the process stream is split so that 4.6 m3/hr (2.7 cfm) is routed through a catalytic oxidizer where low molecular weight compounds are destroyed and then through a post-sorbent bed containing LiOH to remove acid gases produced in the oxidizer [7]. The remainder of the air bypasses the oxidizer and LiOH bed. The trace contaminant control assembly used onboard Mir relies on both regenerable and expendable sorbent beds upstream of a catalytic oxidizer. As is the case with the ISS TCCS, the expendable bed is sized to remove high molecular weight contaminants. The regenerable beds are used to remove more volatile (and therefore more readily desorbed) compounds. The expendable bed contains 1.3 kg (2.9 lb) activated charcoal and each of the regenerable beds contains approximately 7.4 kg (16.3 lb) activated charcoal. Air flow through the assembly is nominally 20 m3/hr (11.8 cfm). Each bed is regenerated after 20 days of operation by exposing it to space vacuum for 60 minutes and maintaining bed temperature between 170 C and 200 C (338 F and 392 F) for 90 minutes [13]. 2.2. CO2 r e m o v a l Regardless of whether it is open-loop or regenerative, an ECLSS must remove CO2 from the habitat atmosphere. NASA has specified that the nominal CO2 partial pressure shall be 5.0 mm Hg on the space shuttles [14] and that the maximum daily average CO2 partial pressure shall be 5.3 mm Hg on the ISS [15]. These values are roughly an order of magnitude higher than the typical atmospheric value of 0.23 mm Hg [16]. In order to meet these specifications, the ECLSS must remove CO2 generated by the crew at a nominal rate of 1.0 kg/person/day (2.2 lb/person/day). On Mercury, Gemini, and Apollo missions, CO2 was removed by LiOH. In these systems, CO2 was absorbed by LiOH and then reacted to form Li2CO3: CO2 + 2 LiOH

> Li2CO3 + H20

This reaction is essentially irreversible. As noted earlier, the mass of LiOH required increases linearly with mission length and crew size. While acceptable for relatively short missions, the mass penalty associated with expendable LiOH canisters was unacceptably large for the longer Skylab missions of the 1970s. For those missions, CO2 was reversibly adsorbed on zeolite material. The Skylab CO2 removal subsystem included two canisters, each containing molecular sieve 13X for dehumidification and 5A for C02 removal. Air entering the adsorbing canister contacted the 13X material before the 5A material. Adsorption of water lowered the dewpoint of the process air and reduced the competition between water and CO2 for adsorption sites on the 5A material. The subsystem operated cyclically so that as one bed adsorbed water vapor and CO2, the other was regenerated by venting to space vacuum [17].

461 The loss of water by the Skylab subsystem is clearly a disadvantage for systems driven toward loop closure. A design for a "water-save" process was proposed even before the Skylab missions [17]. In this configuration, the 13X material in the sorption beds would be replaced with separate desiccant beds containing silica gel. Space vacuum would no longer be used to recover water from the saturated desiccant material; instead, the warm, dry process air leaving the adsorbing 5A bed would be used to strip the water from the desiccant bed before being returned to the cabin. This concept ultimately evolved into the four-bed molecular sieve (4BMS) subsystem that will be used onboard ISS. A schematic representation of the 4BMS is shown in Figure 2. As was true on Skylab, the ISS subsystem operates continuously and works in a cyclic fashion so that as one sorbent bed adsorbs CO2, the other bed is regenerated. Similarly, as one desiccant bed removes water from the saturated air entering the subsystem, the other desiccant bed is regenerated by dry, CO2-depleted air and water vapor is returned to the cabin. The desiccant beds contain zeolite 13X and silica gel, while the CO2 sorbent beds contain zeolite 5A.

~ _ _ _ C02 sorbent i. . . . 4

desiccant 1

process air

II I I I

1

~)-D~~ precooler I blower

return to cabin

9

desiccant3 ]~

Q

()

Q-~ to vacuum

C02 sorbent2

Figure 2. ISS Four-bed molecular sieve subsystem schematic. Darker lines indicate process air flow during half-cycle described in text: air enters desiccant bed 1, continues through CO2 sorbent bed 2, and exits through desiccant bed 3. CO2 sorbent bed 4 is desorbed to vacuum.

As shown in Figure 2, process air laden with water vapor and CO2 enters the adsorbing desiccant bed (bed 1), where the water vapor is removed to protect the downstream CO2 removal bed. Although CO2 is also adsorbed by zeolite 13X, it is displaced by the advancing water front as the desiccant bed approaches saturation. The dry air exiting the bed is drawn through the blower and into the precooler which

462 removes the heat introduced by the blower as well as the heat of adsorption generated in the desiccant bed. The cool, CO2-1aden air passes into the adsorbing CO2 sorbent bed (bed 2), which selectively removes the C02. The air is then directed into the desorbing desiccant bed (bed 3), where the CO2-1ean air stream is rehumidified before being returned to the cabin. During this half-cycle, the second CO2 removal bed (bed 4) is regenerated by a combination pressure/thermal swing method. The bed is pumped down at the beginning of the half-cycle and the residual air is returned to the cabin. The pump is then turned off and embedded electrical heaters are used to heat the bed. As temperature increases, the CO2 is desorbed from the zeolite surface and returns to the gas phase, causing the pressure in the bed to rise. The gas is removed from the desorbing bed by means of a vacuum pump or by exposure to space vacuum. At the end of the half-cycle, the selector valves change position, allowing the newly regenerated beds to become the adsorbing beds and the saturated beds to become the desorbing beds so that the next half-cycle can begin [18]. NASA has gained understanding of and experience with 4BMS operation through extensive testing at Marshall Space Flight Center (MSFC). A combination of subsystem, integrated system, and life testing conducted at MSFC has provided basic understanding of the subsystem and how its performance is affected by various cabin and interface conditions [19-22]. During one series of tests, the 4BMS was challenged by off-nominal flow rates, inlet C02 concentrations, and temperatures [21]. The MSFC program also included a test of the ISS baseline ARS configuration in 1996 [22]. Testing of the 4BMS has also been performed at NASA's Johnson Space Center. In 1992 and 1994, subsystem tests were performed in which inlet CO2 concentration, flow rate, and half-cycle duration were varied [18]. In addition, a 4BMS unit has been used extensively in chamber tests at JSC. In 1996, the subsystem was part of an integrated ARS that provided a breathable atmosphere for a four-person crew living in a 250 m 3 (8824 ft 3) chamber for 30 days. In a second test in 1997, the unit was part of a different integrated ARS that supported a four-person crew living in the same chamber for 60 days. During both tests, the crew participated in normal activities including exercise. The unit's CO2 removal rate varied with the chamber's atmospheric CO2 concentration, which in turn varied with the level of crew activity. The subsystem was able to maintain acceptable C02 levels in the chamber [23,24]. As noted above, the desiccant prevents competition between water and CO2 for the 5A adsorption sites. In order to reduce hardware mass and process complexity, several studies have explored the possibility of using hydrophobic sorbents for CO2 removal. A method of producing amine-functionalized carbon molecular sieves (CMS) was developed in the late 1980s [25]. The untreated CMS material exhibited a high degree of hydrophobicity and the addition of the amine groups increased the material's C02 capacity. The functionalized CMS material has since demonstrated

463 its ability to remove 80% of the CO2 present in a simulated spacesuit air stream without initial dehumidification [26]. Competition between water and COe can also be avoided through use of a material that allows (or even requires) the two compounds to sorb cooperatively. The solid amine used in the extended duration orbiter and described earlier is an example of such a material [2]. The Mir CO2 removal system also relies on a combination of desiccant beds and CO2 sorbent beds. Detailed information on the materials used is not readily available. 2.3. W a t e r r e c o v e r y

As h u m a n space missions have become longer and more complex, the demand for water has increased. During the Mercury program in the early 1960s, the crew for each mission consisted of a single astronaut and missions lasted from 15 minutes to 1.4 days. As NASA progressed through the Gemini and Apollo programs to Skylab and the space shuttle, missions grew longer and included more astronauts. Water supply requirements increased accordingly. On Mercury missions, approximately 2.7 kg (6 lb) local municipal water was supplied for the astronaut's ingestion on each mission [27]. Over the last thirty years, NASA has developed increasingly stringent water quality standards [15]. Allowable concentrations are now specified for ionic species and for classes of organic compounds. Comparison of NASA's standards with the Environmental Protection Agency's (EPA) National Primary and Secondary Drinking Water Regulations [28] shows that for species and compounds appearing in both documents NASA's standards are equal to or stricter than the EPA's. Although wastewater has not yet been recycled for use by crewmembers onboard a U.S. spacecraft, this will occur on ISS. Wastewater streams will include urine, spent hygiene water, and humidity condensate. These streams have distinctly different compositions; urine typically contains high concentrations of solids and ionic species, while spent hygiene water contains large organic molecules contributed by soap and skin oils. Humidity condensate is a relatively clean stream, but it contains volatile, water-soluble organics that can be difficult to remove. The total organic carbon (TOC) concentration of the combined wastewater is expected to be less t h a n 500 ppm. Dozens of species will contribute to the TOC, with concentrations ranging from less than 10 ppb to greater t h a n 10 ppm [29]. In addition to neutral and charged organic species, a number of inorganic ions will be present. In spite of the power penalty associated with phase change operations, they are attractive options for processing waste streams with high solids content. A vapor compression distillation (VCD) unit has been chosen to process urine and flush water onboard ISS. The urine distillate produced by the VCD will be mixed with the hygiene waste and humidity condensate streams before being processed by a

464 combination of technologies, p r i m a r i l y ion exchange, adsorption, a n d catalytic oxidation. These technologies will achieve complete recovery of the w a t e r in the combined w a s t e s t r e a m . A series of ion exchange resins a n d a d s o r b e n t s p a c k a g e d in Unibeds | will remove a p p r o x i m a t e l y 99% of ionic species a n d 95% of organic species [30] p r e s e n t in the w a s t e w a t e r . Each of the Unibed | media h a s been chosen because of its affinity for a specific group of c o n t a m i n a n t s . For example, a w e a k base anion exchange resin is used to remove organic acids. The media c u r r e n t l y baselined for use onboard ISS are shown in Figure 3 and described in Table 1.

influenvt~

200 cm3 MCV-RT

iiiiiii !i !ii,!il i iiii! !i!,o

),;/.. .~i,:.i

:::'ii:;,i: !:",~ii:.....i 695 cma INN-77 275 cma IRN-68

630 cm3 580-26

1325 cm3 APA *****+4 1325 cm3 XAD-4 ~ ~ 4 ............ tl ...... 200 cm3 IRN-150 effiuen

Figure 3. ISS development unit Unibed | media

465 Table 1 ISS Development Unit Unibed | Media Sorbent

Description

Function

MCV-RT (Umpqua Research)

iodinated anion exchange resin

introduces bactericidal iodine into water

IRN-150 (Rohm and Haas)

equal exchange capacity mix of strongly basic anion exchange resin (IRN-77) and strongly acidic cation exchange resin (IRN-78)

removes cations, anions, and some organics

IRN-77 (Rohm and Haas)

strongly acidic cation exchange resin

removes ammonium and other cations

IRA-68 (Rohm and Haas)

weakly basic anion exchange resin

removes organic acids

580-26 (Barneby Cheney)

GAC manufactured from coconut shell

removes organics

APA (Calgon)

GAC manufactured from bituminous coal

removes organics

XAD-4 (Rohm and Haas)

polymeric adsorbent

removes low MW solutes

IRN-150 (Rohm and Haas)

equal exchange capacity mix of strongly basic anion exchange resin (IRN-77) and strongly acidic cation exchange resin (IRN-78)

removes cations, anions, and some organics

note: sorbents are listed in direction of flow Reference: [31]

466 The p r i m a r y a d v a n t a g e of using adsorption for this application is the fact t h a t complete w a t e r recovery can be achieved. Simplicity of operation is an additional benefit. Due to the complexity associated with regenerating the various media in a closed w a t e r recovery loop with limited mass and volume allocations, each Unibed | m u s t be replaced when expended. In this case, the benefits associated with using a variety of media outweigh the mass penalty associated with replacing the beds. Since the selected media have different physical and chemical characteristics, the beds can remove a wide range of contaminants; this gives the Unibeds | the ability to handle variations in the feed s t r e a m with a high degree of reliability and robustness. Since each Unibed | will contain identical media, commonality provides an additional advantage. The Mir hygiene w a t e r processor relies on a system similar to t h a t described above for ISS. The specific n a t u r e and sequencing of the sorbents is proprietary [32].

3.

FUTURE DIRECTIONS

NASA is currently studying the possibility of constructing remote bases on the l u n a r and M a r t i a n surfaces. When the destination of a space mission is a p l a n e t a r y body, it m a y be more economical to process locally available resources into usable products t h a n to carry all consumables from E a r t h or to d e m a n d complete closure of the life support system. This aspect of mission design, called in situ resource utilization (ISRU), could involve extensive use of adsorption processes to achieve separations. The rock and soil on both the moon and Mars are potential sources of oxygen, minerals, and bulk material, but Mars also has a substantial atmosphere. As shown in Table 2, the Mars atmosphere is made up primarily of carbon dioxide and contains significant a m o u n t s of nitrogen and argon. When s e p a r a t e d and compressed, the atmospheric constituents can be used to supply m a n y essential ingredients needed to support life. In addition to being essential for plants, carbon dioxide can also be an i m p o r t a n t source of oxygen. While oxygen's use in life support is obvious, its p r i m a r y use on a M a r t i a n base would be as an oxidant for rockets launched from the planet's surface [34,35]. Nitrogen is needed to make up for leakage of buffer gas from spacecraft and surface structures; argon m a y also be used as a component of buffer gas and in electrical propulsion systems [36]. The atmosphere is presently the only known source of water outside the polar areas, and although it is quite dilute, it m a y have tremendous value both for life support and as a source of hydrogen.

467 Table 2 Model Mars Atmospheric Composition and Conditions Material CO2 N2 Ar 02 CO H20

Composition (vol. %) 95.3 2.7 1.6 0.13 0.07 0.03

pressure: 6 torr (range, 4.5 - 11.5 torr) temperature: 200 - 270 K (range, 130 - 300 K) Conditions and humidity are highly dependent on latitude and season. Reference: [33]

The high costs of electrical power and launched mass give sharp focus to the search for technologies that will perform the separations and compressions needed to utilize the Martian atmosphere. Adsorption technologies are candidates for the same reasons they have been chosen for space missions to date. In addition, it appears possible to take advantage of the large Martian diurnal temperature cycle (70 K on average over much of the year at low latitudes) to effect adsorption separations and compressions that use very little electrical power. Instead, the processes operate as heat engines that take energy from the environment during the relatively warm M a r t i a n day and dump it to the environment during the cold night. Pressure-swing adsorption separations and compressions using the Mars diurnal temperature cycle have been suggested for a variety of potentially important ISRU applications and have been demonstrated in the laboratory [37]. F u r t h e r work in this area could lead to a re-examination of the optimum degree of closure for a surface ECLSS. 4.

SUMMARY

Throughout the history of the U.S. and Soviet (and now Russian) h u m a n space flight programs, adsorption processes have played an important role in controlling atmospheric contaminant concentrations. These processes are currently used in Mir's water recovery, CO2 removal, and trace contaminant control operations and are expected to play similar roles on ISS. Adsorption can also be

468 expected to play a significant role in ECLS systems on planetary bases, where sorbents may be used to process habitat air or to recover useful substances from the local environment. 5.

ACRONYMS

4BMS ARS CMS ECLSS EPA ISRU ISS JSC MSFC NASA SMAC TCCS TOC VCD WRS

four-bed molecular sieve air revitalization system carbon molecular sieve environmental control and life support system Environmental Protection Agency in situ resource utilization International Space Station Johnson Space Center Marshall Space Flight Center National Aeronautics and Space Administration spacecraft maximum allowable concentration trace contaminant control subsystem total organic carbon vapor compression distillation water recovery system

A c k n o w l e d g e m e n t s : The authors thank D. L. Carter, C. Finn, M. Kliss, J. Knox, and J. Perry for their helpful comments on style and content.

REFERENCES

1. Metcalf & Eddy, Inc., Wastewater Engineering: Treatment, Disposal, and Reuse, 3rd ed., McGraw-Hill Inc., New York (1991). 2. F.A. Ouellette, H. E. Winkler and G. S. Smith, The Extended Duration Orbiter Regenerable CO2 Removal System, SAE paper 901292, Society of Automotive Engineers, Warrendale PA, July 1990. 3. M.R. Schwartz and S. I. Oldmark, Analysis and Composition of a Model Trace Gaseous Mixture for a Spacecraft, SAE paper 860917, Society of Automotive Engineers, Warrendale PA, July 1986. 4. B.L. Diamant and W. R. Humphries, Past and Present Environmental Control and Life Support Systems on Manned Spacecraft, SAE paper 901210, Society of Automotive Engineers, Warrendale PA, July 1990.

469 5. P. O. Wieland, Designing for Human Presence in Space: An Introduction to Environmental Control and Life Support Systems, NASA Reference Publication 1324, 1994. 6. J.T. James, Spacecraft Maximum Allowable Concentrations for Airborne Contaminants, JSC 20584, 1995. 7. J. L. Perry, Elements of Spacecraft Cabin Air Quality Control, NASA Research Publication (in press). 8. T.M. Olcott, Development of a Sorber Trace Contaminant Control System including Pre- and Post-Sorbers for a Catalytic Oxidizer, NASA CR-2027, NASA contract NAS-1-9242, Lockheed Missiles and Space Company, May 1972. 9. M. I. Leban and P. A. Wagner, Space Station Freedom Gaseous Trace Contaminant Load Model Development, SAE paper 891513, Society of Automotive Engineers, Warrendale PA, July 1989. 10.J.L. Perry, Trace Chemical Contaminant Generation Rates for Spacecraft Contamination Control System Design, NASA TM-108497, George C. Marshall Space Flight Center, August 1995. 11. T. Olcott, R. Lamparter, B. Maine, A. Weitzmann, R. Luce, G. Olivier, E. Kawasaki, O. Masi and C. Richardi, Design, Fabrication, and Test of a Trace Contaminant Control System, LMSC-D4622467, NASA contract NAS-1-11526, Lockheed Missiles and Space Company, November 1975. 12.D.E. Link and J. W. Angeli, A Gaseous Trace Contaminant Control System for the Space Station Freedom Environmental Control and Life Support System, SAE paper 911452, Society of Automotive Engineers, Warrendale PA, July 1991. 13.R.E. Curtis, J. L. Perry and L. H. Abramov, Performance Testing of a Russian Mir Space Station Trace Contaminant Control Assembly, SAE paper 972267, Society of Automotive Engineers, Warrendale PA, July 1997. 14. Orbiter Vehicle End Item Specification for the Space Shuttle System. Part 1: Performance and Design Requirements, NASA document MJ070-0001, October 1988. 15. System Specification for the International Space Station, NASA document SSP 41000E, prepared by Boeing Defense and Space Group, Missiles and Space Division, Houston TX, July 1996. 16.R.C. Weast (ed.), CRC Handbook of Chemistry and Physics, 58th ed., CRC Press Inc., West Palm Beach FL (1977). 17.T. Coull, Skylab Regenerable Carbon Dioxide Removal System, paper 72-Av-4, American Society of Mechanical Engineers, New York NY, August 1972. 18.M.C. Kimble, M. S. Nacheff-Benedict, L. A. Dall-Bauman and M. R. Kallberg, Molecular Sieve CO2 Removal Systems for Future Missions: Test Results and Alternative Designs, SAE paper 941396, Society of Automotive Engineers, Warrendale PA, June 1994.

470 19.R.G. Schunk, R. M. Bagdigian, R. L. Carrasquillo, K. Y. Ogle and P. O. Wieland, Space Station CMIF Extended Duration Metabolic Control Test Final Report, NASA TM-100362, George C. Marshall Space Flight Center, March 1989. 20.J.L. Perry, R. L. Carrasquillo, G. D. Franks, K. R. Frederick, J. C. Knox, D. A. Long, K. Y. Ogle and K. J. Parrish, International Space Station Integrated Atmosphere Revitalization Subsystem Testing, SAE paper 961519, Society of Automotive Engineers, Warrendale PA, July 1996. 21.J.C. Knox, Performance Enhancement Test Preliminary Report, George C. Marshall Space Flight Center memo ED62(141-96), October 1996. 22. J. L. Perry, G. D. Franks and J. C. Knox, International Space Station Program Phase III Integrated Atmosphere Revitalization Subsystem Test Final Report, NASA TM- 108541, George C. Marshall Space Flight Center, August 1997. 23. S. Brasseaux, M. Rosenbaum, L. Supra and D. E1 Sherif, Performance of the Atmosphere Revitalization System During Phase II of the Lunar-Mars Life Support Test Project, SAE paper 972418, Society of Automotive Engineers, Warrendale PA, July 1997. 24.L.N. Supra and S. F. Brasseaux, Molecular Sieve CO2 Removal Systems: International Space Station and Lunar-Mars Life Support Test Project, SAE paper 972419, Society of Automotive Engineers, Warrendale PA, July 1997. 25. H. A. Zinnen, A. R. Oroskar and C.-H. Chang, Carbon Dioxide Removal Using Aminated Carbon Molecular Sieves, US Patent No. 4 810 266, (1989). 26. S. K. Rose, A. K. MacKnight and D. E1 Sherif, CO2 Removal with Enhanced Molecular Sieves, SAE paper 972431, Society of Automotive Engineers, Warrendale PA, July 1997. 27. R. L. Sauer and J. E. Straub II, Potable Water Supply in US Manned Space Missions, presented at the 43rd Congress of the International Astronautical Federation, Washington D. C., 1992. 28. Code of Federal Regulations, Title 40: Environmental Protection CFR Pilot, Chapter 1: Environmental Protection Agency, Subchapter D: Water Programs, Part 141: National Primary Drinking Water Regulations and Part 143: National Secondary Drinking Water Regulations. 29. H. Cole, M. Habercom, M. Crenshaw, S. Johnson, S. Manuel, W. Martindale, G. Whitman and M. Traweek, The Characterization of Organic Contaminants During the Development of the Space Station Water Reclamation and Management System, SAE paper 911376, Society of Automotive Engineers, Warrendale PA, July 1991. 30. D. L. Carter, D. W. Holder and C. F. Hutchens, International Space Station Environmental Control and Life Support System phase III Water Recovery Test Stage 9 Final Report, NASA TM-108498, George C. Marshall Space Flight Center, September 1995.

471 31.D.L. Carter, Phase III Integrated Water Recovery Testing at MSFC: International Space Station Recipient Mode Test Results and Lessons Learned, SAE paper 972375, Society of Automotive Engineers, Warrendale PA, July 1997. 32.K.L. Mitchell, R. M. Bagdigian, R. L. Carrasquillo, D. L. Carter, G. D. Franks, D. W. Holder Jr., C. F. Hutchens, K. Y. Ogle, J. L. Perry and C. D. Ray, Technical Assessment of Mir-1 Life Support Hardware for the International Space Station, NASA TM- 108441, George C. Marshall Space Flight Center, March 1994. 33. T. R. Meyer and C. P. McKay, The Resources of Mars for Human Settlement, J. Brit. Interplanetary Soc., 42 (1989) 147. 34.R. Zubrin, S. Price, L. Mason, L. Clark, B. Clark and B. O'Handley, An End-toEnd Demonstration of a Full-Scale Mars in-situ Propellant Production Unit, paper IAA-95-IAA-1.3.05, presented at the 46th International Astronautical Congress, Oslo, Norway, 1995. 35. K. R. Sridhar, Mars Sample Return Mission with ISPP, J. Brit. Interplanetary Soc., 49 (1996) 435. 36.K.H. Groh, O. Blum, H. Rado and H. W. Loeb, Inert Gas Radio-Frequency Thruster, paper RIT-10, AIAA/DGLR International Electric Propulsion Conference, Progress in Astronautics and Aeronautics, 79 (1979). 37.J.E. Finn, K. R. Sridhar and C. P. McKay, Utilisation of Martian Atmosphere Constituents by Temperature-Swing Adsorption, J. Brit. Interplanetary Soc., 49 (1996) 423.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

473

P o l y m e r - s u p p o r t e d p h o s p h o r u s - c o n t a i n i n g l i g a n d s for s e l e c t i v e m e t a l ion complexation K. P. Ripperger and S. D. Alexandratos Department of Chemistry, University of Tennessee, Knoxville, Tennessee 37996 USA

Groundwater contamination in the United States is a serious problem with an estimated three million sites requiring remediation [1]. Heightened environmental awareness, depletion of high grade ore reserves, and stricter legislation has generated a strong interest in efficient hydrometallurgical processes [2]. Two methods for the separation and recovery of metal ions from aqueous solutions are solvent extraction with soluble complexants and ion exchange with crosslinked polymer beads. While a wide variety of ligands has been developed for both techniques, this review will focus on phosphorus-based ion exchange resins due to their versatility, wide range of applicability, and environmental compatibility. 1.

P H O S P H O R U S - B A S E D COMPLEXANTS IN S O L V E N T EXTRACTION

In solvent extraction one or more substrates (either ionic or molecular) is transferred from an aqueous phase to a liquid organic phase. The extractant is the liquid organic phase into which the targeted substrate is separated; it is composed of the complexant or extracting agent (defined as the compound responsible for complexation of the substrate) either alone or with a solvent (also known as the diluent). Several terms are used to describe the effectiveness of the extraction system. The distribution coefficient, D, is the ratio of the concentration of the targeted substrate in the organic phase to the concentration in the aqueous phase. The ratio of distribution coefficients or equilibrium constants of two substrates under identical reaction conditions is referred to as the separation factor. The stability constant quantifies the strength of the ligand-substrate complex [3]. Because complexation occurs at the aqueous/organic interface, vigorous mixing of the two layers is required to provide adequate contact of the two phases [3]. The advantages to solvent extraction include high throughput and the numerous complexants that have been developed for many different separations [1]. Disadvantages include complexant loss through dissolution and entrainment by third layer formation resulting from incomplete separation of the phases. This additional phase is due to precipitates, also referred to as cruds, or the formation of

474 inseparable microdroplets during the contact period. These emulsions can be very tenacious and will not separate after hours of centrifugation [2,4]. The choice of ligand is crucial to the success of metal ion separations. Phosphorus ligands demonstrate a high level of selectivity and are widely used commercially for extraction of transition metals, lanthanides and actinides [5,6]. These complexants have a high metal ion loading capacity, are (for the most part) relatively inexpensive, allow for stripping of the metal ion after separation, have suitable rates for extraction and stripping, and are stable in most reaction conditions [2]. The organophosphorus acids, esters, and oxides are thus a popular group of compounds for solvent extraction applications. The phosphoryl moiety may be combined with other functional groups, including carbonyls and amines, to yield bifunctional extractants [6-11]. Complexation with these compounds may or may not involve chelation [9,10]. Phosphorus ligands can complex metal ions through coordination or ion exchange [5]. Choice of an appropriate organic solvent is important because it influences how much of the metal is complexed into the organic phase and the number of ligands in the ligand/metal complex [8,12-14]. Complexants should have low solubility in the aqueous phase, be chemically stable and not form complexes with other components of the aqueous phase which are more stable than the complexes formed with the targeted metal ion [2]. If the extractant is too soluble in the aqueous phase, poor separation and recovery can result [15]. 1.1. M i x e d e x t r a c t a n t s y s t e m s

Combinations of acidic and coordinating extracting agents are used to increase metal ion uptake. Synergism is observed if an increased extracting power results that is greater than the sum of the extracting abilities of the individual reagents [5]. Synergic systems containing phosphoric acids, phosphonates, phosphine oxides, hydroxyoximes, 8-hydroxyquinoline, and diamines have been prepared [5,16-18]. The di(ethylhexyl)phosphoric acid (DEHPA)/8-hydroxy- quinoline system allows for the extraction of lanthanides from phosphoric acid [16]. A combination of DEHPA and 5,8-diethyl-7-hydroxy-6-dodecane oxime (LIX63) displays synergism in the extraction of Cu(II) [17]. The D EHPA/tributyl phosphate (TBP) system is useful for extraction of lanthanides and actinides. The behavior of this system is dependent on the organic solvent, temperature and ratio of the reagents [18]. If the ratio of phosphate to acid is too high, hydrogen bonding between the two compounds decreases the concentration of the free ligand and a decrease in the distribution coefficient is observed [5]. Third phase formation is possible with phosphine oxide/phosphoric acid combinations [18]. Not all extracting agent combinations display synergism. When di-(2,4,4-trimethypentyl)phosphinic acid (DTMPPA) is combined with carbamoylmethylphosphine oxide or a diphosphine dioxide, lanthanide uptake is less than that observed for either of the individual compounds. When steric crowding is decreased using tri-n-octyl phosphine oxide, the distribution coefficient increases but with lowered selectivity [19]. Various water soluble diamines have been used to increase the extraction of U(VI) by D EHPA. The increased hydrophobicity of the extracted complex increases the distribution coefficient of the metal ion [20].

475 2.

ORGANIC P O L Y M E R - S U P P O R T E D R E A G E N T S

Soluble phosphorus-based compounds are found to be versatile metal ion complexants with a wide-ranging applicability. It may thus be expected that combining these important ligands with the advantages inherent to using crosslinked polymers for ionic separations can result in the development of new ionselective polymer-supported reagents. In applying polymer-supported reagents to metal ion separation processes, it is obvious that extractant loss is eliminated because the ligand is covalently bound to the polymer. Both batch and continuous processes are possible. After the polymer is loaded with the targeted metal ion, the metal is recovered by contacting the resin with a stripping agent. Because the volume of the stripping agent is much less than the volume of the original aqueous solution, polymeric reagents are ideal for preconcentration applications [21]. An understanding of polymeric ligand-metal ion chemistry, the complexation kinetics, and the physical properties of the polymer support is important to the design of a successful process [21]. Cross-linked copolymers are preferred because they are insoluble and easily separated from the aqueous phase [21]. Divinylbenzene (DVB) is a common crosslinking agent. Linear polymers, which may be soluble in the metal ion solution, are used less frequently due to the complexity of polymer recovery. The polymer support must allow functionalization with the chosen ligand or else be prepared from functionalized monomers. It should also be physically stable and chemically inert in the aqueous solution. Because of their physical and chemical stability, polystyrene and poly(vinylbenzyl chloride) supports are frequently used. Other supports include acrylic polymers and poly(vinyl chloride) [21,22]. The behavior of the polymer-supported reagent is dependent on several parameters including matrix rigidity, particle size and resin morphology [23]. Macroporous (or macroreticular, MR) polymers allow greater accessibility into the matrix although the volume capacity of the ligand is diminished [23]. Microporous (or gel) polymers must swell to maximize accessibility. In columns, however, swelling of the polymer can disrupt the uniformity of the packing unless precautions are taken [24]. The synthesis and complexation behavior of various organophosphorus polymersupported reagents is described below. They are separated into classes based on the type of ligand. 2.1. P h o s p h o r i c a c i d r e s i n s

Polymer supported phosphoric acids are prepared by the phosphorylation of poly(vinylbenzyl chloride), poly(glycidyl methacrylate), poly(vinyl chloride) or poly(vinyl alcohol) with either phosphoric acid or phosphorus oxychloride (Figure 1) [22,25-27]. The polymers can also be synthesized via polymerization of the vinyl phosphoric acid monomer (Figure 2) [26]. The polymer-supported reagents complex a variety of metal ions including Cu(II), Zn(II), Ni(II), Ag(I), Cd(II), Fe(III), U(VI) and Al(III) [22,25-27].

476

OH

O

O

I

I

CI--P-C1 II 0

HO-P-OH II 0

Figure 1. Phosphorylation of poly(vinyl alcohol).

0

0

O-P-OEt

,

I

O-P-OEt I

CI

O 0 -P-OEt I

9

CI

OH

Figure 2. Synthesis of a phosphoric acid resin via a functionalized monomer.

2.2.

Phosphonic

acid/aminophosphonic

acid/phosphonate

resins

P o l y s t y r e n e a n d p o l y ( v i n y l b e n z y l chloride) r e s i n s are f u n c t i o n a l i z e d via t h e F r i e d e l - C r a f t s reaction w i t h PC1JA1C13 to yield immobilized p h o s p h o n i c acid ligands a f t e r h y d r o l y s i s a n d nitric acid o x i d a t i o n ( F i g u r e 3) [28,29]. P o l y ( v i n y l b e n z y l

--(- C H-~CH --)-1

-(-- C Hz,CH -),-

--(CHIC H ---)-

/P\

HO-P =O I OH

CI --(--C H;CH --)--

CI

~ - CH:,C H -)'

--(- C H:,CH -)-"

CI:P

HO - P

CHiCI

CH21 CI--P-C1 C1

CI

I

OH

CH2

HO-P =O OH

Figure 3. Synthesis of phosphonic acid resins via the Friedel-Crafts reaction.

477 chloride) functionalization leads to both phosphonic and methylenephosphonic ligands [30]. Methylenephosphonic acid ligands are produced exclusively via the Arbuzov reaction on poly(vinylbenzyl chloride) (Figure 4) [31].

O 9

CHuP-OEt I OEt

O -~

H2

-OH OH

Figure 4. Synthesis of phosphonic acid resin via the Arbuzov reaction.

Phosphonic acid resins have demonstrated high metal ion uptake and selectivity toward Yb(III), Fe(III), Hg(II), U(VI), Ag(I), and Mn(II) ions [29,32]. Polymer supported phosphonic acids have been used in the separation of Mo(VI) from U(VI) after both metals are loaded on the resin: elution with sodium acetate removes the Mo(VI) and leaves the U(VI) on the polymer [33]. Alkylphosphonate resins have been prepared from poly(vinylbenzyl chloride) via the Friedel-Crafts reaction with PC1JA1C13 followed by reaction with ethanol in a toluene solution (Figure 5) [34]. The resins coordinate metal ions and can have greater metal uptake t h a n the acid resin given that the phosphoryl moiety of the diester is more basic due to the electron donating ability of the alkyl groups. The diester resin is a better complexing agent for Au(I) from alkaline cyanide solution [34]. The metal is extracted as the Au(CN)2- complex through coordination rather than ion-exchange [35]. Distribution coefficients can be lower, however, with phosphonate diester resins relative to the monester analogues (Table 1) [34]. The monoester ligand may therefore be more useful because it allows for ionic coordination with greater accessibility.

c1 i

o OR

R = H, O E t

Figure 5. Synthesis of the phosphonate resin.

478 Table 1 Distribution coefficients for Ag(I) and Hg(II) from 1N HNOd3N NaNOa by sulfomc, phosphonic, phosphonate monoester and phosphonate diester resins M n+

SO3H

PO(OH)2

PO(OH)(OEt)

Ag(I) Hg(II)

6.03 6.76

34.7 22.9

1230 144

PO(OEt)2 9.77 123

Poly(vinylbenzyl chloride) resins containing dialkyl-N,N-diethylcarbamoylmethyl phosphonates have been prepared with the carbamoylmethyl phosphonate (CMP) moiety attached to the polymer through the carbon alpha to the phosphoryl and carbonyl groups (Figure 6). The resins, like the small molecule analogues, complex lanthanide ions; however, whereas the soluble CMPs show an increase in the distribution coefficients as the aqueous phase acidity increase, the polymersupported reagents show the opposite trend [36].

NEt2

O ~C NEt2

O.. C, CH2-CI

+

"C\

//

,

C H 2__C

/P

H

OEt

O /

"OEt

H

OEt

OEt

Figure 6. Immobilization of CMP on poly(vinylbenzyl chloride).

~

CH2-NH 2 +

R 1=

R,C(O)R 2

~

H, aryl, alkyl

R 2 = H, aryl, alkyl

e l

C HsN =C \ R2

1 O R 2 OH

Figure 7. Synthesis of the aminophosphonic acid resin.

Aminophosphonic acid resins have been synthesized by functionalization of vinylbenzylamine with a ketone or aldehyde and phosphorus acid (Figure 7).

479 An a-amino-benzylphosphonic acid resin has been prepared (Figure 8). Resins crosslinked with 2% DVB have been used for the complexation of calcium; the acid resin performs somewhat better than the analogous ester resin (Table 2). Similar polymers have been prepared with a pyridyl rather than a phenyl ring (Figure 9). They also complex calcium ions, although the difference in uptake between the acid and ester resins is less significant. The pyridyl resins also demonstrate an affinity for Co(II) (Table 2). The metal ions are stripped with 6M HC1 with 85% metal recovery [37]. Poly(aminophosphonic acid) has been used to selectively complex various metal ions including Pb(II), Mn(II), Mg(II), Zn(II), Ni(II), Fe(III), Cu(II), Co(II) and Ca(II) from solutions containing alkali metal salts (Figure 10). Regeneration of the resin with 6 N HC1 allows twenty load/strip cycles without a decrease in metal uptake or selectivity [24].

o

o

CH2-NH C H - P - O H

CH2-NH CH-P-OH

OH

OH

Figure 8. a-aminobenzylphosphonic acid resin.

Figure 9. a-aminopyridylphosphonic acid resin.

Table 2 Metal uptake of a-aminobenzylphosphonic acid resins Resin

Ca 2+ (mmol/g)

Co2+(mmol/g)

0.67 0.33 0.65 0.80

0.00 0.00 0.59 0.26

a-aminobenzylphosphonic acid a-aminobenzylphosphonate a-aminopyridylphosphonic acid a-aminopyridylphosphonate

0 CH2-NH CH-2PI - O H OH

Figure 10. a-aminophosphonic acid resin.

480 2.3.

Phosphinic

acid resins

Resins with primary and secondary phosphinic acid ligands are synthesized via a Friedel-Crafts reaction on polystyrene using PC1JA1C13 followed by hydrolysis (Figure 11) [38]. The polymer shows excellent metal ion uptake and selectivity over sodium ions (Table 3). The metals evaluated include Zn(II), Mn(II), Fe(III), U(VI), Th(IV), Am(III), Pu(IV) and various lanthanides [30-32,39]. A plot of log D (distribution coefficient) versus pH has a slope equal to the charge of the exchanging ion if ion exchange is the only complexation mechanism [40]. Extraction of Fe(III), U(VI), and Th(IV) is only slightly dependent on pH [32,39,41]. This indicates a coordinative mechanism to the extractions, probably through an entropy-driven coordination [42]. Phosphinic acid polymers have better lanthanide uptake than phosphonic acid polymers. The percent Eu(III) and Yb(III) complexed by the phosphinic acid resin is 70% and 82.5%, respectively, from 0.5 N HC1 while the amounts complexed are 25% and 50%, respectively, for the phosphonic acid resin [30].

0 P -Cl I

9

-~

P -OH I

C1

H

Figure 11. Synthesis of phosphinic acid resin.

Table 3 Distribution coefficients at equilibrium from 1 N HNOJ

2.4.

3 N NAN03

solutions

Acid Resin

Fe(III)

Mn(II)

Zn(II)

Co(II)

Phosphinic Pho spho nic Sulfonic

14100. 4570. 5.75

4.67 0.98 4.36

1.58 0.58 3.02

I. 70 0.72 2.39

Phosphine

sulfide resins/quaternary

phosphonium

resins

A phosphine sulfide resin has been prepared for the separation of noble metals from aqueous media [43]. The polymer-supported reagent is prepared by reacting diethyl phosphonate with isobutylmagnesium chloride, followed by the addition of cross-linked poly(vinylbenzyl chloride). The polymer-supported phosphonate is converted to the phosphine sulfide with phosphorus pentasulfide (Figure 12). The resin complexes both Au(III) (present as

AuCl~) and Pd(II) (present as

481

H;--CI ~_ CiMgP(O)(iBu)2 ~

O H~P-iBu " "

S CHuP-iBu I

,,

iBu

iBu

Figure 12. Synthesis of phosphine sulfide resin.

PdCI(SCN)2-), although it has a greater capacity for the latter (0.59 vs. 0.70 mequiv/g, respectively). The resin is selective toward both Pd(II) and Au(III) when evaluated individually in the presence of Pb, Ni and Cu(II) [43]. Complexation of noble metals by phosphines in HC1 solutions occurs via ion-pair formation between the phosphonium ion and the negatively charged noble metal chloride complex [44]. A quaternary phosphonium chloride ion-exchange resin has been prepared by reacting cross-linked poly(vinylbenzyl chloride) with tris(2,6dimethoxyphenyl) phosphine (Figure 13). The resin, evaluated with Au(III), Cu(II), Fe(III) and Pt(IV) in HC1, has low uptake of Cu(II) and Fe(III) but loading capacities of 0.24 and 0.30 mequiv/g for Au(III) and Pt(IV), respectively. The selectivity of the resin allows complete separation of Au(III) and Cu(II). Elution with 0.1 N NH4C1 / 0.5 N NH3 yields 17% Au(III) recovery [44].

CH~O ~CH2--C!

+ P--(~

cl+CH~O _~ )3

" ~CHzP--(

CH3O

/

)3

CH30

Figure 13. Synthesis of phosphonium chloride resin.

2.5. D u a l m e c h a n i s m b i f u n c t i o n a l p o l y m e r s The synergistic extraction of metal ions observed in solvent extraction chemistry when complexants capable of ion exchange are coupled with those capable of coordination, led to the development of a new category of polymer-supported reagents for metal ion separations. Dual mechanism bifunctional polymers (DMBPs) are functionalized with two different types of ligands which operate via two different mechanisms: one ligand, operating via ion exchange, brings metal ions into the polymer matrix in a relatively non-selective manner while the second ligand reacts by different mechanisms with the targeted metal ion. It is this second

482 ligand that is responsible for the observed selectivity and defines each of the three classes of DMBPs [28]. Class I DMBP resins selectively sorb metal ions via a redox reaction. The resin functionalized with primary and secondary phosphinic acid groups is an example. It is prepared by reacting cross-linked polystyrene with PC13 in the presence of A1C13 followed by hydrolysis (Figure 14) [28]. Ion-exchange through the P-O-H ligand allows access of the metal ions into the matrix where they are reduced through the P-H bond of the primary phosphinic acid. Metals with standard reduction potentials greater than 0.30 volt, including Ag(I) and Hg(II), can be reduced. Evidence for the reduction of Hg(II) includes the appearance of liquid mercury in the vial. The resin extracts significantly more Hg(II) and Ag(I) than the phosphonic acid resin; the results are compared in Table 4 [28].

Table 4 Percent Hg(II) and Ag(I) sorbed from aqueous solutions by 2% DVB phosphorus acid resins at constant (4 N NO;) ionic strength Percent Hg(II) Sorbed

Initial ratio (mequiv M n§ / acid ligand)

Percent Ag(I) Sorbed

Phosphinic Acid

Phosphonic Acid

Phosphinic Acid

Phosphonic Acid

99.8 99.8 69.4 60.8

84.7 69.0 52.5 45.8

70.1 75.2 68.5 64.8

30.4 36.0 29.9 26.9

0.10 0.50 1.0 1.5

~

O Pl- C I

b

I,

//

P -C1

Figure 14. Synthesis of phosphinic acid class I DMBP.

P\ OH

483 Class II DMBP resins selectively complex metal ions through coordination. An example is a resin functionalized with phosphorus monoester and diester ligands. The resin is prepared by reacting cross-linked poly(vinylbenzyl chloride) with PC13 in the presence of A1C13 followed by reaction with an ethanol/water solution (Figure 15) [45]. The resin has been used to complex a variety of metal ions including, Hg(II), Ag(I) and Fe(III). The bifunctional resin has a greater Ag(I) uptake than diester and monoester resins (Table 5). The resin also has a greater Ag(I) uptake than a mixed bed containing both monoester and diester resins [45].

C! I

o

0 II

II

-OH CH~P H CI

H~CI

OH

H--CI

Figure 15. Synthesis of phosphonate acid / ester class II DMBP.

Table 5 Ag(I) distribution coefficients from nitric acid solutions Resin Class II DMBP Monoester Resin Diester Resin Mixed Bed (Mono+ Diester)

4 N HN03 2924 439 463 630

2 N HN03 2594 680 373 481

Class III DMBP resins selectively remove metal ions through a precipitation reaction. A representative resin contains phosphonic acid and quaternary ammonium chloride ligands. The resin is prepared by reacting cross-linked poly(vinylbenzyl chloride) with PC13 in the presence of A1C13 at a temperature that allows for partial functionalization. After hydrolysis, the resin is reacted with a trialkylamine (Figure 16). With chloride as the counterion for the trialkylammonium ligand, the DMBP can remove Ag(I) from solution by precipitation as silver chloride. The mechanism probably entails initial ion exchange by the phosphonic acid ligand followed by the silver cation / chloride anion reaction. The results in Table 6 illustrate the improved Ag(I) uptake of the tributyl ammonium salt/phosphonic acid

484 c1

o

I~-Ci Ci

CH-Ci

~

H~Pl

CH -CI

H 2-

"

"

~ "

"

CHuPIII- O E t OEt

P, -

CI

Cl \

\

OEt

Figure 16. Synthesis of the quaternary amine/phosphonic acid class III DMBP.

Table 6 Percent Ag(I) sorbed from solution with a class III DMBP at a 5 min contact time Resin

Percent Ag(I) Sorbed

Monofunctional Phosphonic acid Monofunctional Quat Amine (C1-) DMBP Quat Amine (C1-)/Phosph Acid

31.1

17.3 57.7

resin relative to the monofunctional phosphonic acid and tributyl ammonium salt resins at a 5 min contact time [46].

2.6.

Diphosphonic

acid

polymers

Crosslinked polymers with the diphosphonic acid ligand incorporated into the polymer backbone have been synthesized by polymerizing tetraethylvinylidene diphosphonate (Et4VDPA), styrene, acrylonitrile and divinylbenzene. Beads are prepared with 96.4% incorporation of the diphosphonate monomer. The resin is sulfonated to yield a polymer with sulfonic acid, carboxylic acid, and diphosphonic acid ligands (Figure 17) [47]. The sulfonic acid maximizes ionic accessibility to the selective diphosphonic acid ligand [47,48]. The diphosphonic acid resin, available as Diphonix | from EiChrom Industries, has been used to extract U(VI), Pu(IV), Np(IV), Th(IV), Am(III), Eu(III), and a variety of transition metals [47-49]. The distribution coefficients obtained for the extraction of U(VI), Pu(IV), and Am(III), as well as those of an analogous monophosphonic acid and the monofunctional sulfonic acid resins, are listed in Table 7 and illustrate the higher uptake possible with Diphonix [48]. The dual access/recognition mechanism allows faster uptake than the monofunctional diphosphonic acid resin. The distribution coefficients of the bifunctional and monofunctional resins for the extraction of Eu(III) at a 0.5 h contact time are

485 c o m p a r e d in T a b l e 8 a n d i n d i c a t e a far g r e a t e r r a t e of c o m p l e x a t i o n by Diphonix [47].

EtO OEt P=O /

--~ EtO

+

+ -=k

/Px-=-O OEt

CN EtO

~

HO

I HO

O

I OH OH

~ X"l

I EtO

I OEt OEt

fi ~'1

~-

~_.H

SO-2OH Figure 17. Synthesis of Diphonix |

Table 7 D i s t r i b u t i o n coefficients of Am(III), U(VI), a n d Pu(IV) by D i p h o n i x | acid, a n d sulfonic acid r e s i n s Resin Diphonix | Diphonix | Diphonix | P h o s p h o n i c Acid P h o s p h o n i c Acid P h o s p h o n i c Acid Sulfonic Acid Sulfonic Acid Sulfonic Acid

Metal

0.10 N HNO3

U Pu Am U Pu Am U Pu Am

6.5 x 3.0 x 2.0 x 1.6 x 1.5 x 2.7 x 9.3 x >104 3.8 x

0.10 N HNO3 / 4 N NaNO3

10 5 10 4 10 6 10 4 10 4 10 4 10 3

2.2 2.9 5.6 2.9 1.3 2.4 2.9 2.8 4.9

10 ~

Table 8 P e r c e n t Eu(III) c o m p l e x e d from 1 N HNO3 a t a 0.5 h c o n t a c t t i m e Resin Diphonix | M o n o f u n c t i o n a l D i p h o s p h o n i c Acid Sulfonic Acid

phosphonic

% Complexed 98.3 32.8 44.9

x x x x x x x x x

10 4 10 3 10 3 10 3 10 2 10 10 2 102 10

486 Additionally, Diphonix complexes metal ions from high acid concentrations because of its chelating ability through the diphosphoryl ligand (Table 9) [49]. Water-soluble diphosphonic acids, including l-hydroxy-ethane-l,l-diphosphonic acid (HEDPA), are effective stripping agents [48].

Table 9 Distribution coefficients of Fe(III) and Cr(III) from 0.10, 1.0, and 10 N HNOa Resin Diphonix | Phosphonic Sulfonic

0.10 N

Fe(III) 1.0 N

10 N

1.2 x 10 5 6.6 x 104 5.0 x 104

2.4 x 10 5 2.3 x 103 1.3 x 103

2.4 x 10 4 1.1 x 103 1.2 x 102

0.10 N

Cr(III) 1.0 N

10 N

2.8 x 10 3 2.5 x 103 2.3 x 103

8.2 x 101 1.6 x 101 1.9 x 102

3.6 x 10 3 1.1 x 101 6.4

Diphosphonic acid ligands can also be immobilized on a cross-linked poly(vinylbenzyl chloride) support. The resin is functionalized with the sodium salt of tetraisopropyl methylene diphosphonate (Figure 18) followed by sulfonation to yield a resin with diphosphonic and sulfonic acid ligands [50]. The percent diphosphonate functionalization depends on the crosslink level and the ratio of styrene (St) to vinylbenzyl chloride (VBC) because of the sensitivity of the reaction to steric effects. Maximum functionalization is obtained with a 2:1 St/VBC ratio and 2% DVB crosslinking. The bifunctional diphosphonic acid resin has faster complexation kinetics t h a n the analogous monofunctional resin (Table 10) [50]. The 2% DVB crosslinked resin with a St/VBC ratio of 2:1 has a Eu(III) uptake in 1 N HNO3 comparable to Diphonix | (Table 10) [47,50].

iPrO

\

/

OiPr

iPrO

P =0 + NaCH

' ~

\

/

OiPr

P =0 CH-

P =O /

iPrO

\

P =O

\

I \

OiPr

iPrO

Figure 18. Synthesis of the diphosphonic acid resin from poly(vinylbenzyl chloride).

OiPr

487 Table 10 Percent Eu(III) complexed from nitrate solutions by 2% DVB styrene/VBC (2:1) gel resin Resin Monofunct. Diphosph. Bifunct. Diphosph. Diphonix *

0.5 h contact time 1 N HNOa 1 N HNOa/0.4 N NaNOa 25.1 27.8 93.0 95.1 98.3

24 h contact time 1 N HNOa 1 N HNOa/0.4 N NaNOa 86.0 86.7 92.3 98.1

2.7. T e m p l a t e p o l y m e r s Template polymers have been prepared using organophosphorus ligands. Template polymerization, also known as molecular imprinting, is used to produce cross-linked polymers containing strategically arranged functional groups [51]. The polymer is prepared by contacting a functionalized monomer with the imprint molecule (the species to be extracted in subsequent applications) and polymerizing with other monomers to yield a crosslinked network. Cavities with predetermined structures remain after the imprint molecule is removed. A high degree of crosslinking is required to keep the polymer chains in a fixed arrangement [41]. Surface template polymerization, prepared in a water-oil emulsion, is ideal for metal ion applications because it allows for water soluble imprint molecules (Figure 19) [41,52-54]. The imprinted resin is used to selectively complex the targeted ion from solutions containing several metals [52,53]. In a representative example with a Cu(II)-imprinted resin prepared from copper-complexed methacrylic acid,

R - O , , OP \~ / R-O O

o ~'

o "P"

O ,,\ , O - R P\ / -0 O-R

o "M" "P"

M ,+ "~

R-O R-O

o

O

" "P\ \ /

O O-R "M" "" "P ,, x / 0 0 "O-R

o "x/,..

Figure 19. Template polymerization.

_M,+~

~

d ,p:

o o _ \\p, x).,

488 competition studies were performed with Cu(II) and Zn(II), Cd(II), or Pb(II). The resin demonstrated high selectivity toward Cu(II) (Table 11). Resins with a random functional group configuration prepared without the presence of the imprint ion during polymerization do not demonstrate copper selectivity (Table 11) [53]. It is important to note, however, that the capacities of the resins are all measured in terms of micromoles per gram: the need for a high cross-link level leads to a diminished loading capacity.

Table 11 Competitive loading with Cu(II)-imprinted and non-imprinted resins from acetate buffer solutions at pH 4.7 and a 2.5 h contact time Resin Imprint Imprint Im print Random Random Random

M1

M2

Cu(II) Cu(II) Cu (II) Cu(II) Cu(II) Cu(II)

Zn(II) Cd(II) Pb (II) Zn(II) Cd(II) Pb(II)

[M1]resin(gmol/g) [M2]resin(gmol/g) 46 49 45 11.9 13.2 6.2

1.4 0.70 2.8 3.1 4.4 12.6

2.8. Solvent i m p r e g n a t e d resins Solvent impregnated resins (SIRs) were developed in order to combine the versatility of solvent exchange with the environmental compatibility of polymersupported reagents. SIRs are prepared by sorbing soluble extractants into solid supports [2]. This type of resin can blend the high distribution ratios and rapid mass transfer of solvent extraction with the operational simplicity of immobilized reagents [55]. A variety of supports have been used including lightly cross-linked gel resins, crosslinked macroporous resins, cellulose, silicate, and polyurethane foams [56,57]. The chosen support must not collapse during the impregnation process to ensure good sorption into the matrix [58]. The complexing agent must be a liquid or form a solution with an organic solvent and have very low solubility in the aqueous matrix; the ligating properties of the extractant must remain intact after sorption [58]. Levextrel resins are similar to solvent impregnated resins except that the extractant is added to the monomer solution before polymerization. The copolymer is formed around the extractant and the sorption step is no longer necessary [2,58]. SIRs have been investigated for their metal ion uptake with a variety of organophosphorus extractants including diamyl amylphosphonate, tributyl phosphate (TBP), 2-ethyhexylphosphonic acid mono-2-ethylhexyl ester (EHPNA), di-(2,4,4trimethylpentyl) phosphinic acid (DTMPPA), di(2-ethylhexyl) hydrogen phosphate (DEHPA), and di(2-ethylhexyl) dithio-phosphoric acid (DEHTPA) (Figure 20)

489

O II (CH~):CH CH:CH:--- P --O ~ CH2CH:CH(CH~): -

I

0

C 6I-I13 I

O =P -- C 6I-II3 OH I

I CH~CH~CH(CH.~L Diamyl amylphosphonate

H -C ~P I

II

H

O

- O - C8H17

o-methyl-dihexylp h o s p h i n e oxide o'-hexyl-2e t h y l p h o s p h o n i c acid

GI~.

0 II H9C4-- O-- P--O--C4H 9

o

II H--C--CH;O-P--O--H C~H~

I

CH~

O I

HCsC--C,H ~

GH9

H EHPNA

TBP

CH~ CIH3 0 I " II H~C- C--CH; C--CH~-P-- O--H "

I

I

CH3

H

CH,

C2~

0 II H - - C, - - C H :- O - P -] - O - - H C~H~ 0 I

H--C--CH,

CH~

I

Hs H3C--C--CH ~ I

I

H

CH~

DTMPPA C4H8 II H2CI--O--P--O--CIH2I

CI4H8

S

C2H5

S

C2H5

I

H DEHTPA F i g u r e 20. E x t r a c t a n t s u s e d in s o l v e n t i m p r e g n a t e d r e s i n s .

DEHPA

490 [15,58-66]. Macroporous polystyrene/DVB is a common support. Diamyl amylphosphonate resins have been used in the extraction of U(VI) from 6.1 N HC1. Complete uptake is observed; there is no measureable radioactivity in the aqueous solution after it is contacted with the resin. Three stripping cycles with 0.025 N HC1 permit essentially complete (>99%) recovery [15]. TBP- impregnated resins have been used in the extraction of U(VI) from 3 N HNO3; rapid complexation kinetics are observed (tim=6 min) [58]. Resins impregnated with DEHTPA have been used to extract Cu(II) from HC1 solutions. Thiourea (0.526 N in 1 N HC1) was used to regenerate the resin (76% Cu(II) recovery); the complexing ability of the regenerated resin was maintained (log D values of 2.98 and 2.83 for the initial and regenerated resins, respectively) [59]. The complexing properties of SIRs are often described in terms of the equilibrium constant, Kex, because the distribution coefficient is dependent on both the pH of the aqueous phase and the extractant concentration in the resin, [HL]r. I~x is described by the following equation for a divalent metal operating via ion exchange; q is the degree of aggregation of the extractant and determines the structure of the extracted species.

[.+I =

Kex --D ([HL]!2+q)) The structure of the extracted species was determined from statistical analysis of the experimental data or assumed to be the same as the solvent extraction system [57,61,62]. Resins impregnated with O-methyl-dihexyl-phosphine-oxide O'-hexyl-2-ethyl phosphoric acid have been used to complex Zn(II), Cd(II), and Cu(II) (Table 12); the extractability of the ions increases with increasing pH [61]. Resins containing DTMPPA have been used for the extraction of Zn(II), Cu(II) and Cd(II) (Table 13). Quantitative recovery of the metal ions is possible when the resin is eluted with 0.1 N HC1. Zn(II) can be separated from Cd(II) and Cu(II) due to the different pH

Table 12 Kex values for o-methyl-dihexyl-phosphine-oxide o'-hexyl-2-ethylphosphoric acid (HL) SIRs; 2 h contact time Metal in O.1 N NaNOa Zn(II) Zn(II) Cu(II) Cu(II) Cd(II) Cd(II)

Extracted species ZnLe (HL) ZnL(NOa) CuL(NOa) CuL(NOa)(HL) CdLz (HL)z CdL(NOa)(HL)

Log Kex -2.83 0.33 0.27 -0.49 -2.69 0.64

491 Table 13 Kex values for DTMPPA (HL) SIRs; 2 h contact time Metal in 0.1 N NaNOa

Extracted species

Zn(II) Zn(II) Cu(II) Cu(II) Cd(II)

ZnL2 ZnL2(HL) 2 CuL2 Cu2(HL)2 CdL2

Log K~x -1.13 -0.67 -4.51 -3.11 -5.01

values required to strip each metal (> 3.1 for Zn(II), > 5.5 for Cu(II), > 6.0 for Cd(II)) [62]. D EHPA SIRs have been used to complex Zn(II), Cu(II), Fe(III), Co(II), Ni(II) and Cd(II) (Table 14) [57,60,63-66]. Complexation is dependent on pH and improves as the pH increases [60]. Separation of metals can be accomplished from aqueous solutions of appropriate acidity. For example, Zn(II) can be separated from Cu(II) in solutions with a pH of 2.50 due to the low Cu(II) affinity of the extractant [66]. EHPNA SIRs have been used in Co(II) and Ni(II) complexation although the Kex values are lower t h a n D EHPA (Table 15) [63]. The complexation rate is directly related to the concentration of extractant in the resin and the usefulness of a SIR depends on the amount of extractant lost during complexation and stripping [58,62,65]. The distribution ratio of the extractant between the organic phase and aqueous phase can be greater for SIRs than solvent

Table 14 Kex values for D EHPA (HL) SIRs Metal Zn(II) Zn(II) Zn(II) Zn(II) Cu(II) Cu(II) Cu(II) Cu(II) Cd(II) Cd(II) Co(II) Ni(II)

a20 h contact time

Extracted species ZnL2 a ZnL2(HL)2b ZnL2(HL)2 c ZnL2(HL)2 c CuL2 a CuL2 b CuL2(HL)2 c CuL2(HL)2 c CdL2 b CdL2(HL)a b CoL2 a NiL2 a

b2 contact time

Matrix H20 0.1 N 0.1 N 0.5 N H20 0.1 N 0.1 N 0.5 N 0.1 N 0.1 N H20 H20

NaNO3 SO42. SO42 NaNOa SO42 SO42 NaNOa NaNOa

c24 h contact time

Log Kex -1.44 - 1.29 - 1.21 -1.77 -2.81 -3.23 -3.56 -4.01 -2.70 -3.10 -3.61 -3.62

492 Table 15 Kex values for EHPNA (HL) SIRs, H20 matrix, 24 h contact time Metal

Extracted species

Log Kex

Co(II) Ni(II)

CoL2 NiL2

-4.95 -5.34

Table 16 Distribution ratio of DTMPPA in polystyrene a and tolueneb; 0.05 N NaNO3 Medium

KD

Polystyrene Toluene

1070 44.7

0.2 g resin/20 mL aqueous phase; 3 h contact time b 0.01- 0.075 N DTMPPA in toluene; equal volumes of toluene and aqueous phase.

a

extractant systems because the complexant interacts with the polymer matrix (Table 16) [60,62]. Akita and Takeuchic report negligible DEHPA loss from the SIR with aqueous HC1 (pH 2.50) elution; the resin was evaluated with a cyclic column operation using 1 N HC1 as the stripping agent. The decrease in the total sorption capacity was 7% at the fifth cycle. When the aqueous phase was changed to acetate buffer with a pH of 4.50, a 25% decrease in the total sorption capacity was observed at the third cycle due to greater extractant solubility [66]. The importance of optimizing the solution phase conditions must thus not be overlooked when applying SIRs to metal ion applications. CONCLUSION Phosphorus-based complexing agents have been critical to the development of solvent extraction. Extension of the analogous ligands to polymer-supported reagents will play an ever-increasing role in metal ion separation processes. Their ability to complex metal ions through ion exchange and coordination, depending upon the solution pH, is indicative of their versatility and continuing utility.

ACKNOWLEDGEMENT We gratefully acknowledge the support of our research by the United States Department of Energy, Office of Energy Research, Division of Chemical Sciences, Office of Basic Energy Sciences, through grant DE-FG05-86ER13591. REFERENCES 1. M.C. Kavanaugh, Environ. Progress, 14 (1995) M3.

493 2. L.L. Tavlarides, J.H. Bae and C.K. Lee, Sep. Sci. Technol., 22 (1987) 581. 3. B. Weaver, in: Ion Exchange and Solvent Extraction, Vol. 6; J.A. Markinsky and M. Yizhak (eds.), Marcel Dekker, New York, 1974, Chapter 4. 4. E.P. Horwitz, D.G. Kalina and A.C. Muscatello, Sep. Sci. Technol., 16(1981) 403. 5. A.K. De, S.M. Khopkar and R.A. Chalmers, Solvent Extraction of Metals; Van Nostrand Reinhold, London, 1970, Chapter 9. 6. R.T. Paine, G.S. Conary, A.A. Russell, D.J. McCabe, E.N. Duesler, S. Karthikeyan, R. Schaeffer, R.R. Ryan and D. Cromer, Solv. Extr. Ion Exch., 7 (1989) 767. 7. M.J. Herak and V. Jagodic, J. Inorg. Nucl. Chem., 35 (1973) 995. 8. V. Jagodic and M.J. Herak, J. Inorg. Nucl. Chem., 33 (1971) 2641. 9. J. Rado~evic, V. Jagodic and M.J. Herak, J. Inorg. Nucl. Chem., 39 (1977) 2053. 10. D.G. Kalina, E.P. Horwitz, L. Kaplan and A.C. Muscatello, Sep. Sci. Technol., 16 (1981) 1127. 11. E.P. Horwitz, D.G. Kalina, L. Kaplan, G.W. Mason and H. Diamond, Sep. Sci. Technol., 17 (1982) 1261. 12. J.A. Daoud, M. Gasser and H.F. Aly, Radiochim. Acta, 69 (1995) 197. 13. A. Tok~rov~, O. N~vratil and P. Sl~dek, Collect. Czech. Chem. Commun., 61 (1996) 1177. 14. O. N~vratil and J. Vykoupil, Collect. Czech. Chem. Commun., 42 (1977) 2126. 15. E.P. Horwitz, M.L. Dietz, R. Chiarizia and H. Diamond, Anal. Chim. Acta, 266 (1992) 25. 16. N. Zakareia, S.M. Khalifa, M. Nofal and H.F. Aly, Radiochim. Acta, 47 (1989) 229. 17. E.G. Joe, G.M. Ritchey and A.W. Ashbrook, J. Metals, 18 (1966) 18. 18. A.M. Spasic, N.N. Djokovic, M.D. Babic, M.M. Marinko and G.N. Jovanovic, Chem. Eng. Sci., 52 (1997) 657. 19. Y. Komatsu and H. Freiser, Anal. Chim. Acta, 227 (1989) 397. 20. S.A. Sayed and A.T. Kandil, Radiochim. Acta, 73 (1996) 207. 21. P. Hodge, in: Syntheses and Separations Using Functional Polymers, D.C. Sherrington and P. Hodge (eds.), Wiley, New York, 1988, Chapter 2. 22. J. Kennedy, E.S. Lane and B.K. Robinson, J. Appl. Chem., 8 (1958) 459. 23. S.D. Alexandratos and L.A. Hussain, Ind. Eng. Chem. Res., 34 (1995) 251. 24. M.C. Yebra-Biurrun, A. Bermejo-Barrera and M.P. Bermejo-Barrera, Anal. Chim. Acta, 264 (1992) 53. 25. H. Egawa and T. Nonaka, J. Appl. Polym. Sci., 30 (1985) 3239. 26. J. Furukawa, E. Kobayashi and T. Wakui, Polymer J., 12 (1980) 293. 27. E. Bayer, P-A. Grathwohl and K. Geckeler, Angew. Makromol. Chem., 113 (1983) 141. 28. S.D. Alexandratos and D.L. Wilson, Macromolecules, 19 (1986) 280. 29. H. Egawa, K. Yamabe and A. Jyo, J. Appl. Polym. Sci., 52 (1994) 1153. 30. H. Egawa, T. Nonaka and M. Ikari, J. Appl. Polym. Sci., 29 (1984) 2045. 31. A.W. Trochimczuk and S.D. Alexandratos, J. Appl. Polym. Sci., 52 (1994) 1273.

494

32. S.D. Alexandratos and D.R. Quillen, React. Polym., 13 (1990) 225. 33. A. Jyo, K. Yamabe and H. Egawa, Sep. Sci. Technol., 31 (1996) 513. 34. M. Asker, R.Y. Wan, J.D. Miller, D.R. Quillen and S.D. Alexandratos, Metallurgical Transactions B, 18 B (1987) 625. 35. J.D. Miller, R.Y. Wan, M.B. Mooiman and P.L. Sibrell, Sep. Sci.Technol., 22 (1987) 487. 36. R.T. Paine, S.M. Blaha, A.A. Russell and G.C. Conary, Solv. Extr. Ion Exch., 7 (1989) 925. 37. L. M~nard, L. Fontaine and J-C. Brosse, React. Polym., 23 (1994) 201. 38. S.D. Alexandratos, M.A. Strand, D.R. Quillen and A.J. Walder, Macromolecules, 18 (1985) 829. 39. S.D. Alexandratos, D.R. Quillen and W.J. McDowell, Sep. Sci. Technol., 22 (1987) 983. 40. S.D. Alexandratos, D.R. Quillen and M.E. Bates, Macromolecules, 20 (1987) 1191. 41. M. Yoshida, K. Uezu, M. Goto and F. Nakashio, J. Chem. Eng. Jpn., 21 (1996) 174. 42. S.D. Alexandratos and D.R. Quillen, Solv. Extr. Ion Exch., 7 (1989) 511. 43. M.A. Congost, D. Salvatierra, G. Marquis, J.L. Bourdelande, J. Font and M. Valiente, React. Funct. Polym., 28 (1993) 191. 44. M. Fujiwara and T. Matsushita, Anal. Chim. Acta, 274 (1993) 293. 45. S.D. Alexandratos, D.W. Crick and D.R. Quillen, Ind. Eng. Chem. Res., 29 (1991) 772. 46. S.D. Alexandratos and M.E. Bates, Macromolecules, 21 (1988) 2905. 47. S.D. Alexandratos, A.W. Trochimczuk, D.W. Crick, E.P. Horwitz, R.C. Gatrone and R. Chiarizia, Macromolecules, 29 (1996) 1021. 48. E.P. Horwitz, R. Chiarizia, R.C. Gatrone, S.D. Alexandratos, A.W. Trochimczuk and D.W. Crick, Solv. Extr. Ion Exch., 11 (1993) 943. 49. R. Chiarizia, E.P. Horwitz, R.C. Gatrone, S.D. Alexandratos, A.W. Trochimczuk and D.W. Crick, Solv. Extr. Ion Exch., 11 (1993) 967. 50. S.D. Alexandratos, A.W. Trochimczuk, E.P. Horwitz and R.C. Gatrone, J. Appl. Polym. Sci., 61 (1996) 273. 51. P.K. Dhal and F.H. Arnold, Macromolecules, 25 (1992) 7051. 52. M. Murata, S. Hijiya, M. Maeda and M. Takagi, Bull. Chem. Soc. Jpn., 69 (1996) 637. 53. W. Kuchen and J. Schram, Angew. Chem. Int. Ed. Engl., 27 (1988) 1695. 54. K. Uezu, H. Nakamura, M. Goto, M. Murata, M. Maeda and M. Takagi, J. Chem. Eng. Jpn., 27 (1994) 436. 55. J.L. Cortina, N. Miralles, A.M. Sastre, M. Aguilar, A. Profumo and M. Pesavento, React. Polym., 21 (1993) 103. 56. K. Takeshita, Y. Takashima, S. Matsumoto and H. Yamanaka, J. Chem. Eng. Jpn., 28 (1995) 91. 57. R-S. Juang and J-Y. Su, Ind. Eng. Chem. Res., 31 (1992) 2774. 58. A. Warshawsky and A. Patchornik, Israel Journal of Chemistry, 17 (1978) 307.

495 59. A.G. Strikovsky, K. Jer~bek, J.L. Cortina, A.M. Sastre and A. Warshawsky, React. Funct. Polym., 28 (1996) 149. 60. J.L. Cortina, N. Miralles, M. Aguilar and A.M. Sastre, Solv. Extr. Ion Exch., 12 (1994) 371. 61. J.L. Cortina, N. Miralles, M. Aguilar and A. Warshawsky, React. Funct. Polym., 27 (1995) 61. 62. J.L. Cortina, N. Miralles, A. Sastre, M. Aguilar, A. Profuma and M. Pesavento, React. Polym., 18 (1992) 67. 63. S. Akita and H. Takeuchi, J. Chem. Eng. Data, 37 (1992) 303. 64. R-S. Juang and H-C. Lin, J. Chem. Tech. Biotechnol., 62 (1995) 141. 65. R-S. Juang and M-L. Chen, Ind. Eng. Chem. Res., 36 (1997) 813. 66. S. Akita and H. Takeuchi, J. Chem. Eng. Jpn., 26 (1993) 442.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

497

A p p l i c a t i o n of t h e i o n - e x c h a n g e m e t h o d to r e m o v e m e t a l l i c i o n s f r o m waters and sewages Z. Hubicki a, A. Jakowicz a and A. Lodyga b aDepartment of Inorganic Chemistry, Faculty of Chemistry, Maria Sklodowska University, 20-031 Lublin, Poland

Curie-

bFertilizers Research Institute, Putawy, Poland

Industries throughout the world are faced with the ever-increasing need to reduce the heavy metal content of their discharge streams to acceptable levels. Several methods are available to achieve this objective. Among these methods are those involving the use of ion-exchange and adsorbent resins. This paper indicates the types of resins available, some of the main principles involved, and illustrates the use of a few such methods with specific examples. Particular emphasis is placed on those methods which achieve selective removal of relatively low levels of toxic or noble metal ions. Owing to the ion-exchange method it is possible to remove all ions from the solution or to separate one substance from another which allows for selective removal of ionic impurities and complete deionization. Choice of one of the above possibilities depends first of all, on the composition of the solution and the extent of required impurities removal. Used up plating baths and post-production solutions are sewages containing large amounts of various ions. Such great concentration of contaminating metal ions constitutes a serious threat for the natural environment. The ion-exchange method proves to be effective in solution of this problem. Owing to it, the contaminating ions are removed to such a degree that the used up bath or brine can be applied again and sewages do not impose any t h r e a t for the natural environment after the selective removal of the toxic ions [1-7]. Development and broader application of the ion-exchange method proved possible when modern organic and inorganic ion-exchangers of suitably choosen physicochemical properties were produced and used [5,8-14]. Main trends in the development of conventional ion exchangers (polystyrenesulphonic cation exchangers and strongly basic anion exchangers) referred to great ion-exchange capacity achievement as well as improvement of chemical, thermal and mechanical resistance. Better kinetic abilities of ionexchangers were obtained as a result of macroporous ion exchangers production the seventies [10-13].

498 Rate of ion-exchange and level of concentrations of the purified salts are, as a rule, much smaller compared with those of extraction. That requires a suitably large size of industrial device and high operating costs which limits applicabili@ of ion-exchange processes in production of high quality materials and in selective removal of toxic and noble metal ions from sewages. These inconveniences can be avoided using various impregnants (styrenedivinylbenzene copolymers impregnated by means of liquid ion exchangers or extractants) [15-25]. However, only the method worked out by the firm Bayer, Germany is applied. The disadvantage of impregnants is too slow kinetics resulting from the limited interfacial size and its poor wettability by aqueous solutions. The main drawback is loss of the active substance during purification. Improvement of ion-exchange capacity is possible by so called interpolimerization [10, 26]. Better results should not be expected here. Ion-exchange capacity limitation caused the tendency to improve ion-exchange selectivity achieved by the synthesis of some selective ion exchangers characterized by differentiated affinity for definite ions or their groups. 1.

F A C T O R S A F F E C T I N G ION-EXCHANGE S E L E C T I V I T Y

Selectivity is the ability of choosing some ions rather than others. Ionexchange separations are based on selectivity is of significant importance for predicting the course of ion-exchange reactions and their practical application. The factors affecting the ion-exchange selectivity are, among others, charges of ions being exchanged, types of functional groups as well as specific interactions in the ion exchanger, composition of ion exchanger phase, crosslinking, temperature and formation of complexes in the solution. Functional groups of most commonly used strongly acidic cation exchangers (SOa-) and strongly basic anion exchangers (-CH2-N(R1R2Ra) § show poor complex forming abilities owing to which there are observed ,,normal" affinity series in which the factors responsible for the position of a given ion are its charge, radius and hydration extent [10,12,13,27-30]. The affinity of polystyrenesulphonic cation exchangers for cations increases with the increase of charge. In case of various cations of the same charge their affinity for ion exchangers generally decreases with the increase of the hydrated ion radius [10,12,13,27-30], e.g. for the polystyrenesulphonic cation exchanger affinity of alkali and alkaline earth metals decreases in the series: Cs + > Rb + > K + > Na + > Li + Ba 2+ > Sr 2+ > Ca 2+ > Mg2 + In case of strongly basic anion exchangers with quaternary ammonium groups, the charge of anion exchanger affects the affinity of the anion exchanger as in the case of polystyrenesulphonic cation exchangers. The affinity decreases in the following order:

499 C104- > SCN- > I- > HSO4- > NO3- > Br- > CN- > NO2- > C1- > OH- > FIn case of ion exchangers of different functional groups the problem is more complicated due to the interaction of counterions, with ion exchanger functional groups resulting in formation of ion-pairs and even of covalent bonds. A typical example is very strong affinity of the anion exchangers of the functional groups of primary-, secondary- and tertiary amines for OH- ions contrary to the strongly basic anion exchangers for which the OH- ion is situated almost in the end of the series. The order of affinity of other ions for weakly basic anion exchangers is analogous to strongly basic ion exchangers [10,28-31]. Carboxylic ion exchangers are characterized by strong affinity for the H § ion contrary to strongly acidic cation exchangers [10,28-31]. On carboxylic ion exchangers the affinity series for alkali and alkaline earth metals are as follows: Li § > Na § > K § > Rb § Cs § Mg2 § > Ca2+ > Sr 2§ > Ba2+ that is contrary to those observed for polystyrenesulphonic cation exchangers. Carboxylic ion exchangers are characterized by strong affinity for alkaline earth metals and some transition elements which can be referred to the behaviour of aliphatic or aromatic carboxylic acids forming sparingly soluble salts with these ions. The first resin of the complexing functional groups was synthesized by Shogseid fifty years ago [32]. It was the ion exchanger with dipicrylamine built in the matrix and exhibiting selectivity for potassium. Shogseid used this ion exchanger to remove potassium from the sea water. Based on the solubility series of dipicrylamine with alkali metals, one can expect that the ion exchanger of this type should possess stronger affinity for rubidium and cesium ions. At present there are known several scores of various ion exchangers with complexing groups built in the matrix and exhibiting selectivity for various ions or their groups. Table 1 includes well known and commonly applied both on laboratory and commercial scales selective ion exchangers.

Table 1 Selective ion exchangers Functional group type

Selective for ions

Alamine Amidoxime

U(IV, VI) [34-45], Au(III), Ru(III),

Aminoguanidine 2-amine-4-ketothiazole- 3-acetate

Rh(III), Pd(II, IV), Pt(IV), Ir(III, IV), Cu(II), Fe(III), Cd(II), Hg(II) [46-53] Au(I) [64,65] Ag(I), Hg(II) [66]

Cu(II) [33]

500 Table 1 Selective ion exchangers Functional group type Aminophosphonic

2-amine-4-ketothiazole-3-acetate Anthranilate Arsenazo Arsenic Chitosan Chromotropic Cysteine Diphenylcarbazide Diphenylcarbazone 1,8-dihydroxynaphtalene-0,0-diacetate Dimethylglyoxime Dipicrylamine Dithiocarbamate Dithizone Fluorene Formazane Phosphine Phosphone

Phosphate Glyoxal-bis-2-mercaptoaniline Guanidine 8-hydroxyquinoline o-hydroxyphenylazobenzoate Hydrazide Hydroxamate Imidazole

Selective for ions Sc(III) [30,54], Th(IV) [30,33,55], U(IV,VI) [30,33,56-58], Pb(II),Cu(II), Zn(II), Fe(III) [59-60], Ga(III), In(III) [62-63] Ag(I), Hg(II) [66] Cu(II), Co(II), Ni(II), Fe(II), Zn(II) [67,68] U(VI), Th(IV), Zr(IV), Hf(IV), Ln(III) [33,69] U(VI), Th(IV), Zr(IV), Hf(IV), Ln(III) [30,33] Cu(II), Ni(II), Co(II), Ga(III), Ln(III), Nd(III), Eu(III) [70] Ti(IV) [71], Fe(III), Cu(II), Zn(II)[33] Ag(I), Pt(II), Au(III), Hg(II) [33] Cr(VI) [72-75] Ag(I), Hg(II) [76] Zr(IV), Hf(IV) [77] Ni(II) [78] K(I) [32] Hg(II), Ag(I), Cu(II), Se(IV), Cr(VI), Pb(II) [5,33,79-85] Au(III), Pt(II, IV), Pd(II), Ir(III, IV), Ru(III), Rh(III), Ag(I), Hg(II) [86-88] Sc(III) [89] Pd(II), Ir(IV), Pt(II, IV), Au(III), Ag(I), Hg(II), U(VI)[33,90] Wh(IV), U(VI), Fe(III) [10,33,91] Sc(III) [30,33,54], Th(IV), U(IV, VI) [30,33,44,92,93], Zr(IV), Hf(IV) [94], Fe(III), Zn(II)[30,33,95-97], Ga(III), In(III) [30,33,62,97], Be(II)[30,33] Wh(IV), U(VI) [30,33,98,99] Hg(II) [100] Pd(II), Pt(IV), Au(I) [33,64,65] Cu(II), Fe(III), V(V) [33] Be(II), AI(III), Fe(III) [33] Ag(I), Hg(II) [33] U(VI), Fe(III) [33,85,101-103] Au(I) [104-106]

501 Table 1 Selective ion exchangers Functional group type Iminodiacetate

Crown or cryptands 8-mercaptoquinoline Mercaptide n-methylglucamine Morine Nitriltriacetate r Nitrosorezorcinole Oxime Piridinedicarboxylate Porphyrine Pyrocatechol-0,0-diacetate Rhodamine Salicylate Sarcosinate Tannine Thiazoline Thioglycolate Thiophosphorate Thiourea Thionalide Triazolethiol 2,4,6-triamine- 1,3,5-triazyne

Selective for ions Cu(II), Pb(II), Ni(II) [10,30,33,107,110], Cd(II) [30,33,60,107], Fe(III) [10,30,33,62,107-111], Ln(III) [30,33,107,112-114], Sc(III) [30,33,54], Th(IV) [99], U(VI) [33,98,107], In(III), Ga(III) [30,33,62,115-117], V(IV, V) [118], Se(IV) [119] K(I), Ba(II) [30,33,85] Pd(II), Au(III), Ag(I) [30,33,85] Ag(I), Hg(II), Au(III) [30,33,85] B(III) [33] Hf (IV), Zr(IV), Fe(III) [120] Cu(II), Ln(III) [33] Co(II) [33,107] Fe(III), Cu(II), Co(II), Hg(II)[33] Cu(II), Mo(VI) [33] Cu(II), Ca(II)[33] Ag(I), Cu(II), Co(II), Fe(II)[30,33,84] Zr(IV), Cu(II) [121] Ag(I), Au(III), Pd(II), Pt(IV) [122] U(VI), Fe(III), Cu(II), Hg(II) [10,33,107,123] Cu(II) [33,107] Ge(IV) [33] Hg(II) [33,85] Ag(I), Au(III), Hg(II), Bi(III), Sn(IV) [30,33,85] Hg(II), Pg(II), Cu(II)[33] Ag(I), Au(III), Pt(II, IV), Pd(II), Ru(III), Hg(II) [10,30,33,84,85,107,124], Cd(II), Cu(II), Co(II), Ni(II), Zn(II) [126], As(III, V) [127-128], Sb(III, V) [129,130], Bi(III) [131], Pd(II) [125] Hg(II), Ag(I), Cu(II), Cd(II)[132] Cu(II), Hg(II), Cd(II), Ag(I), P(V), As(V), Cr(VI) [33]

The detailed discussion of physicochemical properties and applications of individual selective ion exchangers presented in Table 1 can be found in many

502 papers [5,8-11,29,30,32-132]. Of various well known chelating ion exchangers only a few types are produced on the commercial scale. These ion exchangers are included in Table 2.

Table 2 Main commercial selective ion exchangers Type

Resin Name

Manufacturer

Amidoxime

Chelite N

Serva, Feinbiochemica GmbH and Co., Germany Rohm and Haas Company, France Not available former USSR

Aminophosphonate

Dithiocarbamate Hydroxyoxime Iminodiacetate

Duolite ES-346 ANKF-2G ANKF-3G ANKF-221 AEF-1 AEF-2 AEF-3 Chelite P Duolite ES-467 Duolite ES-469 Lewatit OC- 1060 Purolite S-940 Purolite S-950 Relite MAC-7 Wofatit ME-55 Nisso ALM-525 Dowex XF-4196 Amberlite IRC-718 ANKB-1 ANKB- 10 ANKB-35 ANKB-50 Chelex 100 Chelite C Diaion CR- 10 Dowex A-1 Dowex XF-4045 Duolite ES-466 Lewatit TP-207 Lewatit TP-208

Serva Feibiochemica GmbH and Co., Germany Rohm and Haas Company, France Bayer, Germany Purolite International Ltd., United Kingdom Resindion, Italy Bayer, Germany Nippon Soda Co. Ltd, Japan Dow Chemical Co., USA Rohm and Haas Company, France Not available former USSR

Bio-Rad, USA; Merck, Germany Serva Feibiochemica GmbH and Co., Germany Mitsubishi Chemical Industries Ltd, Japan Dow Chemical Co., USA Rohm and Haas Company, France Bayer, Germany

503 Table 2 Main commercial selective ion exchangers Type Resin Name Muromac A- 1 Purolite S-930

Isothiouronium

Relite MAC-5 Unicellex UR- 10 Varion CH Wofatit MC-50 Wofatit MC-59 Wofatit Y-66 Duolite ES-345 Ionac SR-3 Ionac SRXL Lewatit TP-214 Monivex Purolite S-920 Relite MAC-3 Srafion NMRR

Picolylamines Poliethylenepolyamine

Polietylenepolyimine Pyridine Thiol

Dowex XF-4195 Dowex XF-43084 Diaion CR-20

Manufacturer Muromachi, Japan Purolite International Ltd., United Kingdom Resindion, Italy Unitika Ltd, Japan Nitrokemia Ipartelepek, Hungary Bayer, Germany Rohm and Haas Company, France Sybron Chemicals Incorporated, USA Bayer, Germany Ayalon Water Conditioning Company Ltd., Israel Purolite International Ltd., U.K. Residion, Italy Ayalon Water Conditioning Ltd., Israel Dow Chemical Co., USA

Mitsubishi Chemical Industries Ltd., Japan Sumichelate MC-10 Sumimoto Chemical Co., Ltd, Japan Mitsubishi Chemical Industries Diaion CR-40 Ltd, Japan Sumimoto Chemical Co., Ltd, Sumichelate CR-2 Japan Serva Feinbiochemica GmbH and Chelite S Co., Germany Rohm and Haas, France Duolite ES-465 Imac TMR Imac GT-73

The selective ion exchangers are produced in the form of granules, membranes, discs, fibres, ion-exchange or ionited paper. As a relatively small number of selective ion exchangers is available commercially, many authors suggest modification of strongly basic anion

504 exchangers by means of aromatic chelating sulpho-derivative reagents like Alizarin S [143-146] Arsenazo I [145], Arsenazo III [145], Beryllon II [145], Bismuthon II [147], Bromopyrogallol Red [143-145], Chromazurol S [145], Chromotrope 2R [148], Chromotropic Acid [144,145], Ferron [144,145], NitrosoR-salt [144,145,149], Pyrocatechol Violet [145], SPADNS [145], 5-Sulphosalicylic acid [145], Tiron [145], Thoron [145], Xylenol Orange [145,150] etc. [151-156]. Complexing organic reagents used for modification transform the anion exchanger into the selective ion exchanger. According to Brajter [143,144] the modified anion exchanger can be characterized by greater selectivity for definite ions or their groups than the corresponding ion exchanger with the analogous functional group built into the matrix. In most cases the separation processes carried out on these ion exchangers are of analytical (separation of micro or miligram amounts, sorption of microgram amounts - trace analysis) or physicochemical character. 2.

S E L E C T I V E REMOVAL OF ION I M P U R I T I E S

It is often sufficient to remove only one type of metal ions from the mixture of impurities contained in sewages. Owing to the ion exchange it is possible to replace undesirable ions with others which are harmless for the natural environment. 2.1. Gold a n d p l a t i n u m m e t a l s Noble metal recovery is a typical example for application of this method. The main aim is total recovery of noble metals but not the reuse of the water which is of less economic importance as the volume of the produced sewages is relatively small. Some used up plating baths contain gold in the form of the anion complex [Au(CN)2-]. Both strongly and weakly basic anion exchangers can be applied to remove this anion from sewages [5,85,105,157]. To remove the complex [Au(CN)2-] in the whole pH range there can be applied strongly basic anion exchangers possessing quaternary ammonium groups:

(R4N+)2SO42- + 2[Au(CN)2]- ~

2R4N+[Au(CN)2]- + 3042-

(1)

The working exchange capacities strongly basic anion exchangers for the complex [Au(CN)2-] are low and equal 55-110 g Au/dm 3 [5,85,105,157]. It is difficult to strip the complex [Au(CN)2-] from the anion exchanger bed by means of alkaline metal hydroxide solutions. Due to high prices of gold, the value of that adsorbed on the anion exchanger exceeds the price of the ion exchanger, ion exchanger combustion at 1000~ in order to recover metal is quite common. Also the complex [Zn(CN)4] 2-, acidified thiourea solution or potassium thiocyanate solution in an organic solvent, can be used for desorption of gold(I) from the anion exchanger bed. The complex [Zn(CN)4] 2- characterized by stronger

505 affinity for the ion exchanger than [Au(CN)2]-is effective for gold recovery from the anion exchanger bed. The regeneration process is as follows: 2R4N+[Au(CN)2] - + [Zn(CN)4]2 ~

(R4N+)2[Zn(CN)4]2- + 2[Au(CN)2]-

(2)

To remove gold from effluents cementation by means of e.g. lead can be used [158]. The cementation process is as follows: 2Au[CS(NHD2]C1 + Pb ~

Pb[CS(NHD2]nC12 + 2Au

(3)

where most probably n equals 4. For desorption of [Au(CN)2]-from the anion exchanger potassium thiocyanate solutions in the water-organic solvent mixtures were used as eluents [159]. The most frequently used mixtures are: water-DMF, water-dimethylacetamide, water-acetone, water-N-methyl-2-pyrrolidine, water-DMSO, water-tetrahydrofurane, water-hexamethylphosphoramide. The 5 M KSCN solution in 50% v/v DMF in water is the most frequently used eluent. The sorption mechanism of the complex [Au(CN)2]- on the weakly basic anion exchanger can be presented in the following way: (R3NH+)2SO4 + 2[Au(CN)2]- ~

R3NH§

- + SO42

(4)

Weakly basic anion exchangers are selective for [Au(CN)2]-, though sorption of cyanide complexes of silver(I), copper(II), iron(II, III), nickel(II) and others can be taken into account [5,85,105,157]. Applicability of some available weakly basic anion exchangers was examined. It proved that the most effective weakly basic anion exchangers are those whose pKa values of functional groups are from 9 to 11. These anion exchangers achieve the maximal sorption of the complex [Au(CN)2]- in the pH range 7-9 which allows for avoiding the problems connected with HCN removal. Regeneration of the weakly basic anion exchanger with deposited gold(I) cyanide complexes is carried out by means of diluted alkaline metal hydroxide solutions. The regeneration process can be presented by means of the following equation: (R3NH+)2SO4 + 2[Au(CN)2]- ~

R3NH+[Au(CN)2] - + 3042.

(5)

The two-column system was applied to remove gold from sewages containing 4-5 mg Au/dm 3 after the plating process [160]. In the first column there was the carboxylic cation exchanger in the H § form (Diaion WK-20) which was probably used to remove other heavy metal ions and to decrease pH. In the second one there was the weakly basic macroporous anion exchanger Lewatit MP-62 on which the gold anion complex adsorbed. The NaCN alkaline solution was used for regeneration of the anion exchanger bed. The regeneration process gave almost 100% gold recovery.

506 In another procedure the post-plating solution containing, besides the gold anion complex, organic acids and their salts was passed through the weakly basic anion exchanger at pH equal about 3 in the three-column system [161]. Gold was adsorbed in the first of them and other substances in the second and third ones. Moreover, selective ion exchangers of various types are used for selective removal of noble metals from sewages. Thus the ion exchanger with the cysteine groups is characterized by high ion-exchange affinity for Au(III) and Pt(II). The ion-exchange capacities for these ions are as follows: Au(III)=l.22 M/kg and Pt(II)=0.39 M/kg [162]. The selective ion exchanger with hydroxamate groups is characterized by great ion-exchange capacity for Au(III) which is 4.0 M Au/kg (pH=l). This sorbent is recommended for selective separation of Au(III) from Cu(II), Ag(I) and Fe(III) and also for removal of Au(III) from sea water and KCN solutions [163]. The ion exchanger with the isonitrosoacetamide groups is selective for Pd(II) [164]. The sorbent of the azoimidazole functional groups is used for sorption of noble metal ions from the solutions of copper(II), nickel(II), cobalt(II), iron(III), aluminum(III) etc. salts [33]. It possesses very great ionexchange capacity for gold which is equal to 660 g Au/kg and is applied for removal of noble metals from the solutions coming from copper and nickel hydrometallurgy as well as from past plating sewages [33]. The sorbent based on polystyreneazorhodamine is stable up to 100~ both in neutral and acidic solutions. Its sorption capacity for the selected noble metal ions in the 1 M HC1 solution is: Pt(II)-19.5 g/kg, Pd(II)-22.2 g/kg and Au(III)-197 g/kg. This ion exchanger is characterized by great selectivity for noble metal ions [33]. The sorbent based on polystyrenazo-8-mercaptoquinoline is stable in the acidic and neutral media. It exhibits great ion-exchange capacity for Pd(II) (which is 300 g Pd(II)/kg in the 1 M HC1 solution) and great selectivity for noble metal ions in the presence of Cu(II), Ni(II), Co(II), Fe(III) etc. [33]. Great selectivity for the complex [Au(CN)2]-is characteristic for the chelating ion exchanger of the guanidine and aminoguanidine functional groups built into various polymer matrices [33,64,65]. It proved that modified hydrophobic copolymers crosslinked with DVB of the expanded structure are the most effective for the sorption of [Au(CN)2]- ions. The ion exchanger of formazane functional groups exhibits great selectivity for gold, silver, platinum metals and mercury ions. Its ion-exchange capacity in 0.01 M HC1 is following: Au(III)0.9 M/kg, Pd(II)-0.75 M/kg, Pt(II, IV)-0.54 M/kg, Ir(IV)-0.43 M/kg,Rh(III)0.1 M/kg and for Ru(III)-0.03 M/kg. This sorbent is recommended for selective removal and separation of noble metal ions as well as for separation of Pd(II) from Ni(II) and Co(II) [33,90,165,166]. Moreover, the selective ion exchangers of dithiocarbamate, thiamide and thioglycole functional groups exhibit great selectivity for gold and platinum metals [30,33,85]. The results of the studies on synthesis and physicochemical properties of the ion exchangers of the dithizone and dehydrodithizone functional groups bonded to the polystyrene matrix by the sulphur atom are very interesting. The ion exchanger of the dithizone functional groups exhibits particularly strong ion

507 exchange affinity for Pd(II), Pt(II) and Au(III). The values of distribution coefficients IQ for Pd(II), Pt(II) and Au(III) in the 0.01-6.0 M HC1 solutions are very high and equal 104-106. This ion-exchanger can be applied both for separation of platinum metals and selective separation of noble metal ions from others [86-88]. Complexes of gold(III) chlorides can be removed from various systems on the sorbents possessing weak ester groups e.g. on Amberlite XAD-7. In this case two types of sorption mechanism can be distinguished [10,167,168]: - solvation: R - CO2 + [nuCl4]- ~ R - CO2[AuC13] + C1-ion-exchange: R - C 0 2 + H20 ~ R - C O O H + + OH-

(6) (7)

R - C O O H + + [AuC14]- ~

(s)

R-COOH+[AuCI4] -

The mixture of hydrochloric acid and acetone can be used for gold(III) desorption. The resins containing thiourea or its derivatives built into matrix exhibit great selectivity for noble metal ions [5,10,30,33,85]. The commercially available ion exchanger selective for noble metals and mercury ions of the isothiourea functional groups (Srafion NMRR) exhibits differentiated ion-exchange capacity for gold and platinum family. Total capacity of this ion exchanger for individual noble metal ions is as follows: Au-340 g/dm ~, Pt-200 g/dm 3, Pd-120 g/dm 3, Ir-120 g/dm ~ and Rh-80 g/dm 3. However, the working capacity of Srafion NMRR for gold ions is small (55-110 g Au/dm3). As the value of the gold retained in the bed exceeds significantly that of the ion exchager used for its recovery, combustion of the ion exchanger at 1000~ is a common procedure, though the 5% NaOH + 5% NaCN solution is also used for its regeneration. The selective ion exchangers of isothiourea groups are used, among others, for separation of platinum metals from other metals. Depending on pH of the medium these goups are of the following forms:

P ~ C H ~ - - S ~ J NH

..

"

P~CH~--S~C

\NH 2 (1)

/NH2 (~ X \NH 2

(9)

(2)

When these groups occur in form (1), they create coordination bonds with metal cations and in form (2) the platinum metal anion complexes are bonded according to the exchange mechanism e.g. [PdC14]2-.

508 NH

2 P ~ CH~-- S--C./ @ 2X

+

_

[ PdC] ]24

\NH 2

(10) / NH P ~ c g T- S - - C \ | NH

H2N \

[PdCJ] 2

H N 2

Srafion NMRR is used for the group removal of noble metals from ores, meteorites, moon, sand, steel, sewages, biological materials etc.[10,33,84,85,124]. 2.2. S i l v e r Silver is another precious metal commonly applied in plating and photographic industries. In plating processes it occurs as the cyanide complex [Ag(CN)2]- with various additions. The ion-exchange methods are recommended for silver recovery from large volumes of sewages of low concentrations of the element being recovered, particularly from the washings. The first technological system for this process was installed in Germany at the beginning of the forties. The silver contained in the washings formed the cyanide complexes which sorbed on the weakly basic anion exchanger in the sulphate form. The NaOH and H2S04 solutions were used successively for the ion exchanger regeneration. The sorption capacity of the ion exchanger was relatively small and was equal to 8 g/dm 3 [169]. Some anion exchangers of various types were applied for the selective removal of silver(I) from cyanide solutions by Riveros [170]. It proved that the anion exchangers like Duolite A-7, Dowex WGR-2, Dowex MWA-1, Dowex XFS-4195 and Amberlite IRA-35 extract silver(I) in large amounts only at pH<8. However, at pH>8 the same ion exchangers extract silver(I) from the solution to a small extent or they do not extract it at all. The comparison of the above mentioned ion exchangers indicates that Duolite A-7 and Dowex MWA-1 possess the highest capacity and effectiveness. Strongly basic anion exchangers like Dowex MSA-1 and Amberlite IRA-910 extract silver(I) in the whole range of pH 4-12. Moreover, it was stated that if the solution under consideration contains only one kind of metal ions, then the given ion exchangers show the highest extraction degree for silver(I). However, if in the feeding cyanide solution the metal ion mixture is present, the percentage of silver(I) removal decreases which depends on the phase contact time. The extent of silver(I) extraction can be increased decreasing the concentation of free cyanides in the solution and also by stopping the extraction before the ion-exchanger reaches the highest saturation [170]. In 1968 Kunin suggested using the strongly basic anion exchanger (type 1) to recover silver(I) being in the form ([Ag($203)2] m from the photographic industry sewages. It was Amberlite IRA-400 in the chloride form regenerated with 1 M

509 hydrochloric acid solution [169]. The studies carried out in the USA in the end of seventies showed that the weakly basic acryldivinylbenzene anion exchanger Amberlite IRA-68 is characterized by better qualities than Amberlite IRA-400 [171,172]. In the first cycles of sorption Amberlite IRA-400 is much better but Amberlite IRA-68 shows greater resistance to impurities. However, its application requires maintaining suitably low values of washings pH (about 5.5). Of many solutions used for silver desorption, 30% (NH4)2S203 solution proved to be the best for regeneration. To recover silver(I) from the post-regeneration solution, the electrolysis was used yielding 98% recovery of silver(I) of 99% purity. The solution remainig after electrolysis was used for another regeneration. The redox ion exchangers were also used for selective removal of silver(I) ions from the photographic sewages. At first the ion exchangers of the polythiomethylenestyrene type were used which enabled achieving exceptionally high capacities (about 22 mmol/g which constitutes about 240% bonded silver in relation to the ion exchanger mass) as for the ion-exchange method. Regeneration was carried out using the solutions Na2SO3 or NaHSO~ [173]. Much better results were obtained using hydroquinoformaldehyde resins of the capacity about 38 mmol/g that is about 400% bonded silver [174]. In this case the ion exchanger was regenerated electrolytically which made it possible to reactivate about 85% initial bed capacity. During the electrolysis silver deposited as pure metal on the cathode after passing into the solution. The sorption effect was 90-98% initial content of silver. Many interesting studies were made using the Russian ion exchanger redox EO-7 which is a polycondensat of sulphonated hydroquinone and phenol with formaldehyde [175]. By its use separation of silver(I) from copper(II), lead(II), zinc(II), bismuth(III), nickel(II), cobalt(II) and aluminium(III) was made [176]. In this case the redox exchange proved to be more selective than the ion-exchange method. Moreover, it enables removal of other components, even trace amounts of silver from the polyionic solutions of high concentrations [177]. Hydroquinone-resorcin-formaldehyde resins were effectively applied in the reactions of various metal ions reduction, among others, of silver(I) ions both in the acidic and neutral media [178]. Sulphonated 2-vinylanthraquinone, styrene and DVB copolymers were used, among others, for selective removal of silver from nitrate solutions [179]. 2.3. Tin a n d c o b a l t Another element used in plating is tin. The sewages containing this element can be purified using a two-column system. One of them is packed with strongly acidic cation exchanger and another one with weakly basic anion exchanger. Tin is retained in the second column and washed with the NaOH solution [5]. Cobalt ions can be recovered from sewages using strongly acidic cation exchangers [5]. Owing to the application of 2 M HNO3 solution for regeneration of the ion exchanger bed in the effluent the cobalt ion concentration reaches about

510 25 g/dm 3. To remove cobalt(II) ions from sewages produced during petrochemical wastes combustion chelating ion exchangers were applied [5]. The chelating ion exchangers of the functional iminodiacetic groups in the sodium form adsorb cobalt(II) ions selectively. The ion-exchange capacity of these ion exchangers for Co 2+ ions is 57 g/kg. In addition, the ion exchange is applied for removal of vanadium from the sewages produced during preparation of the zirconium-vanadium pigment [5]. As well as for recovery of molibdenium ions from sewages using a strongly basic anion exchanger [5].

2.4. Mercury Mercury and its compounds are among the chemical impurities imposing the great t h r e a t for the natural environment due to their ability of translocation in the environment and accumulation in living organisms. Mercury compounds attack the nervous system in organisms. Therefore they are considered to be the most dangerous substances. Its harmful effect consists in inhibiting the protein synthesis in cells. This process is caused by the presence of thiol groups in proteins with great affinity for mercury. Mercury, both metallic and in the form of compounds is widely applied in industry. It is used, among others, in chemical (electrolysis of NaC1 by the mercury method, production of acetic aldehyde, production of some pesticides and plastics), cellulose-paper, pharmaceutical and electronic industries. These branches of industry produce most sewages containing mercury compounds. In the fifties in the area close to M i n a m a t a Bay ( Japanese Sea), serious poisoning with mercury compounds-mainly with methyl mercury took place. Three and half thousand people were poisoned, of whom over 100 died. It was found out that the tragedy was caused by the nearby plastic producing factory which used mercury compounds as catalysts. Post-production sewages with a relatively high content of mercury were dumped into the bay. It has been estimated that only in the period 1951-1970 the factory dumped from 200 to 600 tons of mercury into M i n a m a t a Bay [180]. In the sewages mercury occurs in the forms of metallic, methyl mercury or most frequently of soluble compounds, among others, as undissociated molecules, Hg 2§ and Hg2 e§ ions as well as complex ions. All mercury compounds contained in waters make its self-purification process difficult due to inhibition of biochemical processes. To remove mercury compounds from sewages there are used, among others, reduction, precipitation, extraction and ion-exchange methods [5,10,85,181-196]. Of the above mentioned methods, the ion-exchange is of significant importance because they are technologically simple and enable efficient removal of even trace impurities from solutions. They are particularly useful when it is necessary to treat large volumes of diluted solutions. It is possible to remove mercury(II) from sewages on various types of ion exchangers e.g. strongly acidic cation exchangers, weakly and strongly basic anion exchangers as well as on selective ion exchangers of various types [5,10,85,181-196]. Many West European countries

511 apply the industrial method of mercury ions removal from sewages based on the licence of the Dutch firm Akzo Zout Chemie using the selective ion exchanger Imac TMR of functional thiol groups. This ion exchanger is used for selective sorption of mercury(II) ions from technological solutions, mainly from brines in case of their electrolysis using the mercury method and also from sewages. Imac TMR is a styrenedivinylbenzene copolymer of a macroporous structure. It possesses mainly functional thiol groups and a n u m b e r of sulfone groups [5,182,183,187,191,192]. Owing to the presence o f - S H groups, strong affinity of this cation exchanger for mercury(II) ions can be accounted for by the ability of mercury(II) cations reaction with mercaptans, thiophenols or hydrogen sulfide. Its capacity for mercury (II) is 240 g Hg/dm 3 ion exchanger. Comparing effectiveness of various methods used for mercury removal it was stated t h a t using reduction methods it is possible to decrease the mercury content to 1-3 ppm in sewages, by precipitation of HgS with hydrogen chloride to 1 ppm and applying the ion exchanger Imac TMR to 0.5-5.0 ppb. [5,10,85,182,183,187, 191,192]. Despite of the fact t h a t in the concentrated brines, mercury(II) occurs mainly in the form of the complex ions HgC142-, Imac TMR reacts mainly with Hg 2§ and HgC1 § ions which are in the equilibrium in the solution [5,182,183,191,192]: Hg 2§ + 4Cl-r

HgC1 § + 3C1-r

HgC12 + 2C1- az HgCla- + Cl-r

HgC142- (11)

However, Imac TMR does not react with metallic mercury which can occur in the dispersed form in brines. In this case, metallic mercury should be oxidized with chlorine and then the solution should be deprived of the oxidizer excess before introducing on the column containing Imac TMR as the thiol ion exchanger is readily oxidized according to the reaction: 2 R - S H + oxidizer --> R - S - S - R + 2H § R - S - S - R + oxidizer-~ x R - SOH + yRSO2H + 2 R - SOaH

(12)

(~3)

Where R is the styrenedivinylbenzene copolymer thus losing its precious properties of selective ion exchange. The former reaction can proceed reversibly due to resin reactivation by means of reducing agents but the l a t t e r one gives the ion exchangers of different functional groups which are not selective for mercury (II) ions. Mercury(II) ions sorption on the ion exchanger Imac TMR can be described by m e a n s of the following reactions: 2 R - S H + Hg 2+ -~ R - S - H g - S - R + 2H + R - S H + HgC1 + -~ R - S - H g C 1 + H+

(14)

(15)

512 After the mercury(II) ions sorption, the ion exchanger Imac TMR can be regenerated by means of the concentrated hydrochloric acid solution according to the following reactions: R-S-Hg-S-R + 2HC1 -> 2 R - S H + HgC12 (16) R-S-HgC1 + HC1 -~ R - S H + HgC12 (17) Besides the discussed ion exchanger Imac TMR, the following ion exchangers: Duolite ES-465 (Dia-Prosim, France), Chelite S (Serva, Germany) and Duolite GT-73 (Rohm and Haas, France) possess thiol groups. Of them at present only Duolite GT-73 (Imac GT-73) is produced on a commercial scale [197]. The first industrial system using Imac TMR for mercury ion sorption of 15 m~/h effectiveness was introduced in Delfizji, Holand in 1973. At present the system of 2-60 m3/h effectiveness are in the use in many countries e.g. in Cuba. Akzo Zout Chemie is generally considered to be a leading firm as far as removal of mercury(II) ions from sewages is concerned. Another group of ion exchangers exhibiting great selectivity for mercury(II) are those of isothiourea functional groups [5,10,85] like Ionac SR-3, Imac SRXL (Ionac, USA), Lewatit TP-214 (Bayer, Germany), Monivex (Ayalon, Israel), Purolite S-920 (Purolite, UK), Srafion NMRR (Ayalon, Israel), Sumichelate Q-10 (Sumitomo, Japan), Relite MAC-3 (Residion, USA). Srafion NMRR is most commonly applied for mercury(II) removal from sewages. Its sorption capacity for mercury(II) is 545 g Hg/kg ion-exchanger. 5% thiourea solution containing 0.22% HC1 is used for desorption of mercury (II) from its bed. The ion exchangers of the thiocarbamate functional groups are characterized by strong affinity for heavy metal ions. They allow for the reduction of the ion level from 20-30 ppm to the amount below 1 ppb. Of this type ion exchangers, Nisso ALM-125 ( Society of synthetic Organic Chemistry, Japan) is most commonly applied for selective mercury (II) removal from sewages. Its sorption capacity for mercury (II) is 680 g Hg/kg ion exchanger in the aqueous solution but 340 g Hg/kg ion exchanger in 10% H2SO4 solution. In the continuous process it decreases the mercury (II) content in sewages to 0.1 ppb. The sodium sulphide solution is used for mercury (II) desorption from its bed. In case of the ion exchanger Nisso ALM-126 it was stated that its sorption capacity for mercury(II) ions can be increased by the increase of operating temperature to maximum 50~ However, the bed regeneration is no longer possible. Mercury can be recovered as vapour by the ion exchanger calcination. The ion exchanger with mercury(II) deposited on it can be also mixed with cement in order for its immobilization and storage in such a form. Moreover, it was stated that the macroreticular polystyrene-based resins with the functional aminothiazole, iminothiazole or thiazoline groups exhibit a high selectivity for mercury(II). A thiazoline resin column has been used to concentrate mercury(II) from the sea water adjusted to pH=l with hydrochloric acid. Maximum sorption capacity for mercury(II) was found to be 2.8 M/kg. The

513 sorbed mercury(II) is recovered quantitatively by elution with 5% thiourea containing 0.1 M HC1 [36]. The results of studies on removal of mercury(II) ions from sewages on the anion exchangers modified with complexing organic sulpho-derivative reagents being prototypes of new selective ion exchangers are of significant importance. Immobilization of this type compounds in the ion exchanger phase is possible owing to the affinity of these reagents for anion exchangers. This results from both their aromatic structure and the presence of sulfonate groups [143,144]. For example, strongly basic anion exchangers in the azothiopyrine sulfonate [151], dithizone sulfonate [154] forms and Chromotrope 2R [148], can be effectively applied for removal of mercury(II) ions trace amounts from waters and sewages. Differentiation in mercury(II) ions retention compared with other metal ions on the obtained selective ion exchangers due to their modification is a result of: - affinity of mercury(II) ions for atoms in a ligand, - conditions of proper complexes formation, - stability of the complexes with mercury(II) ions in the ion exchanger phase. These ions exchangers owing to their high selectivity for mercury(II) ions trace amounts can be applied in mercury monitoring in the natural environment. 2.5.

Lead

Lead as a serious pollutant appears from the following sources [3,5]: - waste sludges from petroleum refineries, - waste sludges from the manufacture of alkyl lead compounds, waste solvent-based paint sludges and paint residues, - waste sludges from the manufacture of lead acid batteries, - solvent and water washes from the painting ink production. High toxicity of lead requires that its contents in sewages should be reduced to a minimum (ppb level). For this aim chelating cation exchangers with functional phosphonic or aminophosphonic groups are used [3,5,59,60]. They are selective for lead(II) ions. For example the values of lead(II) distribution coefficients for the aminophosphonic ion exchanger K-AMF depending on the value of pH of the outer solution are as follows: Kd=122 (for pH=2.5) and Kd=1900 (for pH=5) [59,60]. However, the results concerning the kinetics of lead(II) sorption are not very interesting. Weakly basic anion exchangers in the free base or sulphate form can be also applied for selective removal of lead(II) chloride complexes from the solutions of 4-6 pH [3,5]. A cation exchange process combined with precipitation is frequently used for lead removal from sewages [200]. The studies by Dudzinska and Pawlowski [201-204] concerning simultaneous removal from the aqueous solutions of lead(II) and cadmium(II) as well as organic ligands i.e. aminopolycarboxylic acids mainly EDTA in one process on anion exchangers of various types are of significant importance. The anion exchangers Amberlite applied by them differed in basicity of functional groups-

514 strongly basic types 1 and 2 as well as weakly basic ones, matrix porosity: microand macroporous as well as matrix structure: polystyrenedivinylbenzene and polyacrylic copolymers. It was stated that the anion exchangers of functional weakly basic groups exhibit stronger affinity for lead(II) and cadmium(II) complexes with EDTA than strongly basic anion exchangers and that the anion exchangers of a polyacrylic matrix exhibit stronger affinity for the above mentioned complexes t h a n those of the same type with a polystyrene matrix. The authors determined working capacities of selected anion exchangers for lead(II) and cadmium(II) complexes in the presence of various anions. Relatively high values of ion exchange capacity for the above mentioned complexes, their great affinity for the anion exchanger, as well as effective and economical regeneration by means of 1 M NaC1 solution make it possible to use this method in technology of waters and sewages. 2.6. C o p p e r Copper ions due to their toxicity disturb the operation of biological t r e a t m e n t plants and exert a negative effect on self-purification of water and on some organisms living in water reservoirs. Therefore these copper ions should be removed from wastewater by precipitation, electrolysis, ion-exchange and extraction. A very important example of ion exchange application is recovery of copper(II) from sewages formed during leaching which is one of copper production stages. Due to small pH (below 2) of sewages, conventional chelating ion exchangers of functional iminodiacetate and aminophosphonic groups do not practically adsorb copper(II) ions. Special chelating ion exchangers characterized by much greater affinity for copper(II) than for other metal ions were synthesized [5,10,205-208]. These ion exchangers are obtained from N-(hydroxyalkyl)picoliamines. One of them i.e. Dowex XFS-4195 can remove copper(II) even from 1.5 M H2SO4 solution but 5 M sulfuric acid is required for regeneration. Using another chelating ion exchanger Dowex XFS-4196, copper(II) ions are removed from less acidic solutions (pH over 1.5). Dowex XFS-4196 can be easily regenerated by means of sulfuric acid of the concentration 100 g H2SOjdm 3. The post-regeneration liquid contains 33 g Cu2+/dm 3 and 40 g H2SO4/dm 3 and can be applied for copper production using the electrolytic method. Moreover, there was obtained the chelating ion exchanger Dowex XFS-43084 of physicochemical properties similar to those of Dowex XFS-4196 but of stronger affinity for copper ions t h a n for iron ions [5,10,205-208]. Liquid chelating ion exchangers i.e. hydroxyoximes are applied for removal of copper(II) from mine waters, copper ores coming from the mines in P a n g u a n a in New Guinea, Bougainville Copper Ltd [209]. In Germany Copper(II) was also recovered by means of extraction with hydroxyoximes on the commercial scale from the mine waters as well as from the pyrometallurgical processes [210]. Such a recovery was economically justified and financial costs comparable to those of cementation and sorption processes on solid ion exchangers.

515 Copper recovery from the aqueous solutions obtained after adsorption of evaporated off gases originating from chloride calcination of pyrite ashes was proposed [211]. After solution neutralization to p H = l by means of calcium hydroxide and filtering off the sediment, copper(II) was removed using the extraction with hydroxyoximes. Gonczarowa et al. [212] used the ion exchange amphoteric fibres to remove copper(II) ions from acidic sewages. Their structure was based on partially hydrolized polyacrylnitrile and polyethyleneimine. The greatest ion exchange capacity was obtained when the solution pH was increased to 4.5. Ion exchangers of various types were applied for selective removal of copper ions from the ammonium sewages [213-215]. There was studied copper(II) sorption from the ammonium sewages on the following ion exchangers: phenolsulphone, polystyrenesulphone (microporous and macroporous), carboxylic based on acrylic, methacrylic and phenolcarboxylic acids, polyphenol type (Duolite S-30), thiole, chelating ion exchangers of functional iminodiacetate and amidoxime groups as well as on weakly basic anion exchangers of various types. Of the examined ion exchangers, polimerization carboxylic ion exchangers proved to be the most effective for selective removal of copper(II) ([Cu(NH3)4] 2§ from the sewages containing ammonium. It was stated that the maximum concentration of copper(II) 50-70 g Cu(II)/dm 3 in the effluent was obtained during regeneration of these ion exchangers with 2M sulfuric acid solution. Some sewages contain a few grams of ammonium sulphate in a litre and a small number of copper(II) ions. These sewages can be used as fertilizers but copper(II) ions should be removed earlier. The ion exchange method makes it possible to remove copper(II) sulphate solution which can be reused [216]. Carboxylic ion exchangers are most frequntly used for this aim. 2.7. N i c k e l a n d v a n a d i u m Utilization of sewages produced during sulphur removal is another significant example of ion exchange. The contain, among others, nickel and vanadium which can be recovered. Sewages are oxidized, their pH is established from 5 to 10 and then they are passed through the chelating ion exchanger bed on which nickel ions are adsorbed. The solution pH is smaller than 5 after this operation. Then the solution is passed through the chelating ion exchanger bed to adsorb vanadium ions [5,10,216]. Regeneration of beds leads to obtaining concentrated solutions of nickel and vanadium ions which are a source of useful side products. The most serious environmental problem concerning nickel(II) occurs in the metal t r e a t m e n t industry. Nickel is frequently used in the sulphate form. Thus for its removal from sewages there are applied polystyrenesulphone cation exchangers and for their regeneration sulfuric acid. It was stated that under the optimal conditions of cation exchanger bed regeneration with the sulfuric acid solution, a nickel concentration in the effluent can be several scores grams/dm 3. However, ammonium carbonate is used for regeneration of ion exchangers

516 applied in nickel recovery from sewages originating from the nickel refinery [5,10,216]. Polymetallic sewages require chelating ion exchangers e.g. nickel(II) can be selectively removed from ammonium molybdate on the aminophosphonic ion exchanger (Russian ion exchanger ANKF-80). It was stated that its sorption capacity for nickel(If) ions is 19 times as large as that of the conventional cation exchanger. Moreover, chelating ion exchangers are applied in the process of nickel ions selective recovery from the sewages originating from nickel plating and nickel compounds production [5,10,216]. Carboxylic ion exchangers also exhibit great selectivity for nickel(II) ions. Halle et al. [217] applied the macroporous carboxylic cation exchanger Wofatit CA-20 in the sodium form for selective nickel(II) ions removal from washings formed during the nickel plating. For the separation of nickel(II) from metallurgical waste solutions the highest separation efficiency was obtained with an inorganic ion exchanger with sodium titanate, Na4Ti9020 [218,219], which is a layered material, the exchangeable sodium ions being located in the titanium oxide layers. The exchanger takes up nickel(II) at pH values over 5 very efficiently. Column experiments with three types of nickel-containing waste solutions from a metal plating plant showed good performance for sodium titanate in the purification of real waste solution. Metal loadings obtained varied in the range of 0.65 meq/g to 2.7 meq/g and nickel(II) level in the effluent prior to the breakthrough was very low 0.0030.009%. 2.8. C h r o m i u m Removal of chromium(II, VI) from waste waters is absolutely necessary due to toxicity of its compounds [5,10]. Chromium(VI) compounds are believed to be particularly toxic, though all chromium compounds are belived to be carcinogenic. A m a i n source of w a t e r contamination with chromium are sewages from the surface metal t r e a t m e n t plants and from tannery. Sewages coming from baths used for chromium plating, metallic coating passivation and washing already coated surfaces are characterized by low pH caused by the presence of free acids, brown-yellow colour and small transparence. Chromium(VI) concentration in such sewages ranges 5 to 50 mg Cr(VI)/dm 3 (during the periodical change of plating baths it reaches even 200 mg Cr(VI)/dm3). Purification of such a solution consists in passing it through the strongly acidic polystyrenesulphonic cation exchanger and then through strongly basic anion exchanger. Metal ions and other cationic impurities are removed on the cation exchanger and chromates are removed on the anion exchanger: nRSO3H + M + ~-> nRM n+ + nH+ 2ROH + CrO42. ~ R2Cr04 "~- 2 0 H R2CrO4 + Cr042 + H + <--> R2Cr207 + OH-

(18)

(~9) (20)

517 The NaOH solutions are applied for anion exchanger regeneration. To remove chromium(III) from the used up tannery bath (composed of Cr(III)8000 mg/dm 3, Mg(II)-1700 mg/dm ~, Ca(II)-950 mg/dm 3, Na(I)-37000 mg/dm 3, SO42-40000 mg/dm 3, C1-15000 mg/dm 3, HCOO-12000 mg/dm 3 and pH=3.5) there can be applied the polystyrenesulphone cation exchanger Amberlite IR-120. The ion exchange capacity of Amberlite IR-120 does not depend on pH in the range from 1 to 11 and is equal to 35 mg/g. The used up tannery bath contains large number of various ions. Besides chromium(III) ions there are a large quantities of sodium cations. Studies of adsorption selectivity on the cation exchanger Amberlite IR-120 in the H § form showed that the above mentioned cations compete with one another in the process of binding with the ion exchanger functional groups, whereby the ion-exchange capacity of the cation exchanger decreases rapidly with the increase of Na(I) ion concentration in the solution. When the concentration of Na(I) in the solution is about 25 g/dm ~ then the ion exchange capacity of Amberlite IR-120H § decreases to about 5 mg Cr(III)/g, i.e. it decreases by 85% compared with the initial value. The problems of separation selectivity of Na(I) and Cr(III) were overcome using the differences in oxidation reduction properties of both cations. It was shown that the presence of the strong oxidizer which are Cr2072- ions does not cause the loss of ion-exchange capacity by the cation exchanger. The method proposed for removal of chromium(III) from the used up t a n n e r y bath consists of four stages. In the first stage chromium(III) is oxidixed to chromium(VI) by means of sodium persulphate. The obtained solution is passed through the column with Amberlite IR-120 in order to remove sodium, calcium and magnesium ions. Then Cr(VI) is reduced to Cr(III) in the effluent by means of methanol. The last i.e. the fourth stage consists in adsorption of chromium(III) ions on the cation exchanger Amberlite IR-120 which is regenerated with 2 M H2SO4 solution resulting in practically total desorption of chromium(III) [220]. The process of chromium(III) recovery from sewages known as IERECHROM [221] worked out by Petruzzeli et al. is of significant importance. Macroporous carboxylic cation exchangers (Purolite C-106) which, besides chromium ions, retain also trace amounts of other metal ions, among others, of aluminum and iron are used in this process. The regeneration process of the carboxylic ion exchanger with deposited chromium ions proceeds in stages. In the first stage the alkaline H202 solution (pH=12) is used for desorption. As a result the anion forms (chromates and aluminates) being created are quantitatively washed away from the cation exchanger and separated. In the second stage which is purification of the bed in order to wash away iron ions. As a result, iron and aluminum sulphates are formed which are returned into the process as flocculating factors. The remaining chromate solution is then used in the plating industry or in the same t a n n e r y process after reduction of chromium(VI) ions to Cr(III). The ion exchanger of the polyvinylebenzene matrix was synthesized. 1-(4pyridinyl)-2-(1-piperidinyl)ethyl ester of 4-aminobenzoic acid was bonded to this

518 matrix. The ion exchanger is characterized by high selectivity for chromium(VI) ions [222]. The best exchanger for all chemical forms of chromium was found to be a fibrous exchanger FIBAN AK-22 [223]. This selective exchanger contains both carboxylic and imidazole functionalities on polypropylene fibres. This means that the selective exchanger acts both as a cation and an anion-exchanger as well as chelating ion exchanger. At pH=3 Fiban AK-22 removed more than 99.3% of both Cr 3§ and Cr2072 and in the pH range of 5-8 more than 99.6% of CrO42 from 5 ppm Cr solutions. Fiban AK-22 columns took up efficiently chromium from real waste effluents from a metal plating plant. From a lmM solution with pH=6.1 the level of chromium in the effluent was only 0.006% prior to breakthrough. The loading was, however, rather modest 0.4 mmol/g. Liquid anion exchangers of the tertiary amine type can be applied for selective removal of chromium(VI) from the acidic sewages [5,10]. The ion exchangers of this type are produced, among others, by the firms Henkel Co, USA (Alamina 304, Alamina 336), Hoechs, Germany (Hostarex A-324, Hostarex A-327) and others. The mechanisms of Cr2072- ions removal by these ion exchangers is as follows: 2R3N + Cr2072 + 2H § --~ (R3NH)2Cr207

(21)

The liquid ion exchanger and its combination with Cr2072- ions are soluble in kerosene and other organic solvents but in water only in trace amounts. It follows from the above that Cr2072 ions can be extracted from large volumes of the aqueous solution to a small amount of the organic phase by means of such an anion exchanger dissolved in kerosene or in another solvent. Removal of Cr(VI) ions from the organic phase is not difficult. The mechanism of liquid anion exchanger regeneration (removal of Cr2072- ions from the organic phase) is as follows: (RaNH)2Cr207 + 4NaOH -> R3N + Na2CrO4 + 3H20

(22)

Chromium(VI) in the alkaline medium passes back into the aqueous phase. In this way there is obtained a concentrate of the post-regeneration solution Na2CrO4 of the content about 20 g Cr(VI)/dm 3 and the organic solution of liquid anion exchanger is suitable for a direct reuse. Extraction in the liquid-liquid system by means of the liquid cation exchangers like D2EHPA (di(2-ethylhexyl)-phosphoric acid) and Cyanex 272 (di(2,4,4-trimethylpentyl)phosphinic acid) was used for chromium(III) removal from the aqueous solutions of the composition corresponding to the used up tannery bath. The preliminary studies showed that the solutions D2EHPA or Cyanex 272 in kerosene do not extract chromium(III) from the aqueous phase of pH<3. The maximum amount which can be removed from the aqueous phase of pH=5 by means of extraction with 25% D2EHPA solution in kerosens is 35%

519 chromium(III). At higher pH of the aqueous phase there is observed chromium hydroxide removal which has an unfavourable effect on kinetics and extraction yield. These difficulties are omitted using the mixture of D2EHPA and its ammonium salt (1+1) as an extractant and undecanol as a modifier. It was shown that at the phase ratio (O/A=I), extracion with 15% D2EHPA solution in kerosene containing 10% decanol allows for removal of over 95% chromium(III) from the aqueous phase of pH=5 after two minutes. It was stated that under similar conditions, in the presence of p-nonylophenol as a modifier, extraction with the solution Cyanex 272 and its ammonium salt (1+1) makes it possible to remove chromium(III) with about 86% yield. D2EHPA was also used to remove chromium(III) from sewages (composition: 1.4 mg Cr(II), 1.4 mg Fe(III), 1380 mg Mn(II), 820 mg Co(II), 1.4 mg Ni(II), 410 mg Na(I), 127440 mg CH3CO0- of pH=3.1) formed during production of terephtalic acid. The method proposed for chromium(III) recovery from sewages of this type and checked in the semi-technical scale was extraction in the emulsion systems. In extraction there was used the emulsion of W/O type composed of 2 M H2SO4 solution (W) and organic phase ( O ) - m e m b r a n e , containing 90% kerosene, 3% Span 80 (sorbitan monooleate), 5.5% D2EHPA and 0.5-2.9% TBP. The phase ratio W/O in the prepared emulsion was 1/2. The emulsion was dissolved in the solution under studies maintaining an unchanged voluminal ratio. It was found that equilibrium is reached in the studied system after 20-30 minutes and the yield of chromium(III) extraction depends on TBP concentration in the membrane. Maximal yields of chromium(III) extraction from the studied solutions were obtained with the content of 0.5% TBP. The yield of chromium extraction from sewages reached the value over 70% [220,224]. 2.9. Zinc Production of artificial fibres using the viscose method requires during their formation acidic baths whose main components are zinc sulphate, sulfuric acid and sodium sulphate. The sewages containing zinc(II) are the most toxic of all the sewages produced by the viscose fibre industry. The damages caused by their disposal into rivers exceed many times the value of zinc(II) contained in them [5,10,225]. Winnicki et al. [225] carried out the studies of zinc(II) sorption on weakly acidic carboxylic cation exchangers. These ion exchangers exhibit great affinity for hydrogen ions, that is why they cannot be used for zinc(II) sorption in the acidic medium and in the hydrogen form. For such reasons washings must be neutralized before passing through the cation exchanger bed. Regeneration of zinc(II) from the carboxylic cation exchanger proceeds readily which allows for application of acidic baths as a regenerator. However, washing the ion exchanger with water and its transformation from the hydrogen form into a sodium one are necessary. The attemps to apply the carboxylic cation exchanger of the methacryldivinylbenzene KB-4 (Russian) of the total ion-exchange capacity was

520 as high as 6.4 mval/g independently of the zinc and sodium ions concentration ratio [225]. Zinc(II) sorption on polystyrenesulphonic cation exchangers proceeds relatively readily. However, the problem consists in the fact that regeneration should be carried out in such a way that the solutions after the regeneration could be applied in the most economical way [5,225]. The studies of Winnicki et al. [225] are of significant importance. Regenerating zinc from the cation exchanger Amberlite IR-120 with 10% sulfuric acid solution he obtained 72.6 g ZnSO4/dm 3 solution. While using a strongly acidic cation exchanger to remove metal ions from the solutions containing many alkaline metal ions, the interrelation of the amount of the ions adsorbed by the ion exchanger can make recirculation of the post-regeneration liquid impossible. That happens during zinc recovery from the sewages originating from the artificial fibre production (zinc concentraction is 540 mg/dm 3 and that of sodium is 10300 mg/dm 3. To use recylculation the sodium sulphate content should be reduced below 20% of the sum of all cations which can be reached by means of the two-step regeneration: desorption with 7% H2SO4 solution to remove sodium for washing away zinc ions from the bed [225]. A similar effect can be achieved by a preliminary action on the bed with the solution containing Zn(II) ions which remove sodium ions from the cation exchanger and thus improve Zn2+/Na + relation in the bed. Of commercially available ion exchangers selective for zinc(II), only chelating ion exchangers of the functional aminophosphonic groups are produced at present. They are suitable for selective zinc(II) removal from the solutions of pH greater than 4. In the decontamination of waste solutions Duolite C-467 column worked very well. From an alkaline metal plating waste solution having 153 ppm of zinc(II) as a cyanide complex a loading of 1.78 meq/g was obtained and the removal efficiency prior to breakthrough was 99.5% [226, 227]. The Swedish Industrial Plant Svenska Rayon Co. in Valberg along with the Department of Nuclear Chemistry, Chalmers Technical College, Goteborg worked out and applied practically the technology of zinc(II) removal with the yield over 90% from the weakly acidic sewage formed during the artificial silk production using zinc extraction by means of the liquid cation exchanger D2EHPA (so-called Valberg process) [228]. This process makes it possible to recover zinc(II) from the diluted solution about 0.2 g Zn(II)/dm 3 and to improve the environmental protection by reducing heavy metal contents in sewages. Reextraction is carried out with HeSO4 solution and the obtained reextractant is directly returned into the industrial process (into the spinning bath to regulate the speed of cellulose coagulation) or is concentrated to obtain crystalline zinc sulphate. In the process conditions when pH of the water vapour is>2, extraction is very effective and quick remaining below 2 ppm Zn(II) in the raffinate. Iron is the most serious competitor for zinc(II) in the extraction process and can be washed away from the extractant using NaOH. The Valberg process became depreciated within 4-5 years [228].

521 Lately dialkylthiophosphonic acids are of great interest (Cyanex 301 and Cyanex 302) of the following formulae:

R\fs

R\

/ \ R

s

. . R\

/ \ SH

Cyanex 301 pKa=2.61

R

/P OH

R

\SH

Cyanex 302 pKa=5.63

where R is 2,4,4-trimethylpentyl. Liquid cation exchangers like Cyanex 301 and Cyanex 302 will probably replace D2EHPA used for the extractive removal of zinc(II) ions from the weakly acidic sewages formed during the artificial silk production [229]. Zinc salts are also present in the sewages originating from the ceramic industry. A strongly acidic cation exchanger in the sodium form can be applied for their removal [3,5,10]. Regeneration is carried out by means of the NaC1 solution of the ion exchanger concentration 300 g/dm ~. The ion-exchange capacity of the cation exchanger for zinc(II) ions is 32-40 g Zn(II)/dm~. Zinc(II) removal from the solutions used in cooling towers is also troublsome. Then cation exchangers can be also used [3,5,10]. Scott et. al. Suggest application of phosphonic ion exchangers (Duolite ES-63 or Duolite TSAP-40) for selective adsorption of zinc(II) ions from cooling waters. Strongly basic anion exchangers in the chloride form are used for selective adsorption of zinc(II) chloride complexes [ZnCI4]2-. Pure ZnC12 can be obtained washing the used up anion exchanger with water [3,5,10]. The outline of the process makes it possible to isolate zinc(II) from the postregeneration liquid containing the metal ions mixture. A mixture of chlorides is obtained as a result of regeneration of strongly acidic cation exchanger (used to remove metal ions from mixed plating baths). From the obtained postregeneration liquid [ZnCI4] 2 is selectively removed on the anion exchanger. Anion exchangers are used to recover zinc chloride from digesting solutions used in the steel industry.

3. NEW TRENDS IN ION-EXCHANGE METHOD DEVELOPMENT Though removal of metal ions from the given sewage is not difficult from the economical point of view, the applicable methods are those which make the process economical. The same refers to ion-exchange as a method for the sewage components reuse. In this case effective separation of sewages into concentrated salts and water is often too expensive, particularly when the amounts of sewages are large. Thus, besides new technological systems, the efficiency of unit operations, as well as synthesis of new or improvement of physicochemical properties of already known ion exchangers are important.

522 Keeping this in mind, the scientist from CSIRO [5] worked out magnetic ion exchangers which have many advantages. Their commercial names are Sirotherm (e.g. Sirotherm IR-20 contains weak acid and weak base groups) and Siromag (e.g. Siromag 17-strong acid resin, sulphonic acid type and Siromag 57strong base resin, quaternary ammonium type 1) and they are produced by ICI Operations Pty Ltd., Australia. Application of small ion exchanger grain size (about 200 pm.) caused significant improvement of adsorption rate. As the process rate increases, the same effects can be achieved using a smaller number of ion exchangers in smaller systems. Magnetized ion exchangers can be readily separated from the waters purified from metal ions and translocated which creates perfect conditions for their applications in the flow contact apparatus. In this type of apparatus a continuous transport of the ion exchanger was possible and which, in most existing processes, has a periodical character. Very simple devices from the fluidal counter-current bed to the parallel current tube reactors can be applied to work with magnetic ion exchangers. Studies for removal of various elements particularly of uranium(VI) (about 3 pg/dm 3) from the sea water on the ion exchangers of various types are of special importance [34-45]. In case of the amidoxime ion exchanger [36,45] kinetics of uranium sorption and desorption was studied. Of various type ion exchangers tested for uranium(VI) removal from the sea water, the chelating ion exchanger of functional amidoxime groups Duolite ES-346 (3.6 mg U3Os/g) possesses the greatest working capacity for [U02(C03)3] 4-. The working capacity constitutes about 3% total capacity for uranium(VI) and is several times greater than that of hydrated titanium dioxide used in the English and Japanese experimental systems [36,45]. Contrary to hydrated titanium dioxide Duolite ES-346 is characterized by high mechanical resistance as proved by over six month tests of work in the system fluidal bed using the sea water. According to the preliminary economical analysis, the price of i kg U3Os recovered from sea water by sorption is about 5-10 times higher due to great amounts of uranium in the sea water (about 1000 times greater than the ground resources) its recovery on Duolite E S 346 or on another selective sorbent should be expected [36,45]. Barnes et al. applied the synthesized amidoxime and polydithiocarbamide ion exchangers together with the ISP or ASA method to determine trace amounts of several elements in the sea water, fresh water, biological materials, urea, serum, bones, peritoneum dialysis solutions, D-glucose, high purity graphite as well as NaC1, Li2CO~, H3BO~ [47,80-83]. The chelating amidoxime ion exchanger proved to be very effective for removal of metal ions trace amounts [47-48]. Of new selective ion exchangers Diphonix resin is very important. It is a polyfunctional ion exchanger of the structure:

523

0 HO~ lip/OH ~CH2"'-cH~CH2"cH~CH~::~ 2.~

SO3H

H~CH2~CH~

SO3H

It is characterized by high selectivity for some metal ions including lead(II). The properties of Diphonix resin are a result of introducing the additional functional groups into the matrix of the resin. Examples of modified Diphonix resins are the already mentioned Diphonix-A resin containing the same geminally substituted diphosphonic acid groups bonded to a styrenie-based polymer matrix as the regular Diphonix resin, plus strong base anion exchange groups such as the tetraalkylammonium (Diphonix-A, type 1 resin) or the quaternized pyridinium (Diphonix-A, type 2 resin) groups [230]. In our opinion the future development of these ion exchangers will make it possible to decrease ion-exchange costs as a method for recovery and reuse of the sewage components.

REFERENCES

1. 2. 3.

J.A. Tallmadge, Ind. Eng. Chem. Process Des. Dev., 6 (1987) 419. H.W. Prengle, Hydrocarbon Process, 54 (1975) 173. Selective Removal of Heavy Metals from Wastewaters by Ion Exchange and Adsorbent Resins, INF. 75006A, Dia-Prosim, Vitry-sur-Seine, 1975. 4. M.S. Raman, Chemtech, April (1981) 252. 5. B.A. Bolto and L. Pawlowski, Wastewater Treatment by Ion-Exchange, E. and F.N. Spoon Ltd, London, 1987. 6. W. Fries and D. Chew, Chemtech, 23 (1993) 32. 7. A.J. Chaudhary, J. Chem. Technol. Biotechnol., 60 (1994) 353. 8. Recent Developments in Ion Exchange 1, P.A. Williams and H.J. Hudson (eds.), Elsevier Applied Science, London-New York, 1987. 9. Recent Developments in Ion Exchange 2, P.A. Williams and H.J. Hudson (eds.), Elsevier Applied Science, London-New York, 1990. 10. Ion-Exchagers, K. Dorfner (ed.), Walter de Gruyter, Berlin-New York, 1991.

524 11. A. Dyer, M.J. Hudson and P.A. Williams, Ion Exchange Processes: Advances and Applications, Royal Society of Chemistry, Cambridge, 1993. 12. C.E. Harland, Ion Exchange: Theory and Practice, Royal Society of Chemistry, Cambridge, 1994. 13. P. Simon, Ion Exchange Training Manual, Chapman and Hall, LondonWeinheim-New York-Tokyo-Melbourne-Madras, 1996. 14. A. Clearfield, Ind. Eng. Chem. Res., 34 (1995) 2865. 15. B. Strong and R.P. Henry, Hydrometallurgy, 1 (1976) 311. 16. D.S. Flett, Chem. and Ind., 6 (1977) 641. 17. H.W. Kauczor and A. Meyer, Hydrometallurgy, 3 (1978) 65. 18. F. Vernon, Sep. Sci. Technol., 13 (1978) 587. 19. A. Warshawsky, Extraction with Solvent-Impregnated Resins, in: Ion Exchange and Solvent Extraction, vol.8, J.A. Marinsky and Y. Marcus (eds.), Marcel Dekker Inc., New York and Basel, 1981. 20. H. Hommel and A.P. Legrand, Reactive Polymers, 1 (1983) 267. 21. J.L. Cortina, Solvent Extr. Ion Exch., 12 (1994) 349. 22. J.L. Cortina, Solvent Extr. Ion Exch., 12 (1994) 371. 23. R.S. Juang and S.H. Lee, J. Chem. Technol. Biotechnol., 66 (1996) 153. 24. R.S. Juang and M.L. Chen, Sep. Sci. Technol., 32 (1997) 1017. 25. A. Warshawsky, Solvent Extr. Ion Exch., 15 (1997) 259. 26. B.N. Kolarz, The Properties and the Structure of the System of IPN Polymethacrylate Acid-co-divinylbenzene-(polystyrene-co-divinylbenzene), Prac. Nauk. Inst. Technol. Org. I Tworzyw Sztucznych PWr., 6, Wroclaw, 1980. 27. F. Helfferich, Ion Exchange, McGraw Hill Book Company, New York, 1962. 28. Duolite Ion-Exchange Manual, Dia Prosim, Vitry-S/Seine, 1969. 29. J. Minczewski, J. Chwastowska and R. Dybczynski, Separation and Preconcentration Methods in Inorganic Trace Analysis, John Wiley & Sons, New York-Chichester-Brisbane-Toronto, 1982. 30. M. Marhol, Ion Exchangers in Analytical Chemistry, Academia, Prague, 1982. 31. G.S. Libinson, Fizikochemiceskije svojstwa karboksilnych kationitov, Nauka, Moskwa, 1969. 32. A. Shogseid, Dissertation, Norwegian Technol. High School, Trondheim, 1946. 33. G.V. Miasoedova and S.B. Savin, Chelatoobrazujuscije sorbenty, Nauka, Moskwa, 1984. 34. F. Vernon and T.W. Kyffin, Anal. Chim. Acta, 94 (1977) 319. 35. L. Astheimer, H.J. Schenk, E. G. Witte, K. Witte and K. Schwachau, Sep. Sci. Technol., 18 (1983) 307. 36. K. Schwachau, Extraction of Metals from Sea Water, in: Inorganic Chemistry Akademie-Verlag, Berlin, 1984. 37. H. Omichi, A. Katakai, T. Sugo and J. Okamoto, Sep. Sci. Technol., 20 (1985) 163.

525 38. J. Okamoto, T. Sugo, A. Katakai and H. Omichi, J. Appl. Polym. Sci., 30 (1985) 2967. 39. H. Omichi, A. Katakai, T. Sugo and J. Okamoto, Sep. Sci. Technol., 21 (1986) 299. 40. H. Egawa, T. Nonaka, S. Abe and M. Nakayama, J. Appl. Polym. Sci., 45 (1992) 837. 41. H. Egawa and T. Nonaka, Ind. Eng. Chem. Res., 32 (1993) 540. 42. H. Egawa and T. Nonaka, Ind. Eng. Chem. Res., 32 (1993) 709. 43. N. Kabay and H. Egawa, Sep. Sci. Technol., 29 (1994) 135. 44. N. Kabay, Sep. Sci. Technol., 29 (1994) 375. 45. R. Khamizov, D.N. Muraviev and A. Warshawsky, Recovery of Valuable Mineral Components from Seawater by Ion-Exchange and Sorption Methods, in: Ion Exchange and Solvent Extraction, vol.12, J.A.Marinsky and Y.Marcus (eds.), Marcel Dekker Inc., New York-Basel-Hong Kong, 1995. 46. Duolite ES-346, Dia-Prosim, Vitry, 1978. 47. C.S. Siggia and R. Barnes, Anal. Chem., 52 (1980) 967. 48. R. Dybczynski, Z. Hubicki and K. Kulisa, Solvent Extr. Ion Exch., 6 (1988) 699. 49. P.R. Devi, T. Gangaiah and G. Naidu, Anal. Chim. Acta, 249 (1991) 533. 50. A.G. Latha, B.K. George, K.G. Kannan and K.N. Ninan, J. Appl. Polym. Sci., 43 (1991) 1159. 51. H. Egawa, T. Nonaka and K. Tsukamoto, Polymer Journal, 23 (1991) 1037. 52. A. Boudakgi, J. Jezierska and B. Kolarz, Makromolekulare ChemieMakromolecular Symposia, 59 (1992) 343. 53. Z. Hubicki, H. Hubicka and B. Lodyga, Adsorption Sci. Technol., 14 (1996) 5. 54. Z. Hubicki, Hydrometallurgy, 23 (1990) 319. 55. H. Eccles, The Removal of Thorium Isotopes from Nitric Acid Liquors, in: Ion Exchange Technology, D. Naden and M. Street (eds.), Ellis Horwood Limited, London, 1984. 56. G. Gonzalez-Luque and M. Street, The Recovery of By-Product Uranium from Wet Process Phosphoric Acid Solutions Using Selective Ion Exchange Resins, in: Ion Exchange Technology, D. Naden and M. Street (eds.), Ellis Hoorwod Limited, London, 1984. 57. Z. Hubicki, Chemia Stosowana, 33 (1989) 205. 58. M. Matsuda and Y. Akiyoshi, Nippon Kagaku Kaishi, 4 (1991) 336. 59. W. Szczepaniak and J. Siepak, Chem. Anal., 18 (1973) 1019. 60. W. Szczepaniak, Chem. Anal., 19 (1974) 869. 61. H. Leinonen, J. Lehto and A. Makela, React. Polym., 23 (1994) 221. 62. Z. Hubicki, Rudy Metale, 30 (1985) 54. 63. Z. Hubicki, Hung. J. Ind. Chem., 17 (1989) 51. 64. B.N. Kolarz, J. Jezierska and D. Bartkowiak, React. Polym., 23 (1994) 53.

526 65.

66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85.

86.

87. 88. 89. 90. 91. 92. 93. 94.

B.N. Kolarz, D. Jermakowicz-Bartkowiak, W. Apostoluk and A. Trochimczuk, The Sorption of [Au(CN)2]- on the Chelating Ion Exchangers, in: II Kongres Technologii Chemicznej, Wroclaw, September, 1997. J. Ciba, Z. Gregorowicz and B. Janoszka, Zesz. Nauk. Polit. SI., Chemia, 63 (1973) 55. E. Jenckel and H. Lillin, Kolloid Z., 146 (1956) 159. V. Sykora and F. Dubski, Coll. Czech. Chem. Commun., 27 (1962) 350; 28 (1963) 2139. J. Chwastowska, Chem. Anal., 32 (1987) 643. K. Kondo, H. Sumi and M. Matsumoto, Sep. Sci. Technol., 31 (1996) 1771. A. Lewandowski and W. Szczepaniak, Chemia Stosowana, 7 (1963) 603. W. Szczepaniak, Chem. Anal., 8 (1963) 843. W. Szczepaniak, Chem. Anal., 9 (1964) 481. W. Szczepaniak, Chem. Anal., 9 (1964) 761. W. Szczepaniak, Chem. Anal., 9 (1964) 1115. W. Szczepaniak, Chem. Anal., 10 (1965) 1199. E. Blasius and G. Kynast, J. Radianal. Chem., 22 (1969) 55. W. Szczepaniak and G. Kubera, Chem. Anal., 15 (1970) 1009. J . F . Dingman, Anal. Chem., 46 (1974) 774. R.M. Barnes and J. Spero Genna, Anal. Chem., 51 (1979) 1065. H.S. Mahanti and R.M. Barnes, Anal. Chim. Acta, 151 (1983) 409. P. Fodor and R.M. Barnes, Spectrochim. Acta, 388 (1983) 229. R.M. Barnes, P. Fodor, K. Inagaki and M. Fodor, Spectrochim. Acta, 388 (1983) 244. J.H. Hodgkin, Chelate-Forming Polymers, in: Encyclopedia of Polymer Science and Engineering, vol.3, John Wiley & Sons, London, 1985. A. Warshawsky, Chelating Ion Exchangers, in: Ion Exchange and Sorption Processes in Hydrometallurgy, M. Streat and D. Naden (eds.), John Wiley & Sons, Chichester-New York-Brisbane-Toronto-Singapore, 1987. M. Grote and A. Kettrup, The Separation of Noble Metals by Ion Exchangers Containing s-Bonded Dithizone and Dehydrodithizone as Functional Groups, in: Ion Exchange Technology, D. Nadne and M. Streat (eds.), Ellis Horwood Limited, London, 1984. M. Grote and A. Kettrup, Anal. Chim. Acta, 172 (1985) 233. M. Grote and A. Kettrup, Anal. Chim. Acta, 175 (1985) 239. J. Seidl, I. Stamberg and E. Hrbkova, J. Appl. Chem., 12 (1962) 500. M. Grote, P. Wigge and A. Kettrup, Z. Anal. Chem., 310 (1982) 369. G. Schmuckler, Talanta, 12 (1965) 281. K.M. Saldadze and W.D. Kopylowa-Valova, Kompleksoobrazujuscie ionity, Chimia, Moskva, 1980. K.N. Sabharwal, P.R. Vasudeva Rao and M. Srinivasan, Solvent Extr. Ion Exch., 13 (1995) 561. Z. Hubicki, Solvent Extr. Ion Exch., 6 (1988) 183.

527 95. R. Bogoczek and J. Surowiec, J.Appl. Polym. Sci., 26 (1981) 4161. 96. A.A. Efendiev, D.D. Orujev and I.I. Sidorchuk, React. Polym., 1 (1983) 295. 97. Z. Hubicki, Hung. J. Ind. Chem., 16 (1988) 217. 98. Z. Hubicki, H. Hubicka and S. Jusiak, Materials Science, 3 (1977) 53. 99. A. Sozanski, Industrial Methods for Separation of Thorium from Rare Earth Elements, Pr. Nauk. Inst. Chem. Nieorg. PWr., Wroclaw, 1981. 100. E. Bayer, Angew. Chem., 76 (1964) 76. 101. F. Vernon and T. Shah, React. Polym., 1 (1983) 301. 102. M.J. Chen, J.D. Fang and C.Y. Liu, J. Chin. Chem. Soc., 39 (1992) 121. 103. C.Y. Liu, M.J. Chen, N.M. Lee, H.C. Hwang, S.T. Jou and J.C. Hsu, Polyhedron, 11 (1992) 551. 104. B.R. Green, React. Polym., 8 (1988) 221. 105. B.R. Green, Unconventional Weak-Base Anion Exchange Resins for the Extraction of Metals, Especially Gold, in: Ion Exchange for Industry, M. Street (ed.), Ellis Horwood Chichester, 1988. 106. H.H. Kotze, B.R. Green and P. Ellis, React. Polym., 14 (1991) 129. 107. R. Hering, Chelatbildende Ionenaustauscher, Akademie, Berlin, 1967. 108. Z. Hubicki, Przem. Chem., 64 (1985) 435. 109. Z. Hubicki, Przem. Chem., 64 (1985) 611. 110. Z. Hubicki, Hydrometallurgy, 16 (1986) 36. 111. Z. Hubicki, Przem. Chem., 65 (1986) 479. 112. H. Hubicka and Z. Hubicki, Hung. J. Ind. Chem., 20 (1992) 249. 113. H. Hubicka and Z. Hubicki, Hung. J. Ind. Chem., 20 (1992) 255. 114. H. Hubicka and Z. Hubicki, Hung. J. Ind. Chem., 22 (1994) 171. 115. Z. Hubicki, Hung. J. Ind. Chem., 16 (1988) 217. 116. C.R.M. Rao, Anal. Chim. Acta, 318 (1995) 113. 117. Z. Gong and Z. Zhang, Anal. Chim. Acta, 325 (1996) 201. 118. T. Soldi, M. Pesavento and G. Alberti, Anal. Chim. Acta, 323 (1996) 27. 119. T. Ferri and P. Sangiorgio, Anal. Chim. Acta, 321 (1996) 185. 120. A. Lewandowski and W. Szczepaniak, Chem. Anal., 10 (1965) 961. 121. E. Blasius and G. Kynast, Z. Anal. Chem., 203 (1964) 321. 122. A. Lewandowski and W. Szczepaniak, Z. Anal. Chem., 202 (1964) 321. 123. F. Vernon and H. Eccles, Anal. Chim. Acta, 72 (1974) 331. 124. A. Warshawsky, Modern Research in Ion Exchange, in: Ion Exchange: Science and Technology, A.E. Rodriguez (ed.), Martinus Nijhoff Publishers, Dordrecht-Boston-Lancaster, 1986. 125. K. Terada, K. Matsumoto and Y. Taniguchi, Anal. Chim. Acta, 147 (1983) 411. 126. J. Chwastowska, W. Zmijewska, E. Sterlinska and W. Golebiewska, Chem. Anal. (Warsaw), 40 (1995) 879. 127. K. Terada, K. Matsumoto and T. Inaba, Anal. Chim. Acta, 158 (1984) 207. 128. J. Chwastowska, E. Sterlinska, W. Zmijewska and J. Dudek, Chem. Anal. (Warsaw), 41 (1996) 45. 129. K. Matusi, K. Matsumoto and K. Terada, Anal. Claim. Acta, 193 (1987) 381.

528 130. H. Fukunda, J. Tsunoda, K. Matsumoto and K. Terada, Bunseki Kagaku, 36 (1987) 683., C.A., 10S (1988) 26739w. 131. K. Terada and K. Matsumoto, Anal. Sci, 5 (1989) 319. 132. A. Sugii, N. Ogawa and Y. Higawara, Talanta, 31 (1984) 1079. 133. H. Egawa, T. Nonaka and H. Maeda, Sep. Sci. Technol., 20 (1985) 633. 134. H. Yoshida, T. Kataoka and S. Fujikawa, Chem. Eng. Sci., 41 (1986) 2517. 135. H.J. Hudson and Z. Matejka, Sep. Sci. Technol., 24 (1989) 1417. 136. K. Inoue, K. Yoshizuka and Y. Baba, Solv. Extr. and Ion Exch., 8 (1990) 309. 137. S.D. Alexandratos and P.T. Kaiser, Solv. Extr. and Ion Exch., 10 (1992) 539. 138. H. Eccles and H. Greenwood, Solvent Extr. and Ion Exch., 10 (1992) 713. 139. F. Mijangos and M. Dias, Ind. Eng. Chem. Res., 31 (1992) 2524. 140. A.K. Sengupta and Y. Zhu, AIChE.J., 38 (1992) 153. 141. R. Chiarizia and E.P. Horwitz, Solvent Extr. and Ion Exch., 12 (1994) 847. 142. E.P. Horwitz, R. Chiarizia and S.D. Alexandratos, Solvent Extr. and Ion Exch., 12 (1994) 831. 143. K. Brajter, J. Chromatogr., 102 (1974) 385. 144. K. Brajter, The Aplication of an Organic Complexing Reagents to the Reparation of Ionic Metal on an Anion Exchangers, Post Ph.D. Dissertation, UW, Warszawa, 1974. 145. Z. Hubicki, Hung. J. Ind. Chem., 17 (1989) 51. 146. S. Przeszlakowski, Chem. Anal., 37 (1992) 285. 14 7. M. Nakayama, K. Itoh, H. Chikuma, H. Sakurai and H. Tanaka, Talanta, 31 (1984) 269. 148. O. Hernandez, V. Castro and J.J. Arias, Anal. Scie., 7 (1991) 342. 149. A. Flieger and S. Przeszlakowski, Chem. Anal.,32 (1987) 85. 150. K. Brajter and E. Olbrych-Sleszynska, Talanta, 30 (1983) 355. 151. M. Nakayama, M. Chikuma, H. Tanaka and T. Tanaka, Talanta, 29 (1982) 503. 152. D.G. Berge and J.E. Going, Anal. Chim. Acta, 123 (1981) 19. 153. M. Nakayama, M. Chikuma, H. Tanaka and T. Tanaka, Talanta, 30 (1983) 455. 154. M. Chikuma, M. Nakayama, J. Kunimasa, Y. Suzuki, Y. Kurisaka and H. Tanaka, Reac. Polym., 3 (1985) 163. 155. A. Flieger, Chem. Anal., 33 (1988) 411. 156. A.K. Singh and S.K. Dhingra, Analyst, 117 (1992) 889. 15 7. A. Mehmet and A.M. Te Riele, The Recovery of Gold from Cyanide Liquors in a Countercurrent Contactor Using Ion Exchange Resins, in: Ion Exchange Technology, D. Naden and M. Streat (eds.), Ellis Horwood Limited, Chichester, 1984. 158. I.N. Plaksin and S.A. Tetark, Gidromietallurgia s primienienijem jonitov, Metallurgia, Moskwa, 1964. 159. H.H. Law, W.L. Wilson and N.E. Gabriel, Ind. Eng. Chem. Process. Des. Dev., 24 (1985) 236.

529 160. N. Masanori, U. Yoshitoshi, K. Noboru, H. Yukio and T. Tomoko, Kenkyu Hokoku-Ksmamoto-ken Kogyo Shikenjo 1979, 8, C.A., 94 (1979) 6282. 161. E. Yabe, Treatment of Plating Waste Solutions, Jap. Patent No. 7 975 159 (1977). 162. L. Chuen-Ying and S. Peng-Yong, Anal. Chim. Acta, 132 (1981) 187. 163. F. Vernon and W.M. Zin, Anal. Chim. Acta, 123 (1981) 309. 164. V. Sykora and F. Dubsky, Collect. Czech. Chem. Commun., 37 (1972) 33, 1504. 165. M. Grote and A. Kettrup, Z. Anal. Chem., 295 (1979) 366; 300 (1980) 280; 310 (1982) 369; 316 (1983) 247. 166. M. Grote, U. Huppe and A. Kettrup, Talanta, 31 (1984) 755. 167. R.I. Edwards, A.K. Haines and W.A. Te Riele, The Separation of Gold from Acidic Leach Liquors with Amberlite XAD-7, in: The Theory and Practice of Ion Exchange, Society of Chemical Industry, London, 1976. 168. L. Elci, S. Isildar and M. Dogan, Anal. Chim. Acta, 293 (1994) 319. 169. K. Kunin, Amber-hi-lites, 102 (1967) 1. 170. P.A. Riveros and C.C. Cooper, Solvent Extr. and Ion Exch., 3 (1985) 357. 171. H.W. Chou, J. Appl. Photogr. Eng., 6 (1980) 14. 172. R. Mina, J. Appl. Photogr. Eng., 6 (1980) 120. 173. A.B. Davankov, J.A. Aptova and Z.M. Giterman, Z. Prikl. Chim., 34 (1961) 1852. 174. A.B. Davankov and J.A. Aptova, Z. Prikl. Chim., 35 (1962) 2171. 175. B.A. Muchitdinova and E.E. Ergozin, Tr. Inst. Chim. Nauk Akad. Nauk Kaz. SSR, 37 (1973) 3. 176. N.F. Podorvanova, Z. Anal. Chim., 26 (1971) 818. 177. N.F. Podorvanova and N.A. Kovarski, Z. Prikl. Chim., 45 (1972) 993. 178. A.V. Gordajevski and A.E. Renard, Z. Prikl. Chim., 36 (1963) 264. 179. G. Manecke, G. Kossmehl, G. Hartwich and R. Gawlik, Angew. Makromol. Chem., 2 (1968) 86. 180. H.Z. Wrembel, Postepy Fizyki, 42 (1991) 38. 181. W.C. Gardiner and F. Munoz, Chem. Eng., 23 (1971) 57. 182. G.J. De Jong and G.J.N. Rekers, J. Chromatogr., 102 (1974) 443. 183. The Akzo Imac TMR process for the removal of mercury from waste water, Akzo Zout Chemie, Hengelo, 1975. 184. A. Warshawsky and S. Erlich-Rogozinski, Michrom. J., 22 (1977) 362. 185. R. Bogoczek and E. Kociolek, Przem. Chem., 58 (1979) 273. 186. M. Ocenasek, Z. Matejka and V. Kriz, Sb. Vys. Sk. Chem.-Technol. Praze. Technol. Paliw, 41 (1980) 43, C.A. 95 (1981) 120459. 187. J.H. Vis, Chem. Age India, 31 (1980) 481. 188. H. Heanni, B. Righetti and E. Salami, Removing Mercury from Waste Waters, Brazil Patent No. BR 8 201 651 (1983). 189. R. Bogoczek and E. Kociolek-Balawejder, Polimery, 30 (1985) 439. 190. R. Bogoczek and E. Kociolek-Balawejder, Chemia Stosowana, 30 (1986) 591. 191. R. Bogoczek and E. Kociolek-Balawejder, Przem. Chem., 65 (1986) 368.

530 192. R. Bogoczek, E. Kociolek-Balawejder and A. Kogut, Przem. Chem., 68 (1989) 83. 193. F.H. Wang and T. Jung, Huan Ching Pao Hu, 12 (1989) 118; C.A. 113 (1990) 197063d. 194. H. Yoshida and T. Kataoka, AIChE.J., 35 (1989) 318. 195. H. Yoshida and T. Kataoka, Ind. Eng. Chem. Res., 29 (1990) 2152. 196. K.A. Larson and J.M. Wiencek, Ind. Eng. Chem. Res., 31 (1992) 2712. 197. Duolite GT-73 Complexing Resin, Rohm and Haas, Paris, 1995. 198. Anon. Chem. Eng., 82 (1975) 49. 199. A. Sugil, N. Ogawa and H. Hashizume, Talanta, 27 (1980) 623. 200. K. Lee and J. Hong, AIChE J., 41 (1995) 2653. 201. M.R. Dudzinska, Ph.D. Dissertation, M. Curie-Sklodowska University, Lublin, 1992. 202. M.R. Dudzinska and D.A. Clifford, React. Polym., 16 (1991/1992) 71. 203. M.R. Dudzinska and L. Pawlowski, Chemistry for Protection of the Environment 1991, Proc.VIII Int. Conf., L. Pawlowski and W.J. Lacy (eds.), New York, 1992. 204. M.R. Dudzinska and L. Pawlowski, Anion Exchange Removal of Heavy Metal-EDTA Complex, in: Ion Exchange Processes: Advances and Applications, A. Dyer, M.J. Hudson and P.A. Williams (eds.), Royal Society of Chemistry, Cambridge, 1993. 205. K.J. Jones and R.R. Grinstead, Chem. Ind., (1977) 637. 206. R. Grinstead, W.A. Nasutavicus and R.M. Wheaton, New Selective Ion Exchange Resins for Copper and Nickel, in: Extractive Metallurgy of Copper, J.C. Yannopoulos and J.C. Agarwal (eds.), vol.2, New York, 1976. 207. R.R. Grinstead, J. Metals, 31 (1979) 13. 208. J. Melling and D.W. West, A Comparative Study of Some Chelating Ion Exchange Resins for Applications in Hydrometallurgy, in: Ion Exchange Technology, D. Naden and M. Streat (eds.), Ellis Horwood Limited, Chichester, 1984. 209. J. Szymanowski, Hydroxyoximes and Copper Hydrometallurgy, CRC Press, Roca Baton, 1993. 210. G. Barthel, Tech. Mitt. Krup. Werkber, 35 (1977) 73. 211. G.M. Ritcey, B.H. Lucas and K.T. Price, Extraction of Copper and Zinc from Chloride Leach Liquors Resulting from Chlorination Roast-Leach of TineGrained Sulphides, Proc. Int. Solvent Extr. Conf. Liege, 3 (1980) 80. 212. N.A. Goncharova, N.A. Strukova, I.M. Smirnova, E.G. Mabarakshin and L.V. Emets, Z. Prikl. Khim, 55 (1982) 2095. 213. Z. Hubicki and S. Jusiak, Materials Science, 4 (1978) 17. 214. Z. Hubicki, Rudy Metale, 30 (1985) 338. 215. Z. Hubicki and L. Pawlowski, Environment Protection Engineering, 12 (1986) 5. 216. B. Bolto and L. Pawlowski, Przem. Chem., 64 (1985) 569.

531 217. K. Halle, K. Fischwasser and B. Fenk, Technol. Umweltschutz, 25 (1982) 120. 218. A. Clearfield and J. Lechto, J. Solid State Chem., 73 (1988) 98. 219. J. Lehto, R. Harjula, H. Leinonen, A. Paajanen, T. Laurila, K. Mononen and L. Saarinen, J. Radioanal. Nuclear Chem., 208 (1996) 435. 220. T.F. O'Dwyer and B.K. Hodnett, J. Chem. Tech. Biotechnol., 62 (1995) 30. 221. D. Petruzzelli, R. Passino and G. Tiravanti, Ind. Eng. Chem. Res., 34 (1995) 2612. 222. M. Heininger and C.E. Meloan, Solvent Extr. and Ion Exch., 10 (1992) 159. 223. V.S. Soldatov, A.A. Shunkevich and G.I. Sergeev, React. Polym., 7 (1988) 159. 224. B. Wionczyk and W. Apostoluk, Rudy Metale, 41 (1996) 339. 225. M.A. Gostomczyk, T. Winnicki, M. Manczak and A. Poranek, Pr. Nauk. Inst. Chem. Nieorg. PWr., 17 (1973) 473; T. Winnicki, M. Gostomczyk, M. Manczak and A. Poranek, Environ. Protection Eng., 1 (1975) 37. 226. H. Vejima, M. Hirai and T. Ishibashi, Prog. Wat. Tech., 9 (1977) 871. 227. H. Leinonen, J. Letho and A. Makela, React. Polym., 23 (1994) 221. 228. H. Reinhardt, Chem. Ind., (1975). 229. W.A. Rickelton and R.J. Boyle, Solvent Extr. Ion Exch., 8 (1990) 783. 230. R. Chiarizia, E.H. Horwitz, S.D. Alexandratos and M.J. Gula, Sep. Sci. Technol., 32 (1997) 1.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

533

P h o s p h o r u s r e m o v a l b y slag: e x p e r i m e n t s a n d m a t h e m a t i c a l modeling S. Vigneswaran a and H. Moon b Faculty of Engineering, University of Technology, Sydney, PO Box 123, Broadway, NSW 2007, Australia

a

b Faculty of Applied Chemical Engineering, Chonnam National University, 300 Yongbong, Kwangju 500-757, Korea

1. INTRODUCTION Phosphorus is one of the essential nutrients needed for the growth of plants and animals. When plants and animals excrete wastes or die, microorganisms mineralize the various forms of organic phosphorus in the decaying matter. Both Sewage Treatment Plant(STP) effluent and urban run-off also lead to phosphorus enrichment to the waterways. In country sides, where agriculture and animal husbandry are main industries, wastes from these activities contribute to the accumulation of P in soil and water bodies. These phosphorus compounds dissolved in surface or ground waters are responsible for the eutrophication in the closed water system, especially in lakes and highly enclosed bays where water is stagnant. At present, chemical treatment, biological processes, and land treatment are used to remove phosphorus from water and wastewater. Among them, the land treatment is an attractive solution if the substratum has significant adsorptive capacity of phosphorus [1-3]. Sydney Water, Australia, recommended the land application of sewage effluent in areas where there are serious eutrophication problems (for example in Richmond area in New South Wales, Australia). On the other hand, a constructed wetland system has been suggested as an appropriate technology for P treatment in many parts of the world [4]. In this respect, phosphorus removal by slag media (waste by-products from steel industries) may be a good solution. Several researchers have investigated the utilization of natural soils and amended sand with iron oxides for phosphorus removal [1,5]. These studies were confined mainly to investigate the removal efficiency using those materials. No study, however, have attempted to deal with the P transport in substrata to provide basic information for the design of efficient land treatment facilities.

534 A detailed study on the phosphorus removal using soil and slag media from steel industries [6,7] indicated that slag media have higher sorption capacities for P compared to other substrata, such as natural soils. In particular, the batch and column experiments conducted with natural soil and slag media showed that adsorption was dominant at low pH values less than 6 while chemical precipitation, dominant at higher pH values over 8. The slag particles contain significant amount of soluble metal ions such as magnesium and calcium. Those metal ions are responsible for chemical precipitation and complexation at high pH. The section 2 of this chapter presents the results of column experiments to show the dynamic behavior of P in the column with slag at various operating conditions. Non-equilibrium dynamic models based on surface or pore diffusion inside slag media were used for simulating the adsorption behavior of P. Comparing simulated results with experimental breakthrough curves of P assessed these models. In particular, the effect of kinetic model parameters on breakthrough behavior will be rigorously discussed. At high pH, the phosphorus retention by chemical precipitation is significant. The soluble metal ions react with phosphorus in the solution to form insoluble precipitates such as calcium phosphate. A number of early researches on P chemistry postulated the formation of various insoluble inorganic phosphates of Fe, A1, Mg, and Ca through precipitation reactions. Aluminum ions combine with phosphate ions to form aluminum phosphate. Both ferrous (Fe 2+) and ferric (Fe 3+) ions are also responsible for the precipitation of phosphorus. With Fe a§ the reaction is similar to that of aluminum ion. Fe 2§ shows more complicated reactions, which are not fully understood. In the case of Fe 2§ ferrous phosphate is usually formed at pH values about 8. The precipitation of phosphate from wastewater by calcium is common. The batch precipitation experiments conducted by Lee [7] indicated that a considerable precipitation of P was observed at pH greater t h a n 8. Further, the contribution of precipitation was found to increase with the increase of pH. According to Lee [7], the dominant removal mechanism of P at pH greater than 10 is precipitation. Jenkins et al. [8] observed species such as PO43-, HP042-, P30105-, and P2074- at high pH. Aulenbach and Meisheng [9] showed that calcium present in the wastewater enhanced precipitation of P as a crystal of hydroxy apatite: 3PO34- +5Ca 2+ + O H - --+ Ca5(PO4)3OH At high pH, when phosphorus removal is both by adsorption and precipitation, in order to obtain the exact adsorption isotherms, one has to use the real adsorption amount after subtracting the amount of precipitation from the total uptake. One can also use equilibrium data obtained with washed slag media. In

535

the washed slag media most of most soluble metal ions are removed ( or washed out ) from the adsorbent particles. Because of the presence of the reactive species in the solution, the adsorption model has to incorporate the chemical reaction to calculate the phosphorus removal in slag media. For simplicity reasons, in the present work, the precipitation reaction of phosphate with soluble cations was represented by a simple bimolecular reaction. Since the precipitation reaction occurs both in the bulk solution and in the pore solution, a pore diffusion model incorporates the chemical reaction was developed to interpret the kinetic data of P in batch and column adsorption experiments. The model details are given in Section 3.

2.

PHOSPHORUS

A D S O R P T I O N : C O M P A R I S O N OF K I N E T I C M O D E L S

2.1. K i n e t i c m o d e l s 2.1.1. D y n a m i c m o d e l b a s e d o n s u r f a c e d i f f u s i o n A surface diffusion model(SDM) with external mass transfer resistance was used because of its simplicity and adequacy in describing the adsorption of P from aqueous solutions onto slag media. For model formulation, we assume an isothermal adsorption column, packed with porous spherical particles. The flow pattern is described as an axial dispersed plug-flow model. Another assumption involved in the model is the fast intrinsic adsorption kinetics, resulting in instant equilibrium between the solid phase and liquid phase concentrations at the external surface of particle [10]. If the surface diffusion is dominant, the mass balance inside a spherical porous particle can be described by the following equation:

at

~

~r--2 + -r-~r-r

/

(1)

with the initial and boundary conditions: q(r,t=0)=0

(2)

aq =0 a r r=0

(3)

aq

: kf(C-Cs)

DsPp-~-r r=Rp

(4)

536 where D~ is the surface diffusion coefficient, pp is the particle density, and kf is the external film mass transfer coefficient. The mass balance equation and the relevant initial and boundary conditions for the liquid phase in the column are as follows:

_Da x 02C

0(uC) 0C ( 1 - ~ ; b ) 3 k f ( c C ) 0 + + - s = 0z 0t ~b Rp

a--~ +

C(z,t = 0)= 0

(5)

(6)

- - (Clz o -Clz 0+)

Dax-~-z z:0

aC =0 8z z=L

(8)

The final term surface of particles concentration, Cs, equilibrium theory

of Eq.(5) represents the mass transfer rate to the external which is proportional to the driving force, C - Cs. The surface can be evaluated from the corresponding isotherm or [11].

2.1.2. Pore diffusion model(PDM) If pore diffusion is dominant inside the particle, the intraparticle diffusion model can be represented by the following equations aCp

aq

+ pp

aCp ~pDp Or

=~

lO(

r2Dp

aCp)

=kf(C-Cs) r

(9)

(10)

=Rp

aCP I =0 a r r=0

(II)

Cp(r,t : 0 ) : 0

(12)

where Dp is the pore diffusion coefficient and % is the particle porosity.

537

2.1.3. Simplified m o d e l A simplified model can be formulated from the postulate that the uptake rate by a pellet is linearly proportional to a driving force, defined as the difference between the surface concentration and the average adsorbed-phase concentration. This is called "Linear Driving Force Approximation(LDFA)". According to this approximation, the intraparticle diffusion can be simplified as follows [12]: 0~ = 3k____L(C_Cs)= ks(q s _ ~ ) 0t Rppp

(13)

Here, ks is the solid-phase mass transfer coefficient. For spherical particles, the average adsorbed-phase concentration, ~, is defined as Rp

_ q=

3 Rp

3

]rr2qd r

(14)

0

In the case that the surface diffusion is the controlling mechanism of intraparticle mass transfer, this can be estimated from the following equation. ks

~

15Ds ~ Rp 2

(15)

This relationship has been mathematically derived based on the assumption that the adsorbed-phase concentration profile is parabolic. When adsorbent particles are very small, this simplified version has been successfully applied. The coupled parabolic second-order partial differential equations, Eqs. (1) and (5), can not be solved analytically. Thus, a numerical method must be employed. In this work, a method based on orthogonal collocation [13-15] was used. The method of orthogonal collocation was used to reduce these coupled second-order partial differential equations to a set of first-order ordinary differential equations using several interior collocation points. This technique combines the classical procedure of orthogonal collocation with the high accuracy of the finite element method. The entire column was divided into a finite number of elements, in which the space variables are discretized. The set of ordinary equations was integrated using an integrator named LSODI. LSODI employs the variable-step size, variable-order, and the predictor-corrector techniques that are suitable for stiff equations.

538

2.2. E x p e r i m e n t a l Slag media (dust and cake - waste materials obtained from a steel industry) were used as adsorbents in column experiments. The dust was t a k e n from the secondary bag house deducting system used to control fugitive emissions from vessels in the steel industry while the cake is the gas scrubbing slurry from the blast furnace. In the steel production process, they are produced independently. Slag media were air dried at 105~ for a day and the fraction less t h a n 2 m m size was obtained by sieving. Slag samples were mixed well before being used in the experiments. P solution was prepared by dissolving Na2HPO4 in distilled water. The properties of dust and cake were analyzed by s t a n d a r d procedures [16]. A pycnometer was used to determine the particle and bed densities. Permeability was d e t e r m i n e d by the constant head method. The surface area and pore size were m e a s u r e d by a nitrogen adsorption method (BET, Micromeritics ASAP2400). Total element concentrations were determined in an acidified solution by ICPAES (Inductively Coupled P l a s m a Atomic Emission Spectrometer). Field conditions were simulated in the laboratory-scale column packed with slag media. Column experiments were carried out mainly to find the concentration b r e a k t h r o u g h p a t t e r n under various conditions. A phosphorus solution was supplied to the columns under gravity from overhead containers (each with a capacity of 20 liters). Constant head was m a i n t a i n e d by providing an overflow tube above the adsorbents. Effluent flow was controlled by a flow meter. Concentration of POn-P in the tap w a t e r was less t h a n 0.05 mg/L. Forty liters of the tap w a t e r was infiltrated through the column prior to the introduction of the P solution to ensure a uniform compactness of the adsorbents. Care was t a k e n not to allow any air bubbles in the under drainage section and in the media layer. It is evident t h a t the solid phase concentration (q) and the liquid phase concentration (C) m u s t coexist in equilibrium at the b r e a k t h r o u g h point. The pH of the P solution in the influent was m a i n t a i n e d in the range of 6 to 7. Six columns of 9 cm d i a m e t e r were packed with the dust and cake adsorbents to thickness of 1 and 3 cm, respectively. Depth of the media was m a i n t a i n e d shallow to achieve an early breakthrough. The entire effluent concentration history was recorded. To e s t i m a t e the solid phase concentration, the soil and slag in the column was t a k e n out for air drying at the b r e a k t h r o u g h stage of the adsorbents. It was further dried in an oven at a t e m p e r a t u r e of 105~ for 24 hours. Then the preweighed a m o u n t of the adsorbents was extracted using nitric acid. After extraction, they were diluted with a required volume of distilled w a t e r and the solution was decanted. Then the samples were analyzed for P concentration using a spectrophotometer (Milton Roy Spectronic 20D). 2.3. E x p e r i m e n t a l r e s u l t s and m o d e l v e r i f i c a t i o n 2.3.1. C h a r a c t e r i s t i c s of the a d s o r b e n t s Physical and chemical properties of the adsorbents were m e a s u r e d since P t r a n s p o r t is influenced by the intrinsic characteristics of the media. The physical

539 characteristics of the slag media tested are listed in Table 1. The chemical properties are also shown in Table 2. More detailed properties are given elsewhere [7]. The values of the permeability of slag media are in the range of 1.8 - 3.2"10 .7 m/s. These values are very low so t h a t the slag media are not suitable in rapid infiltration systems. This result would be expected because two slag media used in this study have slit (<0.06 mm) fraction more t h a n 80 %. For rapid infiltration, larger particles with high permeability are recommended as s u b s t r a t a [17].

Table 1 Physical properties of dust and cake

Average particle d i a m e t e r Particle density Packing density Bed porosity Surface area x 103* Average pore radius * Permeability x 107

Dust

Cake

Unit

18 2,780 1,029

50 1,950 760 0.63 22.0 20.0 3.2

~m m3/kg m3/kg 0.61 m2/kg nm m/s

7.0 9.22 1.8

* from BET method based on N2 adsorption

Table 2 Chemical properties of dust and cake Constituents

Dust

Cake

Exchangeable cation I (cmol/kg) A1 Mg Ca K Na CEC (Cation Exchange Capacity)

0.22 0.32 67.96 1.88 3.00 73.40

<0.01 20.92 27.53 1.80 46.88 97.10

Total Elements 2 (%) P A1 Fe

0.07 0.19 17.70

0.05 0.21 34.40

1 Exchangeable cations in a 0.01 M BaC12 leachate, determined by ICPAES (InductivelyCoupled Plasma Atomic Emission Spectrometer). Exchangeable A1 is determined only if pH < 5.04 2 Total elements determined by acid digestion and ICPAES

540 The BET analysis based on N2 adsorption shows t h a t dust and cake have very small surface area, which is comparable to those of metal oxides [18]. Their average pore diameters are in the range of 9.2 - 22.3 nm, which can be classified into mesopores.

2.3.2. Adsorption i s o t h e r m s Equilibrium data of P on dust and cake were fitted with two well-known isotherms, L a n g m u i r and Freundlich equations as shown in Figure 1. qmbC q=~ l+bC

(16)

q : KC 1/n

(17)

10

o

Dust

a

Cake

......

7//o ~

Langmuir

/~.,..

Freundlich

.-" """

, -""

~:r 4

0

I

I

I

I

I

I

iO

20

30

40

50

60

C (mg/L) Figure 1. Isotherms of P on cake and dust at 20~

Equilibrium data were obtained from column experiments. Their isotherm p a r a m e t e r s are given in Table 3. According to the average percent deviations given in Table 3, the Freundlich equation gives better fit t h a n the L a n g m u i r equation. This fact can be justified by the high surface heterogeneity of slag media. Since slag media used in this study contains a lot of metal compounds

541 that are responsible for effective cation exchange capacity (CEC). The major contributing factors for CEC are calcium for dust (68 cmol/kg) and sodium for cake (46.9 cmol/kg). Large proportions of metals exposed on external and pore surface may enhance the adsorption of P.

Table 3 Adsorption equilibrium parameters of P on dust and cake Type Langmuir Equation

Freundlich Equation

---N--k=l

qexp

qm b E(%) k n E(%)

Dust

Cake

21.82 0.0104 2.69 0.263 1.130 0.140

10.28 0.0072 1.82 0.083 1.094 0.100

k

The adsorption capacities of dust and cake are found to be much higher than that of a sandy loam soil obtained from North Sydney (0.019 - 0.033 mg/g in the same concentration range) [7] even they were washed thoroughly prior to use. The dust has approximately 200 times of capacity for P uptake comparing with the soil (if the effect of precipitation is included). These results imply that slag media from steel industries are very suitable adsorbent for P removal from water bodies.

2.3.3. D e t e r m i n a t i o n o f k i n e t i c p a r a m e t e r s In addition to equilibrium isotherms, information on sorbate transport is essential for analyzing the dynamic behavior in the adsorption-based facilities. Lee [7] used a simple dynamic model to predict his experimental breakthrough curves of P obtained from column experiments. His model is quite similar to the simple model described in this paper, which contains two kinetic parameters, an axial dispersion coefficient and a solid-phase mass transfer coefficient. He extracted two parameters directly from one set of his own data and predicted all other breakthrough curves successfully. Since these kinetic parameters can compensate each other during their optimization process, it is not easy to get reliable parameters simultaneously from a set of data. In this work, the axial dispersion coefficient and the external film mass transfer coefficient were estimated from proper correlation reported in the literature. Therefore only the surface diffusion coefficient was obtained from a set of experimental data.

542 The Reynolds n u m b e r calculated from experimental conditions used in this study is in the order of 10 .4 - 10 -5. For low Reynolds number, one can approximate the axial dispersion coefficient from the fact that the axial Peclet n u m b e r is of the order of 0.3 - 1.0 [12]. However we use a correlation suggested by Wakao and F u n a z k r i [19] since this correlation has been used successfully for a liquid-phase system [20].

Dax = Re Sc +

2~Rp

(18)

The external film mass transfer was also estimated from Wakao-Funazkri equation [12]. Sh =

2k

fR_________=~p2 + 1.1Re ~ Sc ~ (19) Dm where Sc and Re denote Schmidt and Reynolds numbers, respectively and Dm is the molecular diffusion coefficient that was estimated by Wilke-Chang equation [21]. Under the experimental conditions used here, the estimated molecular diffusion coefficient of P (as PO3) is 1.34"10 .9 m2/s. The surface and pore diffusion coefficients of P in dust and cake were calculated from a set of column data (Run No. Dust 30 and Cake 30 in Table 4) by minimizing the object function defined as

lOO FlOex,call]

E(%) = ~

j=l

[

C exp

(20)

j

Figure 2 clearly shows a minimum of the object function in terms of the surface diffusion coefficient. When the simplified version was used, the solidphase mass transfer coefficient, ks, is required. This value was calculated from Ds using Eq.(15). All kinetic p a r a m e t e r s obtained are listed in Table 4 with column conditions. The axial dispersion coefficient estimated is about 2.70.10 -s m2/s, which is very low. This could be due to small particle size and low interstitial flow rate. Consequently, it can be expected that the effect of axial dispersion is negligible under these experimental systems. The external film mass transfer coefficient obtained is in the range of 0.54 - 1.57.10 .4 m/s. These values are much larger t h a n those of other system, namely organic species on activated carbon and resins [18]. This implies that the mass transfer resistance through the external film is small. The low value is due to relatively small particle size. The surface diffusion coefficient is very small (10 -15 m2/s) compared to that of phenols in activated carbons [18]. This result may be explained using the concept of an activation process for P transport in dust and cake. In fact, the adsorption

543 Table 4 Kinetic p a r a m e t e r s * e s t i m a t e d at 25~ R u n No.

Co [mg/L]

L [m]

u'106 [m/s]

Dax'108

[me/s]

kf'104 [m/s]

D~-1015 Dp.10 la [me/s] [me/s]

ks" 105

[l/s]

Dust Dust Dust Dust Dust Dust

05 10 20 30 40 50

5 10 20 30 40 50

0.01 0.01 0.01 0.01 0.01 0.01

35.3 35.3 35.3 35.3 4.28 4.98

2.72 2.72 2.72 2.72 2.69 2.68

1.57 1.57 1.57 1.57 1.51 1.52

0.229

4.50

4.25

Cake Cake Cake Cake Cake Cake

05 10 20 30 40 50

5 10 20 30 40 50

0.03 0.03 0.03 0.03 0.03 0.03

1.18 1.18 1.81 5.32 6.25 6.94

2.69 2.69 2.70 2.70 2.71 2.71

0.54 0.54 0.55 0.55 0.55 0.56

5.25

9.50

7.25

i000

I00

i

%,

I0 0

.

~ l l

0.I

!"

i, r

~

Dust

"r ~

~--

_,u

I

~2 'i' 'I'

i.

~

..

I ~ -

,

.... Cake :

-+

.

.

.

.

4

!

!

.

.

.

.

.

. .

. .

. .

" : : : '

!

,

:

,

i

0.01 I0

-16

I0

-15

Surface diffusion coefficient, D~, m2/s

Figure 2. Determination of Ds from experimental data.

i0

-14

544 35

25 -

~o

~176176176176176176176176176176176176176 .....................

0 15 s

"/

10

.... Dax = I0 ~

,"'/~

. . . . Dax

= 1 0 "9

5 0

I

I

I

I

I

20

40

60

80

I00

T i m e (hr)

Figure 3. Sensitivity of axial dispersion coefficient on BTC. 35 30

..'" ........................................

25

~,20 "m ~Z5 ro 1~ I

/

5 I/ 0 ,- ..... 0

/;

- - - - D s =10 -16

.'"

- ' - " Ds :10 "17 I

I

I

I

I

20

40

60

80

i00

Time (hr) Figure 4. Sensitivity of surface diffusion coefficient on BTC.

545 of P on dust and cake is likely to be chemisorption r a t h e r t h a n physical adsorption. Therefore the u p t a k e rate is usually very slow. The pore diffusion coefficient is also small (10 -13 m2/s). In batch systems, it was confirmed t h a t more t h a n 100 hours is required to reach its equilibrium state [7]. Figure 3 and 4 show the sensitivities of Dax and Ds on the b r e a k t h r o u g h curve of P in dust column (Run Dust 30). To check the sensitivity of each kinetic p a r a m e t e r , the value was varied by one order less and more t h a n t h a t of the determined one. As expected above, the effect of Dax was negligible on the b r e a k t h r o u g h curve. On the other hand, the variation of Ds affected the b r e a k t h r o u g h curve significantly. This result implies t h a t the diffusion controls the sorption rate and its d e t e r m i n a t i o n is crucial in s i m u l a t i n g the dynamic behavior of P sorption. 2.3.4. S i m u l a t i o n o f c o l u m n d y n a m i c s Finally, b r e a k t h r o u g h curves (BTC) of P from dust and cake columns were simulated using the three models described in section 2.1. Figures 5 to 8 show the experimental data and the predicted b r e a k t h r o u g h curves. The solid and broken lines r e p r e s e n t the prediction by the full dynamic models based on surface or pore diffusion mechanism, respectively, and the dotted lines represents those by

0

O0 0

0

O0 O0 0

o.'"'"~176 ......................

~176 .-i

00 0

0

0"" 0'''"

~ 3 vE o

1

0

o/~ ,~,

o

......Simple S DM

,-'I' o

7

"~^0 0

-

PDM

..................

20

40 Time

60

Figure 5. Breakthrough curves of P (Run Dust 05).

80

I00

546 35 3o

-1-

20

~

./j

oOo o

r

_

15 [

,~/ ,q/

1

Exp. Data

o

. . . . . . Simple

5 0 ,-, 0

I

I

I

I

I

20

40

60

80

i00

Time (hr) Figure 6. Breakthrough curves of P (Run Dust 30).

~176 .................................................

5 ~

4 i 0

,

0

9"

~03

0 0

9 9

0 0

9 t 9

o

o ,

o

~SDM

O

O O

1

Exp. Data

. . . . . . Simple

o

2

0

0

0

//~r

PDM

0 O0 0

to 0

0

I

20

4O Time (day)

Figure 7. Breakthrough curves of P (Run Cake 05).

I

I

60

80

547

3O ~176176176176176 ......... .o"0~ 0 25

--

(5" ~20

",~/

~15-

o

Exp. Data

...... Simple SDM

i0-

PDM 5-

0

I

-

0

5

i0

15

Time (day) Figure 8. Breakthrough curves of P (Run Cake 30).

the simplified version. It is very interesting to note that the full models predict all b r e a k t h r o u g h curves satisfactorily while the simplified model could not predict the adsorption behavior of P in the column of cake, particularly at low concentrations. At first, it was anticipated that both models predict all b r e a k t h r o u g h curves since the particle sizes of dust and cake are small and the intraparticle diffusion is extremely slow. This unexpected result may be due to the low flow rates r a t h e r t h a n low concentrations. Other cases, which are not shown here, revealed also the same tendency at low flow rates. In Figures 5 - 8 , all b r e a k t h r o u g h data are somewhat scattered, showing multiple stages which are typical for multiple species systems [22,23]. Phosphorus has a very complex chemistry [8]. Even one used an orthophosphate to prepare the solution; many different ionic species can be formed in the solution depending on the solution pH. They have different affinities eventually competing with each other for specific adsorption sites. This may be a reason why b r e a k t h r o u g h data show multiple stages. If a proper solution characterization technique is available and the fraction and isotherm p a r a m e t e r s of individual ionic species are known, one can predict these abnormal b r e a k t h r o u g h curves more accurately [23,24]. However such an approach would not be so important from the engineering points of view.

548

3.

A D S O R P T I O N OF P A C C O M P A N Y I N G WITH P R E C I P I T A T I O N REACTION

3.1. T h e o r e t i c a l d e v e l o p m e n t 3.1.1. P o r e d i f f u s i o n m o d e l for b a t c h s y s t e m s The rate of chemical reaction between P (denoted by 1) and the cation (namely Ca 2+, denoted by 2) is t a k e n into account as a simple bimolecular reaction [10]. r 1 = krCblCb2

(21)

r 2 = krVCblCb2

(22)

where kr is the reaction constant and v is the stoichiometric ratio between P and Ca e§ If the precipitation reaction forms mainly hydroxy apatite, the stoichiometric ratio is 1.667. In this work, it was assumed t h a t precipitation occurs both in the bulk solution and the pore fluid at the same time. When the solution with a known initial concentration is contacted with a given a m o u n t of adsorbent particles, the mass balance of P in the bulk solution can be written as follows V dCbl = - V k r C b l C b 2 - k f l a f ( C b l - C s l ) dt

(23)

Cbl(t =0) = C 0bl

(24)

where V is the volume of solution, kfl is the external film mass transfer, and af is the external surface area of adsorbent particles, af can be represented by the following equation 3w

af = - ppRp

(25)

Here, w is the mass of adsorbent, Rp is the particle radius, and pp is the particle density. The mass balance for the cation is also represented by a similar equation. V dCb2 = -VvkrCblCb2 - kf2af(Cb2 - C s 2 ) dt

(26)

0 Cb2(t = O) = Cb2

(27)

549 When the intraparticle mass transfer is due to only the diffusion of adsorbate molecules through the pore fluid, the macroscopic conservation equation of P inside a spherical adsorbent particle is given by OCpl t P P 0 q l . Dpl . c3. ( r 2. 0 C p l / 0t gp 0t r 2 Or ~, )0r

krCplCp2

(28)

with two following boundary conditions

c3Cpl Ir=R = kfl(Cbl -Csl ) p

epDpl ar

0Cpl I =0 Or r=0

(29)

(30)

The corresponding initial condition is given by

Cpl (t = 0) = C ~

(31)

Similarly, the conservation equation for the cation is given by

c3Cp2 + ~ ~ Pc~q2 P 0t gp 0t 0Cp2 epDp2 Or

c3Cp2 I 0r

= DP----~2 r 20r c32 ( r

r=Rp

=0

c3Cp2 l-krVCplCp20r

= k f2 (Cb2 - Cs2 )

(32)

(33)

(34)

Ir=0

Cp2(t = 0) = C~

(35)

where Dpi is the pore diffusion coefficient of the i th species and ~p is the particle porosity. 3.1.2. Dynamic model for column adsorption When the flow through the column is assumed to be axially dispersed and a chemical reaction is involved in the adsorption system, the mass balance

550 equation and the relevant initial and b o u n d a r y conditions for the i th species in the liquid phase of the column are also r e p r e s e n t e d as follows: 92 aCbl 3(1- gb)kfl c3Cbl = Dax Cb-------!-I- u -krCblCb2 (Cbl - C s l ) 0t az 2 0z ~bRp

(36)

Cbl(t=0)=0

(37)

Dax c3Cbl z = - u ( C b l Iz:0--Cbl[z=0+ ) c3z =0

(38)

~Cbl

=0

c~z

(39)

z=L

where Dax is the axial dispersion coefficient, v is the interstitial flow rate, ~b is the bed porosity, and L is the bed length. Similarly, the mass balance equation for the cation becomes c3Cb2 _ Da x c32Cb2 ~ - t ) c3t c3z 2

0Cbl 3(1 - gb)kf2 - krVCblCb2 (Cb2 - C s 2 ) 0z ~bRp

Cb2(t =0) = 0

DaxC3Cb20z z=0 =

0Cb2 0z

(40)

(41)

-"(Cb2z:0--Cb21z:0+)

=0

(42)

(43)

z=L

3.1.3. N u m e r i c a l s o l u t i o n t e c h n i q u e

To obtain dimensionless model equations, some dimensionless variables were introduced as follows

z

s = -L

r

x = -Rp

(44)

(45)

551 t)t =~ L

(46)

Cbi Ybi = CO

(47)

Cpi

(48)

Ypi = Co

where Co is a unit concentration. Then, the dimensionless model equation for column becomes c3Ybi 0x

1 02Y bi Pe c3s2

c3Ybi c3S

~iYblYb2 - ~

- Ysi)

Ybi(l: = O) = 0

Ybil

(?s [s=O

(49)

(50)

--Pe( bi s-O- - Ybits_-O+)

(51)

OYbi =0 8s s=l

(52)

Also, the dimensionless equation inside the particle is given by {1 + qf, (Ypi)}0Ypi O~ =7i

02

} Ypi + .2 .C3Ypi . . 2 X C3X Ox

~iYplYp2

(53)

C3Ypi = Ki(Ybi-Ysi) Ox x=l

(54)

CgYPi ] =0 Ox x=O

(55)

Ypi(X =0) = y 0pi

(56)

Here, dimensionless groups are defined as follows

552

Pe

uL Dax

(57)

Ppqo

(58)

- - - ~

q _ . _

EpC0 ~i =

krLC0vi

(59)

3(1- ~:b)kfL ~bRpO

(60)

DpiL Rpu

(61)

Rpkfi ~i

=

(62)

~

spDpi

The derivative of isotherm is defined as d~gi f' (Ypi): dypi

(63)

where the dimensionless solid phase concentration, ~pi,is defined as ~gi = q__L qo

(64)

Here, qo represents a unit concentration. All the parabolic second-order partial differential equations above can not be solved analytically. In this work, a numerical method based on orthogonal collocation was used as in the section 2.1. The method of orthogonal collocation was used to reduce these coupled secondorder partial differential equations to a set of first-order ordinary differential equations using several interior collocation points [13,14]. The entire column was divided into a finite number of elements, in which the space variables are discretized. According to a collocation technique, the first and second partial derivatives can be represented as follows ayb /

t, as )i

NF+2

~A~,jybj j=l

(65)

553 / NF+2 a2yb = Z Bi,JYbJ 0s2 i j=l

(66)

Here Ai,j and Bi,j are the first and second derivative coefficients. From a boundary condition of the column at s = 1 /

(•sb

NF+2 NF+I = ZANF+2,JYbJ = ZANF+2,JYbJ + ANF+2, NF+2ybNF+2 = 0 NF+2 j=l j=l

(67)

Then, the dimensionless concentration at s = 1 is obtained as NF+I YbNF+2 = - E ANF+2'J Ybj j=l ANF+2, NF+2

(68)

After substituting this relation into Eqs. (65) and (66), one obtains / NF+I . 0Yb = Z Ai,JYbJ ~S i j=l

(69)

NF+I . 02yb = Z Bi,JYbJ 0 s2 i j:l

(70)

]

Here, two modified derivative coefficients are defined as Ai,j = Ai,j -

Bi, j = Bi, j -

Ai,NF+2ANF+2,j

(71)

ANF+2,NF+2 Bi,NF+2ANF+2,j

(72)

ANF+2,NF+2

From the inlet boundary condition, / NF+I . . NF+I . OYb = Z A I , j Y b j = A I , l Y b l + Z A l , j Y b j = - P e ( C i n - Y b l ) as 1 j=l j=2 From this equation, one can evaluate the concentration at the inlet section as

(73)

554

Ybl = -

NF+I PeCin 1 ~ * . - . Al,jYbj A1,1-Pe A1,1-Pe 9

(74)

The set of first-order ordinary differential equations is obtained by substituting Eqs. (71), (72), and (74) into dimensionless partial differential equations. Equations for the radial direction are not listed here. The detailed set of equations are elsewhere [25]. This set of ordinary equations was integrated using an integrator, LSODI as in the section 2.1.

3.2. E x p e r i m e n t a l Series of batch sorption experiments were conducted to study the P sorption in the static system. The batch adsorbers were 500 mL conical flasks. In each flask, 400 mL of a known concentration of P solution was mixed with a known amount of adsorbent (0.4 g for dust and 2 g for cake, respectively). The mixture was kept in a completely static condition without stirring throughout the experiment. The experiments were carried out at various initial concentrations. The column experiments were carried out mainly to obtain the breakthrough curves of P under various conditions. A phosphorus solution was supplied to the columns under gravity from overhead containers. The constant head was maintained by providing an overflow line. 3.3. E x p e r i m e n t a l results and m o d e l v e r i f i c a t i o n 3.3.1. C o n t r i b u t i o n of p r e c i p i t a t i o n It is known that the precipitation contributes to the removal of phosphorus when slag media are used as adsorbent, particularly at high solution pH. As mentioned in Section 2, original slag particles contain a lot of soluble cations, namely magnesium and calcium ions which are responsible for chemical precipitation and complexation. However, it is not simple to separate the contribution of precipitation from the total P removal because of its complex chemistry [8]. In designing the efficient land treatment facility for P removal, it is very important to quantify the precipitation or complexation effect. In order to account for only the contribution of adsorption, one can use washed slag media, from which all soluble cations are extracted to eliminate the contribution of precipitation. In Section 2, we determined adsorption isotherms of P on dust and cake from equilibrium data obtained from the column packed with washed slag media. Using these isotherms, we can predict the concentration history curve of P in the batch experiments. Here only the adsorption contributes to P removal. Figure 9 shows experimental and predicted concentration curves when dust was used as an adsorbent for P removal (in the batch experiments). This experimental data were obtained by contacting 400 mL of the P solution of 20 mg/L initial concentration with 0.4 g of dust (Run BD-20). The dotted line represents only the contribution of adsorption, calculating from the isotherm of P determined in

555 Section 2. The solid line denotes the variation in P concentration when adsorption and precipitation occur at the same time. In this calculation, the a m o u n t of soluble cations (equivalent to Cae+), which was d e t e r m i n e d from experimental d a t a and isotherm, was used. The difference between two curves m a y be a s s u m e d to be the contribution of precipitation or complexation of P. According to the results shown in Figure 9, it can be concluded t h a t the contribution of precipitation is equally i m p o r t a n t as t h a t of adsorption.

25

20

?

15

~

i0 O

O

o 5

O

O

Exp. Data

. . . . . . Adsorption

--

Ads + Pre 0

I

I

I

I

50

i00

150

200

Time (hr) Figure 9. Contributions of adsorption and precipitation to P removal by dust. 3.3.2. E v a l u a t i o n o f t h e a m o u n t o f s o l u b l e c a t i o n s The a m o u n t s of soluble cations were e s t i m a t e d from static batch experimental data and corresponding isotherm as mentioned above. All experimental conditions for batch experiments are listed in Table 5. Figure 10 shows how the a m o u n t of soluble cations affects the concentration history curve of P for Run BD20. Since it is a s s u m e d t h a t the precipitation of P follows an irreversible bimolecular reaction, the final (or equilibrium) concentration in the batch adsorption decreases almost linearly with the a m o u n t of soluble metal ions in the batch adsorption. For Run BD-20 as shown in Figure 9, the a m o u n t of soluble ions dissolved from 1 g of dust was 15.67 mg/g (or 0.39 mol/kg). This value is much lower t h a n the cation exchange capacity (CEC) of dust, 0.734 mol/kg shown

556 in Table 2. This result implies t h a t all metal ions dissolved from dust are not responsible for precipitation. Similarly, the a m o u n t of soluble cations on cake was e s t i m a t e d by comparing the experimental data for Run BC-20 with the calculated result. In the case of cake, the a m o u n t was 3.34 mg/g (0.083 mol/kg). In Figure 1 l, the broken line represents only the adsorption contribution while the solid line accounts for the total removal a m o u n t both by adsorption and precipitation.

Table 5 E x p e r i m e n t a l conditions for batch adsorption Run No.

Type of adsorbent

V x 103 (m 3)

w x 103 (kg)

BD-10

Co (mg/L) 10

BD-20

Dust

0.40

0.4

20

BD-30

30

BC-10

10

BC-20

Cake

0.40

2.0

20

BC-30

30

24

20

16

"'\\

\\

"'-. .................................................

-....

c,.) 0 mg/g 5 mg/g . . . . . . 10 mg/g ....

15 mg/g

I

I

I

I

50

i00

150

200

Time (hr) Figure 10. Effect of initial amount of soluble cation on concentration curve.

557

24 20

16

6"'~...................................................... o

8 Data

4 +

. . . . . . Adsorption ~

0 I

0

Ads + Pre

I

I

200

400

600

Time (hr) Figure 11. Contributions of adsorption and precipitation to P removal by cake.

3.3.3. D e t e r m i n a t i o n o f r e a c t i o n r a t e c o n s t a n t Prior to d e t e r m i n i n g the reaction rate constant, kr, one needs to check the effect of kinetic p a r a m e t e r s on the corresponding concentration curve. In this work, the pore diffusion coefficient of P was e s t i m a t e d from a b r e a k t h r o u g h curve (Run Dust-30 in Section 2) by comparing with results predicted from the dynamic model based on the pore diffusion mechanism. The pore diffusion coefficient of the ion was a s s u m e d to be the same as t h a t of phosphate ion. This assumption is reasonable because two ions have similar effective sizes in aqueous solution. Figure 12 shows the sensitivity of the pore diffusion coefficients on the concentration decay curve. When one order of m a g n i t u d e was t a k e n less or more t h a n the pore diffusion coefficient of P, the variation in concentration curve is significant. This implies t h a t the effect of the pore diffusion coefficient is very critical within this range. However, the sensitivity of reaction rate constant for dust is relatively small as shown in Figure 13. In this calculation, the order of m a g n i t u d e in the reaction rate constant was also changed. According to the simulated results, the reaction rate constant for the bimolecular reaction between P and the cation is approximately of the order of 10 .2 for dust. The sensitivity of the reaction rate constant for cake is shown in Figure 14. Exact values of the

558

24-

20\ \

16-

~12-

\

8--

~Dp=4.5xl0

12

. . . . . . D p = 4 . 5 x l 0 -13 _

....

0

I 0

50

D p = 4 . 5 x l 0 -14

I

I

I

100

150

200

Time (hr)

12. Effect of pore

Figure

diffusion coefficient on concentration curve.

2ll

20

16

.

. . . . . . Kr = 3.0x 10 -1 Kr = 3.0xl 0 -2

__

Kr = 3.0x 10 .3 0

0

I

I

I

I

50

100

150

200

T i m e (hr) Figure 13. Effect o f reaction rate constant on concentration curve o f P on dust.

559

24

20

16

o

. . . . . . Kr = 6.3x10 -3 Kr = 6.3x10 -4

4 --

Kr = 6.3x 10 -5

0 0

I

I

I

200

400

600

T i m e (hr) Figure 14. E f f e c t o f r e a c t i o n rate c o n s t a n t on c o n c e n t r a t i o n cu rv e o f P on cake.

i00

g o

10

b m

i 0.0001

0.001

0.01

9

Kr

Figure 15. D e t e r m i n a t i o n o f r e a c t i o n rate constant,

kr,

on dust and cake.

i

560 reaction rate constant can be obtained by minimizing the object function t h a t is defined as the average percent difference between experimental and predicted concentration curves. In this work, we obtained 3.10 .2 and 6.3"10 .4 L/mg s for dust and cake, respectively. Figure 15 shows the average percent error in t e r m s of the reaction rate constant to find a minimum. It should be noted t h a t two o p t i m u m reaction rate constants for dust and cake are quite different since they have different types of metal ions incorporated in them.

3.3.4. S i m u l a t i o n

of concentration

curves

When both adsorption and precipitation contribute to the P removal, it is essential to check their relative effect on the total uptake. According to the calculation, the precipitation reaction is quite fast compared to the rate of adsorption. This is due to the fact t h a t most metal ions released from slag particles were consumed at early stage of P removal process. W h e n the a m o u n t of P is over the limiting value (which is defined as the stoichimetric a m o u n t for reaction), the concentration of cations was extremely low or not detectable as shown in Figure 16. For an initial P concentration of 10 mg/L, the concentration of cations in the bulk solution was not detectable throughout the experiment. However, in the case of low P concentration (7.5 and 5 mg/L), the concentration of

1 2

- -

. . . . . . 10. m g/L

10-

7.5 mg/L _

.

.

.

.

.

5.0 mg/L

_

4

c..)

-

~".

P

Cation

~ i~ ~ .~.....'""" ................ .

/

--2

0

'/"~9

o. o. o

I

I

I

I

50

i00

150

200

Time (hr) Figure 16. Concentrations of P and cation in terms of Co of P.

561 cations increased monotonously with time after consuming all P. This results show t h a t the precipitation is very fast and dominates the initial stage of P removal in the batch adsorption. This is a reason for initial steep decrease in the concentration curve. If the kinetic p a r a m e t e r s of P removal processes are evaluated without introducing the precipitation effect, the m a g n i t u d e of the p a r a m e t e r s will be u n r e a s o n a b l y high compared to the real value. Using the a m o u n t of soluble cations and the reaction r a t e constant, the concentration history curves of P in the batch adsorption were evaluated. Figure 17 shows the results for P removal with dust. Even there is a slight deviation between the experimental and predicted results, all experimental concentration curves could be predicted satisfactorily by the adsorption model which incorporate a precipitation reaction term. Similarly, the concentration curves of P with cake were also predicted and their results are shown in Figure 18. There are deviations between experimental and predicted results after a certain time of about 20 hours in Figure 18. These discrepancies may be resulted from the r e t a r d a t i o n of pore diffusion since precipitates formed in pore fluid can block pores.

35 o

30

[] 20 mg/L zx 30 mg/L Prediction

25"

~

A A

~~ 1 5 ~ i0

10 mg/L

A

A

[]

_

[] _

_

_

u

u

u

0

0

5 0 0

i 50

0

I

I

I

I00

150

200

Time (hr) Figure 17. Experimental and predicted concentration curves of P on dust.

562

35 o 10 mg/L o 20 mg/L A 30 mg/L

30 25 ~20

a

Prediction

.~ []

~

A

A

A

i0 5 ~ 5

"0

0 0

0

0

0

0

0

0

I

I

i

200

400

600

Time (hr) Figure 18. Experimental and predicted concentration curves of P on cake.

3.3.5. S i m u l a t i o n o f c o l u m n d y n a m i c s Some breakthrough curves of P from columns which were packed with dust or cake were also predicted by the proposed model in order to simulate the effect of precipitation on the breakthrough behavior of P. This is very important because the wet-land system will usually use the unwashed slag media. The simulated breakthrough curves of P with kinetic parameters obtained show that the precipitation contributes significantly the total P removal even in column systems unless the reaction rate constant is extremely low or the pore diffusion coefficient of cations are extremely high. Figure 19 shows the effect of pore diffusion coefficient on breakthrough curves of P in the column in which dust particles are packed. The solid line denotes the predicted breakthrough curve with a pore diffusion coefficient of 4.5.10 -13 m2/s. When this value was increased by one order, the breakthrough curve becomes a steep but S-shaped curve which is typical for fast adsorption. However, the predicted breakthrough curve with a low pore diffusion coefficient represents a little complicated phenomenon, a steep increase at earlier stage and a progressive increase later on. These unusual results may be due to a combined effect of diffusion and reaction kinetics. In other words, the dominating

563 35 30 o

25

-

~20

~

................

!

--

~15"%)

0-12

/

I0--

/

....

4.5x10 -13

/

5--

d/

0 0

.':

I

20

40

60

80

i00

Time (hr) Figure 19. Effect of pore diffusion coefficient, Dp, on breakthrough curve of P. 30

25

"-

-

o,"~176176

-

20

3

/

10

....

Kr=0.03 .003

20

40

60

80

i00

Time (hr) Figure 20. Effect of reaction rate constant on breakthrough curve of P.

564

mechanism can be changed during the process. One may conclude that the reaction dominates at earlier stage but the diffusion becomes dominant later. The effect of reaction rate constant on breakthrough curves of P was also studied here. In Figure 20, the breakthrough curves calculated with different reaction rate constants are shown. Comparing with the effect of pore diffusion coefficient, the reaction rate constant does not affect much. When the reaction rate constant is over a certain limiting value, the corresponding breakthrough curve does not change because of diffusion limitation. Finally, two experimental breakthrough curves of P (Run Dust-30 and Cake-30 in Section 2) were simulated using dynamic model in conjunction with precipitation term (Figures 21 and 22). The broken lines denote the breakthrough curves predicted by the model when the precipitation term is incorporated. When the effect of precipitation was taken into accounts, the breakthrough time increased significantly. This means that the slag media with soluble cations have longer lifetime for P removal in constructed wetland systems. However, this results is valid will be valid only when precipitates formed are insoluble and retained in the pore or interparticle structures.

5

-[-

30-

/

25-

~

~20-

ooo%~ ...'"" o~ % /../-""

~15-

S 10-

. ' " . . . o Exp. Data

c9~

/.-""

oJO

_

....'" '

0

I .........

20

"

~

Adsorption

...... Ads + Pre

I

I

I

40

60

80

I00

Time (hr) Figure 21. Experimental and predicted breakthrough curves of P from dust column.

565 30a 25"

-

Exp. Data

A

Adsorption

/

A

20-

15-

10-

.

0

5

i0

15

Time (day) Figure 22. Experimental and predicted breakthrough curves of P from cake column.

4.

CONCLUSIONS

Equilibrium data of P on dust and cake were found to fit well both by Langmuir and Freundlich equations. Even dust and cake were washed, the sorption capacities were observed to be more than 20 - 50 times that of a sandy roam soil. This implies that slag media from steel industries can be used as effective adsorbents to remove P from water and wastewater. The axial dispersion coefficient and the external film mass transfer coefficient appearing in the model presented in Section 2 were estimated from the experimental conditions using correlative relationships given in literature. Only the intraparticle diffusion coefficient was obtained from breakthrough data. From the sensitivity analysis, it was found that the intraparticle diffusion is a very slow process that is the rate-controlling step. In column experiments, all breakthrough data were observed to be somewhat scattered, showing multiple stages that are typical for multicomponent systems.

566 These results may be due to the fact that different phosphate ions are present in the solution according to a complex chemistry of phosphorus. They can compete with each other for specific adsorption sites. Breakthrough data of P obtained from dust and cake columns were predicted from the dynamic model based on the intraparticle diffusion. They can successfully be used in the design of adsorption-based P removal facilities packed with slag media. The simplified version, on the other hand, could not predict the breakthrough curves from the column of cake, particularly at low flow rates. The precipitation was observed to be significant at high pH values. In this study, the precipitation was represented by an irreversible bimolecular reaction and the total amount of metal ions released from slag media was represented in terms of an equivalent amount of Ca 2§ The amount of metal ions was evaluated by subtracting the predicted amount by isotherms (those were measured from column experiments packed with washed slag media)from the total uptake. The amount of soluble metal ions from 1 g of dust was 15.67 mg/g while that from cake was 3.34 mg/g. These values are much lower than the cation exchange capacities of slag media. This implies that not all metal ions dissolved from dust or cake are responsible for precipitation. The pore diffusion coefficient of the cation was assumed to be equal to that of phosphate ion and the reaction rate constants were obtained by minimizing the deviation between experimental and predicted concentration curves. The reaction rate constants obtained are 3-10 -2 and 6.3"10 -4 L/mg s for dust and cake, respectively. The reason for the difference in the reaction rate constant may be explained by the fact that two slag media have different types of soluble cations (as shown in Section 2). The model predicted all concentration history curves of P in batch adsorption satisfactorily. The model was formulated based on pore diffusion mechanism by combining the effect of precipitation as well as adsorption. According to calculated results, it may be concluded that the precipitation is very fast and signifies the initial stage of P removal processes. The model prediction shows that the dominating mechanism changed from reaction to adsorption during P removal processes. The column experiments were also simulated by the proposed model. As observed in batch systems, the effect of pore diffusion coefficient was also larger than that of reaction rate constant in column systems. When the contribution of precipitation as well as adsorption was considered, the breakthrough time increased significantly. However, these results will be valid, when precipitates formed are insoluble and remained in the column. If unwashed slag media are used as adsorbents for P removal, at the initial stage, the precipitation can contribute to P removal more than the adsorption. This model forms an excellent first step in the simulation and design of the constructed wetland systems.

567

ACKNOWLEDGEMENT This work was supported by Australian Research Council (ARC) as a joint program between ARC and Korea Science and Engineering Foundation(KOSEF). H. Moon gratefully acknowledges the support of ARC as a 1996 international research fellow at University of Technology, Sydney.

NOMENCLATURES af

Aij b Bij C Dax Dp Dm Ds E f K kf kr

ks L n N Pe q qm r

Rp Re s Sc Sh t T V w x

external surface area defined in Eq.(25) (m 2) first order derivative coefficient Langmuir constant (L/mg) second order derivative coefficient concentration in the fluid phase (mg/L) axial dispersion coefficient (m2/s) pore diffusion coefficient (m2/s) molecular diffusion coefficient (m2/s) surface diffusion coefficient (m2/s) percent error (%) derivative of isotherm Freundlich constant film mass transfer coefficient (m/s) reaction rate constant (L/mg s) solid phase mass transfer coefficient (l/s) bed length (m) Freundlich constant number of data point Peclet number defined as vdp/Dax concentration in particle phase (mg/g) monolayer coverage (mg/g) radial distance (m) particle radius (m) Reynolds number defined as 2vRp/v dimensionless axial coordinate Schmidt number defined as v/Din Sherwood number defined as 2kfRp/Dm time (s) temperature (~ volume of solution (m 3) weight of adsorbent (kg) dimensionless radial coordinate

568 dimensionless concentration axial distance (m) Greek Letters

~b

@ Y 2 I/ V

pb Pp T,

dimensionless group defined in Eqs.(60) dimensionless group defined in Eq.(59) bed porosity particle porosity dimensionless group defined in Eq.(61) dimensionless group defined in Eq.(58) dimensionless group defined in Eq.(62) kinematic viscosity or stoichiometric coefficient interstitial velocity (m/s) bed density (kg/m 3) particle density (kg/m 3) dimensionless time defined in Eq.(46) dimensionless adsorbed phase concentration defined in Eq.(64)

Super and subscripts 0 b cal exp i,j p s -

initial or inlet values bulk or bed calculated experimental species pore or particle surface or interface average

REFERENCES

1. R.J. Bruce, C.R. Martin and A.F. George, Water Environ. Res., 64 (1992) 699. 2. H. Yamada, M. Kayama, K. Saito and M. Hara, Wat. Resour. Res., 20 (1986) 547. 3. H. Yamada, M. Kayama, K. Saito and M. Hara, Wat. Resour. Res., 21 (1987) 325. 4. R.A. Mann and H.J. Bavor, Wat. Sci. & Tec., 27 (1993) 107. 5. G.E. Ho, M. Kuruvi!la and R. Gibbs, Water Resour. Res., 26 (1992) 295. 6. S.W. Lee, S. Vigneswaran and H. Moon, Separ. Purif. Technol., 12 (1997) 109. 7. S.H. Lee, Ph.D Thesis, University of Technology, Sydney (1995). 8. D. Jenkins, T.F. Ferguson and A.B. Menar, Wat. Resour. Res., 5 (1971) 369. 9. D.B. Aulenbach and N. Meisheng, J. Wat. Pollu. Control Fed., 60 (1988) 2089.

569 10. S.W. Park, H.S. Park, W.K. Lee and H. Moon, Sep. Tech., 5 (1995) 35. 11. H. Moon and C. Tien, Ind. Eng. Chem. Res., 26 (1988) 2042. 12. C. Tien, Adsorption Calculations and Modelling, Butterworth-Heinemann, Boston (1994). 13. J. Villadsen and M.L. Michelsen, Solution of Differential Equation Models by Polynomial Approximation, Prentice-Hall, Englewood Cliffs (1978). 14. E. Suwondo, L. Pibouleau, S. Domenech and J.P. Riba, Chem. Eng. Comm., 102 (1991) 161. 15. N.S. Raghavan and D.M. Rutheven, AIChE J., 29 (1972) 922. 16. A. Klute, Am. Soc. of Agronomy, Inc., Soil Sc. Soc. of Am., Inc. Madison, Washington, USA (1986). 17. G.S. Steiner and R.J. Freeman, Configuration and Substrate Design Considerations for Constructed Wetlands Wastewater Treatment, ID. Hammer (ed.), Lewis Publishers Inc., Michigan, (1989) 363. 18. D.M. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley & Sons, NewYork (1984). 19. N. Wakao and T. Funazkri, Chem. Eng. Sci., 33 (1978) 1375. 20. J.W. Lee, H.C. Park and H. Moon, Separ. Purif. Technol., 12 (1997) 1. 21. R.C. Reid, J.M. Prausnitz and B.E. Poling, 4th ed., The Properties of Gas and Liquids, MaGraw-Hill Co., New York (1994). 22. H. Moon, H.C. Park and C. Tien, Chem. Eng. Sci., 46 (1991) 22. 23. H. Kage and C. Tien, Ind. Eng. Chem. Res., 26 (1987) 284. 24. M.C. Annesini, F. Gironi and L. Marrelli, Ind. Eng. Chem. Res., 27 (1988) 1212. 25. J.W. Lee, Ph. D. Dissertation, Chonnam National University, Kwangju, Korea (1996).

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

571

Adsorption and chemisorption of organic pollutants on solid aerosol surfaces V. A. Pokrovskiy ~, V. I. Bogillo a and A.D~browski b aInstitute for Surface Chemistry, Academy of Sciences of Ukraine, Prospect Nauki 31, 252022 Kiev, Ukraine bFaculty of Chemistry, M. Curie-Sklodowska University, 20-031 Lublin, Poland 1.

INTRODUCTION

Up to date m a n y sources of environmental pollution are not well understood. This is valid, in particular, for solid aerosols of industrial origin, and for toxic organic chemicals adsorbed on their surface. T h e r m a l and photochemical reactions of adsorbed and chemisorbed species m a y result in their transformations into more toxic forms. Those can be delivered to the h u m a n organism with the respirable fraction of aerosols or through drinking water. In the W e s t e r n Europe t r a n s p o r t aerosols receive most of the attention while in Central and E a s t e r n Europe the industrial dust is dangerous atmosphere pollutant posing h a z a r d for h u m a n health. The distribution of industrial dust pollution in Europe, showing two m a x i m a at G e r m a n - P o l a n d boarder and at E a s t e r n U k r a i n e with dust concentration about 20-40 ~g/m ~ and more [1], confirms the suggestion t h a t industrial dust microparticles originate mainly from burning coal. In Ukraine, for example, energy production provides about a half, heavy i n d u s t r y about a q u a r t e r of industrial dust pollution. During the initial stage of formation in the fire-chamber, the fly ash particles are h e a t e d to several h u n d r e d s centigrade and then they are exposed to visible, ultraviolet and particle-beam irradiation while wandering in troposphere. The surface of fly ash microparticle is covered by incomplete combustion products known as precursors to synthesis of strong toxins, carcinogens and mutagens. Such a system seems to be an ideal catalyst chamber and close attention should be paid to results of uncontrolled catalytic reactions on the aerosol surfaces. Migration of solid aerosols to areas with other types of organic and inorganic pollutants can create unexpected combinations of chemicals in the atmosphere. The systematic study of the adsorption and chemical reactions of organic pollutants on the solid aerosol surfaces was not performed till now, and the present methods of solid aerosol control are limited mainly by monitoring of its inorganic components. In addition, the effect of industrial solid aerosols on the

572 h u m a n organism has not been studied in detail. Recently the renewed interest in the health risks of aerosols has been generated by the finding of a correlation between increased mortality and the concentration of fine airborne particles in the urban atmosphere of USA [2]. But detailed mechanisms by which this trend is realized are not clear. It is known that adsorption of drugs on the solid surfaces cause their prolonged action [3-5]. In the same vein, adsorption of the toxic compounds on dust microparticle surface may lead to prolonged exposure to toxins. There has been recognized that the fine fraction of the atmospheric aerosols is probably much more involved in the health effect of the general population than previously thought. A number of epidemiological studies has been published over the past 5 years which reported an association between low ambient particle concentration and acute morbidity and even mortality in the elderly. It was hypothesized that the ultrafine anthropogenic atmospheric aerosols with particle sizes less and much less than 1 pm present in the urban atmosphere at high number concentrations may be causally involved. Some recently published model animal inhalation studies support this hypothesis. The growing concentration of industrial solid aerosols in the stratosphere has raised concern about their potential impact on the ozone cycle. Although some modeling studies have considered in case of the heterogeneous chemical reactions of atmospheric reservoir species on solid propellant rocket motors-exhaust particles, little attention has been paid to the possibility of gas phase atmospheric species reacting directly with solid aerosol surface. Considering that most metal oxides contain a variety of acid/base active sites that contact with organic pollutants, such as halogenated hydrocarbons, the solid aerosols could very well promote the decomposition of the important trace species in the stratosphere. The high flux of solar ultraviolet radiation in the stratosphere may possibly provide a photochemical mechanism for generating reactive surface sites of solid aerosols. Since halomethane photolysis is the rate-limiting step in the entire halogencatalyzed ozone depletion cycle, surface-mediated processes promoting carbonhalogen bond cleavage could potentially have a profound impact on the ozone cycle. The adsorption and surface reactions of organic compounds on the solid aerosols are of great interest lately as a possible means of deposing of stockpiles of these compounds in the atmospheric processes, particularly of chlorofluorocarbons whose production has been banned by international treaty. In addition, the surface reactions may be the reason for appearance of anomalous high concentrations of some organic compounds, such as formaldehyde, which origins cannot be explained in usual manner. Heterogeneous reactions of gaseous molecules on aerosol particles have recently been recognized to significantly alter the chemistry of the atmosphere. However, the actual reaction mechanisms and their rates are still poorly quantified. Because of very complex processes in atmosphere it is impossible to distinguish between single reaction steps. Many processes run parallel, and may influence each other or depend on meteorological conditions (e. g. the relative

573 humidity), on season, on origin of aerosols and pollution level. The theoretical evaluation of single processes is only possible by comparing the individual rate and equilibrium constants of these reactions. In the proposed chapter the main regularities in the organic compounds adsorption and surface reactions on various inorganic oxides and carbonaceous materials are discussed. Similar study of adsorption and surface reactions on the model components of solid aerosols is a commonly employed approach in heterogeneous atmospheric chemistry. Main attention is directed to the relationships between the structure of important organic pollutants and main constituents of solid aerosols, on the one side, and their activity and reactivity in the adsorption processes and surface reactions, on the other. Because the reviews devoted to problems of heterogeneous atmospheric organic chemistry are absent, this paper presents one of attempts to connect various fields, such as analytical chemistry of atmosphere, theories of adsorption equilibria and surface reactions on the heterogeneous solids, quantitative "structure-activity" and "structurereactivity" relationships in these processes, in order to estimate and predict the adsorption behavior and reactivity of atmospheric organic pollutants with the surface of solid aerosols, mainly of industrial origin. The results presented here are also of interest for other fields of adsorption, in particular to applications in environmental sciences, such as development and choice of selective adsorbents for removing organic pollutants from air, adsorption of organic substances by soils, etc. 2.

SOLID AEROSOLS COMPOSITION

Particles in the atmosphere boundary layer have a direct impact on climate and weather. They influence the radiation budget of the atmosphere because of their ability to absorb and scatter solar radiation. The rise of atmospheric pollutions and related entire changes in E a r t h climate attend the great interest to composition and structure of atmospheric aerosols, as main components of atmospheric pollutions. Main sources of solid atmospheric aerosols and their efficiencies are presented in Table 1. The aerosols may be separated in small (fine, 0.1 - l~m) and coarse (1 - 100 ~m) particles and exhibit a characteristic distribution. The fine particles mainly originate from chemical processes, whereas the coarse particles are caused by wind soil erosion and sea spray. The composition of the aerosols depends on the production mechanism and the chemical environment. The aerosols may be considered as complicated mixture of natural and anthropogenic chemical substances, containing sulfates, soluble inorganic compounds, including solid minerals and carbon containing compounds. Some data about the aerosols composition are given in Tables 2 and 3.

574 Table 1 Range for a m o u n t of p r i m a r y solid aerosols in the E a r t h atmosphere (in million tons per year) received from various sources (taken from [6]) Source Soils and rocks

Troposphere 130

+ 8000

Ocean (mainly NaC1)

300

-

Volcanos(ashes)

200

+ 1000

Forest fires (carbon black) Cosmic dust

Stratosphere -

1300

4.0

3 - 360 0.25

+ 14

+ 50 -

0.25

+ 14

Industrial plants

5 + 27

-

Plowed fields

2 - 80

-

10 + 133

-

Fuel consumption plants Transport

1.0

0.01

+ 0.1

Aerosols contain salts and such insoluble particles as metal oxides. Iron, aluminum, manganese, calcium, silicium, zinc, cobalt, copper are the most a b u n d a n t metals in the atmosphere. In the case of such mixed particles as atmospheric dust and fly ashes, their composition is a dominant factor t h a t influences the overall surface transformations of inorganic or organic atmospheric gases and volatile substances. The black and organic carbons are also included as i m p o r t a n t constituents of the atmospheric aerosols. Some data on composition of the organic carbon in the aerosols originated from various sources are given in Table 3. The p r i m a r y fraction of the organic carbon is essentially emitted by combustion processes. Also, the analysis of the organic material in the aerosols shows strong evidence of secondary organics resulting from the condensation, adsorption and catalytic transformations of gaseous volatile organic compounds, produced frequently from photochemical oxidation reactions in the atmosphere. It is evident t h a t the higher molecular weight organics in atmospheric aerosols are predominantly of polar nature. The fatty acids are mainly present in the marine aerosols, whereas phenols were found in the continental samples. P r i m a r y compounds as alkanes, alkenes and polycyclic aromatic hydrocarbons constitue the major organic fraction in u r b a n aerosols. Heterocyclic compounds, alkenes, mono and polycyclic aromatic and alcohols constitue the major fraction in biomass burning aerosols. Secondary compounds as esters, dicarboxylic acids and ketones are the major fraction in the polar aerosols. The humic-like substances are also i m p o r t a n t constituents of the organic carbon present in atmospheric particulates. During the lifetime of aerosols a certain a m o u n t of metal ions can be leached from the particles by the polar

575

organic substances, and participate oxidation processes.

in homogeneous

transition

metal-catalyzed

Table 2 C h e m i c a l c o m p o s i t i o n of s o l i d a e r o s o l s (in ~ g / m 3) f r o m v a r i o u s s o u r c e s . ( A v e r a g e d c o n c e n t r a t i o n s of t h e c o m p o n e n t s a r e t a k e n f r o m d a t a p r e s e n t e d i n [6]) Component Organic carbon * Si A1 Fe Ti Zn Cd Ca Sr Mg Na K Mn Cu Cr Ni Pb Hg Mo Th Ce

Ta Br C1 As NH4 NO3 304

1

2

3

4

1.4 + 70 0.1+0.7 0.01+0.02 0.01+0.07

0.02+0.1

5

6

7

0.1 + 1.8

2+10 0.1 + 4.0 0.9 + 15 0.1 + 1.0 0.06 + 5

0.2+40 0.3 + 15 1.0 + 3.0

300+2000 100+ 400 50 +180

0.5+16

1.7+6.0

0.1 + 0.3 0.1 + 2.1 30+200

1.0 + 7.0

0.1 + 1.2

0.5 + 70 0.5 + 29 0.5 + 44 0.0003 0.2 + 3.8

0.2 +1.5 0.15 + 12 0.2 + 50

0.7+150

0.06+14

1.0 + 26 1 . 7 + 7.3 0.4 + 1.8 0.04 + 7 0.05+10 0.05+ 0.3 0.2+0.7

0.1 + 3.2

0.02 + 7

0.5 + 5.0

0.2 + 0.5 0.3 1.0

0.01+0.05

3+25 0.03+0.5 0.006+1

0.003

0.07+0.3 0.4+3.0

0.3+9.7 0.5 + 4.8 3+25 0.06+0.1 0.1+0.43 0.2+1.2 0.04 + 0.1

0.5+20 1.0+25

5.0 0.03 + 2 0.006+6 103+ 0.6 0.01+1.0

1.6 + 3.4 1.9 + 3.2 0.7+2.8 0.1 + 2.6

0+0.003 0.007 0+0.04 0+3.4 0.03 +0.1

0.02+ 0.5 0.03+5 0.01+ 0.1 0.02+ 0.5 0.1+0.2

0 . 0 7 + 3. 0.1 + 0 . 8 0.05+0.2 < 0.04 0.005+0.1 0.01 0.01 + 0.4 0.3 + 0.8 0.3+0.7 0.2

0.4 + 1.9

Definition of columns in the Table" (1) Stratosphere; (2) Urban atmosphere; (3) Near ground layer; (4) Volcano (Kamchatka); (5) Near water layer (Atlantic ocean); (6) Troposphere; (7) Industry dusts in troposphere (Leningrad region). Data are taken from [7].

576

Table 3 Organic components of solid aerosols Group of organic compounds

Near w a t e r layer

n -alkanes:

2.7 + 10 ng m -3

C26H54 + C35H72

[9,10]

Polycyclic aromatic hydrocarbons Alcohols: C 12t-12~0 H -

C34H690H

Fatty acids C,TH35COOH + C3zH65COOH Esters of aromatic and fatty acids, aromatic hydroxyderivatives

Urban atmosphere

Forest atmosphere

4 + 150 ng m -3 [11] 5.5 + 405 ng m -3 [11, 12]

0.01 + 10 ng m -3 [10, 13, 14, 15] 0.001 + 30 ng m -3 50 + 200 ng m -3 [10, 13, 14, 15] [161

*[17]

198 +1930 ng m -3 [7] 90 + 390 ng m -3 [7]

*[17,18]

*Presence of these substances in solid aerosols is conformed by data of qualitative analysis in [17,18].

In Table 4 the chemical composition of fly ashes of several electric power stations in Ukraine is presented which allows to estimate the chemical composition of oxide component of industrial dust microparticles. It is clear from the Table 4 t h a t the oxide component of fly ashes of electric power stations is silica accounting for more t h a n half, alumina about a q u a r t e r of the total mass. Fe oxides occupy the third place. As a whole, the industrial dust microparticles should be considered as alumina silicates, with admixture of some other oxides. Among them, presence of titanium dioxide known as efficient photocatalyst is to be emphasized, in a m o u n t of up to 1% in some samples. In view of possible surface reactions on the aerosols, their specific adsorption area is an i m p o r t a n t parameter. The surface area of most fly ashes is in the range of about 1 - 30 m 2 g-l, which limits the adsorption capacity toward organic pollutants. At the same time, the low adsorption capacity of the aerosols toward various atmospheric inorganic gases is explained frequently as occupation of their active sites by adsorbed large organic molecules. Accordingly, main attention in this paper will be given to the surface processes taking place on various solid parent and mixed inorganic oxides and carbonaceous materials modified by grafted organic groups, as main model constituents of solid aerosols. The adsorption, chemisorption and some photochemical transformations of organic pollutants on the surface of these solids will be considered in the forthcoming paragraphs.

577 Table 4 Oxide composition of fly ashes and slams of electric power stations in Ukraine, in wt % [8] K20 Power station SiO2 A1203 Fe203 FeO CaO MgO + SO3 Na20 Burshtynskaya 46.3 22.4 15.9 5.3 2.7 2.5 2.6 0.8 49.4 23.1 6.1 12.3 4.4 1.6 2.2 0.2 Pridneprovskaya 46.8 25.9 15.4 2.3 2.5 2.3 2.7 0.3 49.4 21.7 4.4 16.4 3.9 1.0 2.2 0.4 Dobrotvorskaya 45.7 26.7 11.5 2.0 4.1 2.8 1.2 1.1 55.5 23.1 3.0 10.4 2.3 2.0 2.3 0.5 Zaporozhskaya 43.8 23.3 18.7 2.9 4.4 2.2 1.8 1.0 53.6 21.8 12.1 9.2 2.8 1.0 3.2 Ladyzhynskaya 44.8 22.7 14.5 4.5 6.6 2.9 2.3 0.9 52.2 25.0 5.2 10.7 3.0 1.0 3.3 Starobyshyvskaya 46.8 21.5 14.8 4.8 6.3 2.5 2.0 1.2 55.4 26.8 - 11.0 1.8 0.7 3.2 Dzerzhinskaya 47.7 24.9 11.9 2.9 3.7 1.9 3.6 0.8 53.6 24.4 3.3 9.5 2.9 2.1 4.3 Mironovskaya 47.0 20.3 11.0 2.2 5.4 1.2 2.5 1.8 56.0 24.1 0.4 10.1 2.8 1.8 2.3 0.7 Voroshilovgradskaya 43.6 24.8 10.8 2.0 7.7 2.2 2.1 2.0 55.5 23.1 3.0 10.4 2.3 2.0 2.3 0.5

0

A D S O R P T I O N OF ORGANIC C O M P O U N D S F R O M GAS P H A S E ON HETEROGENEOUS AEROSOL SURFACE

All components of solid aerosols such as silica, alumina and other metal oxides, metal salts, carbonaceous parts in the presence of adsorbed inorganic anions (SO 2- , NO~, CI-, Br- etc.) and long-chain organic compounds possess surface and structural heterogeneities, which play a dominant role in adsorption and reactions of organic pollutants on their surface especially at low pressures. The main sources of the surface heterogeneity are as follows: different types of functional groups on the inorganic oxide and carbonaceous material surfaces, various types of crystal planes, edges and corners, irregularities in crystallographical structure of a surface and impurities strongly bounded with the surface [19]. The assumption of solid heterogeneity introduces considerable difficulties in the theoretical description of adsorption equilibria, chemisorption kinetics and surface reactions on solid aerosols. We present here relatively simple expressions for physical adsorption from gas phase on heterogeneous solids,

578 which allow to provide a semi-quantitative description of this phenomena on complex aerosol surfaces. Let us consider solid surface having L-types of adsorption sites. The overall adsorbed from gas phase amount, nt, is equal to L n t = ~ ni i=1

(1)

where ni is amount adsorbed on adsorption sites of the i-th type with an adsorption energy Ei. Defining O~i as a ratio of the number of adsorption sites of the i-th type (noi) to the total number of adsorption sites (no), the equation 1 can be written as follows L O(P, T)= Z O~i0i(P' T, E i ) i=1

(2)

where 0i = ni/noi, is relative coverage of adsorption sites of the i-th type and | = nt/n0 is relative monolayer coverage of the entire surface, P and T are equilibrium pressure and temperature of the adsorbate, respectively. Equation 2 defines the overall adsorption isotherm for surfaces with a discrete energy distribution. This normalized distribution may be expressed as L

z(E)=Z

~

(3)

- Ei )

i=l where

5(E_Ei)=fl

forf~

i

(4)

The normalization condition is L Z ( x i =1 i=1

(5)

Let us assume that ~ i quantity is related to the relative concentration of i-th component of the solid aerosol surface. Strictly speaking, the surface of the most components (subsurface) also is heterogeneous and the adsorption energy Ei

579 presents the first initial moment (average adsorption energy, E(av)i) of the adsorption energy distribution for surface of i-th component (i-th subsurface). Then, the patchwise topography of the adsorption sites on the aerosol surface is supposed. Also, the simplest localized Langmuir adsorption model without lateral interaction between adsorbed species on the surface is taken for description of relative coverage of adsorption sites of the i-th type [20] 0i(P):

KiP 1+KIP

(6)

where Ki is the Langmuir's constant for i-th type of the surface sites. This constant may be expressed as K i = K0 (T)exp/R~ /

(7)

where K0(T) is pre-exponential factor which depends on rotational, vibrational and translational degrees of freedom of adsorbed polyatomic molecule [21]. In present case, simple Jaroniec equation may be used in order to estimate this preexponential factor [22]

K0 = Ps exp

AHvap / RT

(8)

where AHvap is the heat of vaporization of the adsorbate and Ps is its saturated vapour pressure at temperature T. Therefore, problem for prediction of relative monolayer coverage of entire surface for solid aerosol with known surface composition reduces to the estimation of the adsorption energy Ei or adsorption free energy, AGi - - RT ln(Ki) for a given pair: i-th subsurface/organic pollutant. The routes for estimation of these quantities will be given below. It should be noted that usually adsorption of organic pollutants on an aerosol surface results from multicomponent gas mixture in the atmosphere. Above approach based on the Langmuir equation and adsorption energy discrete distribution may be extended to adsorption from multicomponent gas mixtures in accordance with [23]. The total relative surface coverage for adsorption of the m-component gas mixture on a homogeneous surface, 0m, is m 0m - E O j j=l

(9)

580 where 0j is partial relative coverage by j-th component of the gas mixture. The partial and total adsorption isotherms on the discrete heterogeneous surface may be expressed as follows rn

ZKjiPj 0i m(P,T)= ,

j=l

(10)

m

1 + Z KjiPj j=l and L O(P, T)= Z (zi~ m (P, T) i=1

(11)

where Zji is the Langmuir constant for interaction of j-th component of gas mixture with i-th subsurface and pj denotes the partial pressure of j-th m component at temperature T, P = Pj. j=l 4. "STRUCTURE-ACTIVITY" RELATIONSHIPS IN ADSORPTION OF ORGANIC COMPOUNDS ON INORGANIC OXIDE AND CARBONACEOUS MATERIAL SURFACES Two main routes exist for estimation of stability of adsorption complexes on the solid surface. Those are computation of formation heats of the complexes by quantum chemical and molecular mechanics methods, and application of various empirical quantitative "structure-activity" relationships for this purpose. Precision data for the structure of surface sites and for the potentials of intermolecular interactions are needed to use the first route. This restricts its application to the calculation of these quantities for the interaction of relatively simple small compounds (noble gases, n-alkanes, etc.) with well-characterized solid surfaces (single crystals of metals, their simple oxides and graphite). However, the surface of solid aerosols is heterogeneous. This heterogeneity is caused by their amorphous nature, possible microporosity and of the existence of several types of adsorption sites. It is extremely difficult to perform correct quantum chemical and molecular mechanic calculations of the adsorption processes due to the presence of large scale surface defects, impurities and especially a large number of different molecular configurations. It is evident that because the surface of various atmospheric solid aerosols has very complex composition these methods cannot be used in prediction of the adsorption complex

581 stability on their surface. In such a case various correlative approaches based on the combination of the thermodynamic functions of adsorption equilibria with physicochemical properties of adsorbates by simple relationships can be applied. Quantitative "structure-activity" relationships (QSAR) studies are unquestionably of great importance in chemical engineering. To obtain a significant correlation, it is crucial that appropriate descriptors should be employed, whether they are theoretical, empirical, or derived from readily available experimental characteristics of the structures. There are two groups of descriptors for possible relationships of those with thermodynamic functions of adsorption equilibria on the components of the solid aerosol surfaces. One of them is the group of empirical descriptors. They are derived from the experimental data for standard processes in solutions or in gas phase. However, for most organic compounds, these empirical parameters are lacking. An additional point to emphasize is that the correlations with these parameters are poorly justified theoretically and the intrinsic relationships exist between them. Second group of parameters is the quantum chemical or molecular mechanic descriptors. There are parameters of frontier orbitals (energies of high occupied (EHoMO) and lowest unoccupied (ELuMO) molecular orbitals, their maximum coefficients; aHOMOand aLUMO), electrostatic parameters (maximum positive and negative charges in the molecule (q+maxand q-max), electrostatic field averaged on the van der Waals surface of the molecule; FA1), electronic polarizability ((Ze) and dipole moment (~) [24]. In the case of physical adsorption, the coefficients of QSARs reflect the activity of surface sites for certain types of adsorption interaction, for example, their acid/base characteristics, polarizability or dipolarity. Most of quantitative results for adsorption of organic compounds on the surface of various inorganic oxides and carbonaceous materials (Henry constants, adsorption free energies and adsorption heats) have been evaluated from data of inverse gas chromatography at infinite dilution and at finite concentrations. Inverse gas chromatography is one of the most convenient methods for the determination of the surface properties of powders [25-29]. This method allows to examine a given material in terms of surface free energy and acid/base characteristics. When non-polar probes (n-alkanes) are used only London interaction exist between adsorbate and solid surface. The dispersive component of the surface free energy of any solid in Henry region may be determined through the measurements of the adsorption free energies for n-alkane series. This method was applied in the examination of several inorganic oxides, polymers and fibres. The adsorption energy (EA) and adsorption free energy (AGA) determined from retention data by this method for given pair polar adsorbate/polar solid surface may be presented as the sums of non-specific (mainly dispersion) and specific (donor/acceptor or acid/base) contributions F.nsp +EAsp EA ='~A

(12)

582 (13)

AG A : A O ~ sp + A G ~

Accordingly, the problem of EA estimation reduces to calculation of these contributions.

4.1. D i s p e r s i o n i n t e r a c t i o n s Dispersive forces are common to the adsorption interaction of all molecules whether or not they possess a p e r m a n e n t dipole moment. For non-polar adsorbates, e. g. aliphatic and aromatic hydrocarbons, dispersive forces are sole interacting forces between them and functional groups of solid surface. The London's dispersion term of intermolecular interaction between unidentical molecules 1 and 2 can be expressed as follows [30] 3 z

vI v 2

wd = --:h

Vl + V2

(101(102

NA (4~0

(14) G

where (10 is the molecular deformation (induction) polarizability, h is the Planck's constant, ~o is the permitivity of vacuum, r12 is the distance between spherical molecules 1 and 2, NA is the Avagadro' number, v is the characteristic vibrational frequency of electron expressed by [31]

lle2

v=~ 2~

(15)

a0m e

e and me the elementary charge (1.602x10 -19 coulomb) and mass (9.109x10 -31 kg) of electron, respectively. The distance of separation of molecules is nearly equal to 0.3 nm when the heat of vaporization of given adsorbate has the order of magnitude of 31 kJ mo1-1 or more. Therefore, energy of dispersion interaction between organic molecule (1) and functional groups of an solid surface (2) is proportional to term (hv 1)1/2(1ol. However, this p a r a m e t e r is u n k n o w n for most organic compounds. In such a case the molar deformation polarization (PD) calculated using a simple additive scheme for examined compounds can be used

PD =I0

M(n -l)

(16)

d(n~ + 2) where M is the molecular weight in g, d is the liquid density at 293 K in g cm -3 and nR is the refractive index of the liquid at 293 K. The p a r a m e t e r PD for most

583

popular organic solvents is related to t h e (hvl) 1/2 a 0 1

quantity by a simple

relationship [24] PD

=

(4.21+0.32)(hVl) 1/2 0~01 - ( 9 . 3 +- 2 . 6 ) cm

(17)

(N=ll;(~=0.545cm3"R=0.975), Therefore, non-specific interaction contribution into adsorption free energy of polar and non-polar adsorbates on solid surface may be written as follows -

sp = K p

(18)

• PD +

where Kp coefficient is proportional to average polarizability of the surface adsorption sites and ~ is a constant characteristic for a given solid surface. The Kp and ~ parameters for solid surface may be determined using adsorption free energies for n-alkanes series undergoing only dispersion interaction with a given solid surface. Let us consider an approach to estimate these parameters. It is based on use of known dispersive components of surface free energies for various solids evaluated from data of inverse gas chromatography at infinite dilution (in the Henry region). The dispersive component of surface free energy of solid (7~) in the Henry region may be evaluated from gas chromatographic measurements of the adsorption free energy for n-alkanes series on a examined solid surface [32]

d. ~'S

1[

. . . 4~/CH2 NAaCH2

2

(19)

where 7CH2 is the surface energy of methylene group (35.6 mJ m -2 at 293 K). The variation of 7CH2 with temperature is given by 7CH2 = 35.6 + 0.058• - T), (in mJ m -2) [28]. " a C H 2 " is the area occupied by a methylene group (aCH2 = 0.06 rim2), NA is the Avagadro' number and AGA(CH2) is the difference between adsorption free energy of n-alkane having nc+l carbon atoms in its molecule and n-alkane with nc carbon atoms, i. e., adsorption free energy of a methylene group. Eq. 19 may be transformed in the following relationship between the adsorption free energy at infinite dilution for n-alkanes series and their carbon number (nc) - AGA = n c • 2NAacH2

(Td~/2 S

)1/2 +b (7CH2

(20)

where b is a constant depending on the choice of a reference state of the adsorbed substance. From the other side, the molar deformation polarization for n-alkanes series (C5-C10) at 293 K is related to their carbon number as follows

584 PD = (46.83 + 0.31)n C + 16.25 + 2.427 cm3

(21)

(N = 6; (~ = 1.313; R 7= 0.9999) Substituting Eq. 20 and 21 into Eq. 18, the expressions for Kp and ~ p a r a m e t e r s may be written as Kp = NAaCH2 (YCH2)1/2 (Td ~ /2 23.42

(22)

and = b - 0.694 • NAaCH2

CH 2

S

(23)

The constant b depends on the choice of standard adsorption state, b = RT ln(ps,g/n), where ps,g is the adsorbate vapor pressure in gaseous s t a n d a r d state and n is the s t a n d a r d pressure. In the case of definition of L a n g m u i r constant in cm a per gram of an adsorbent, denominator of later expression contains the product of specific adsorption area of the surface multiplied by the weight of examined adsorbent. Various choices exist for the definition of standard states. De Boer and Druyer [33] define the standard surface pressure as that when the distance of separation between molecules in the adsorbed state equals that in the s t a n d a r d gaseous state, i. e. ps,g 1.013x105 Pa; n = 3.38x10 -4 N m -1, whereas Kemball and Rideal [ 3 4 ] suggested a different value for = 6.08x10 -4 N m 1. Because the values of molar deformation polarization for most organic molecules exceeds 100 cm 3, effect of (y~) quantity on the ~ value is low in comparison with its effect on the product PDxKp. It leads to a simple relationship (18) between AGA and PD quantity. Example of such plots for adsorption of n-alkanes series and polar organic compounds on the surfaces of rutile and mixed Ti, A1 and Si oxide is given in Figure 1. Therefore, Eq. 18 may be used for calculation of specific interaction contribution into adsorption free energy of polar organic compounds on the examined solid surface =

AG~polar ) = AGA(polar) - Kp x PD(polar) -

(24)

It should be mentioned that most frequently used descriptors of polar organic compound in correlation with the contribution of its dispersion interaction to the heat of adsorption or adsorption free energy are logarithm of s a t u r a t e d vapor pressure (log Ps) or vaporization heat (AHvap) [27]. However, the use of log Ps or AHvap quantities for polar compounds leads to overestimation of this interaction due to the considerable contributions of dipole-dipole and hydrogen bonding interactions between the probe molecules by itself to these quantities. Then,

585

30

~- ~

25

"7

Polar probes ~-

O ,at

TiOJA1203/Si~

20

~"

~ r ~f C10

~',

r

<

r~ <1

15

/

I

10

/

,/, 100

/

C9j/

%/ CVsgf

c6.,

TiO

, 200

300

400

5t

PD (cm3) Figure 1. Plot of adsorption free energies for n-alkane series and polar probes on the surface of rutile and titania/alumina/silica vs. molar deformation polarizability of the adsorbates. Figure is taken from [43].

more correct molecular descriptors for this purpose are to be the molecular induction polarizability or the molar deformation polarization. In addition, the sum of indexes of dispersion interactions for functional groups in the examined polar molecule, ID = EKsxPD• -6 where Ks is the coefficient of the intramolecular screening and r is the van-der-Waals' radius of this group, may be used in the calculations of the Esp or AG~p quantities for polar adsorbates [35,36]. The AHvap quantity for n-alkanes and polar adsorbates may be used for prediction of the lower limit for their Ensp value. Let us consider the data about Ys values for various initial and modified inorganic oxides as well as for some carbonaceous materials which are possible components of solid aerosol surfaces. These data are shown in Table 5. From this Table it follows that dispersive components of surface free energies for various nonporous and mesoporous solids lie in the range 30 + 130 mJ m -2, i. e. slope of plot AGA vs. PD may vary twofold. The dispersive components of surface free energy are proportional to overall polarizability of surface sites, their ionization energy and to maximum partial charges on atoms of these sites. Maximum y~ values are observed for graphite, graphitized black (Carbopack) and most parent and mixed Fe, Zn, Mg, A1, Ti oxides. This may be explained by high polarizability of extended graphite planes consisting of condensed aromatic nuclei and high partial charges on Lewis acid/base sites of the parent and mixed metal oxide surfaces. Pyrogenic silica possess minimum Ys among the inorganic oxides.

586 Table 5 Dispersive c o m p o n e n t s of surface free e n e r g y of various solids d e t e r m i n e d by u s i n g the i n v e r s e gas a d s o r p t i o n c h r o m a t o g r a p h y at infinite dilution for n - a l k a n e s series Solid

T, oC Yds' Ref. m J m -2

B a r i u m sulfate Silochrom $120 (S = 102 m 2 g-l) C a r b o n a t e d $120 (S = 96 m 2 g-l) Carbopack (S = 8.7 m 2 g-l) Silica gel (S = 80 m 2 g-l) Aminosilica gel (S = 72 m 2 g l ) G u a n i d i n o e t h a n t i o l - s i l i c a gel (S=75 m 2 g-~) Fluorinated graphite Carbopack B (S = 100 m 2 g-l) Pyrogenic silica (S = 170 m 2 g l ) Pyrogenic a l u m i n a (S = 140 m 2 g-l) Pyrogenic t i t a n i a (S = 70 m 2 g") Pyrogenic alumina/silica 30 wt% A1 in SiO2 (S - 170 m 2 g-l) Pyrogenic titania/silica 20 wt% Ti in SiO2 (S = 72 m 2 g-l) T i t a n i a (rutile) (S = 12 m 2 g-l) T i t a n i a / a l u m i n a / s i l i c a (4 wt% A1 a n d 2 wt% Si on the rutile surface) ( S - 14 m 2 g-l)

150 180 180 180 130 130 130 150 321 130 130 130 130 130 100

35.5 a 26.3 a 55.1 a 106.2 a 22.5 a 24.4 a 30.6 a 90.6 a 247.4 a 28.6 c 51.8 c 35.2 c 32.5 c 44.9 c 43.1

37 38 38 38 39 39 39 40 41 42 42 42 42 42 43

100

56.1

43

U n t r e a t e d c a r b o n fiber Oxidized carbon fiber Oxidized c a r b o n fiber coated with a n epoxy r e s i n C a r b o n black Graphite Calcium c a r b o n a t e Zinc oxide M a g n e s i u m oxide Smectite y-Fe203 y-Fe2Os, modified by stearic acid 7-Fe20~, modified by o c t a d e c y l a m i n e Pyrogenic silica (Aerosil 130) ),-A1203 Pyrogenic silica (Aerosil 200) B o e h m i t e (A1OOH) Pyrogenic silica (Aerosil A130) M e t h o x y l a t e d silica (Aerosil A130) H e x a d e c y l a t e d silica (Aerosil A130)

100 100 100 30 44.5 30 25 25 180 100 100 100 110 100 100 80 20 20 20

53 56 36 42.8 129 44.6 85.3 95.6 159 91.7

44 44 44 45 46 47 48 48 49 a

50

31.0

a

50

35.0

a

50

46.5 115 65 172 75 70 38

32 51 51 51 52 52 52

587 Table 5 Dispersive components of surface free energy of various solids determined by using the inverse gas adsorption chromatography at infinite dilution for n-alkanes series Solid

T, ~

7ds' mJ m -2 Ref.

Precipitated silica (Spherosil XOB 75) Alumina/silica (1050 ppt SiO2 in AleO3) Alumina/silica (45 ppt SiOz in A120~) Synthetic activated carbon obtained by carbonization of polyfurfuryl alcohol (S = 408 m 2 g-l)

20 100 100

80 42 100

60

560.5 55 (37.9) b

53 54 54

S is the specific adsorption area of the solid surface Calculated from dependence of adsorption free energies for n-alkane series on the number of carbon atoms in these n-alkanes; the adsorption free energies were calculated on basis of the retention volumes or Henry constants for n-alkanes presented in cited studies. b Dispersive component of surface free energy for flat surface, evaluated from original ,/ds value taking into account radius of the carbon micropores and the adsorbate molecule size [55]. c Average dispersive component of surface free energy in the monolayer region was determined from data of inverse gas adsorption chromatography for n-alkanes series at their finite concentrations [4.2]. a

Mixed Ti, A1 and Si oxides in some cases exhibit positive synergistic effect, i.e. they show more high polarizable properties in comparison with parent oxides. Modification of the inorganic oxide surfaces by organic functional groups gives decrease of the ),~ value and this effect increases with elongation of alkyl chain in the grafted group. From the other side, carbonization of grafted organic groups leads to increase of the surface polarizability. Oxidation of the carbonized layer enhances the 7sd value, whereas its coating by organic polymer leads to decrease of the polarizability. The most high 7~ values are observed in the case of microporous inorganic oxides, such as smectite and boehmite and carbonaceous materials, as activated carbons and Carbopack B. It was shown that these high values may be explained by effect of the average micropore radius on the apparent ysa values [55]. The dispersive components of surface free energy for the flat activated carbon surface (7~ (0)), evaluated from original 7as values taking into account average radius of the carbon micropores and the adsorbate molecule size correspond to the usual 7~ range for nonporous solids. However, in the predictions of Kp values for microporous solids, the apparent y~ values are to be used.

588 It is known that the ~,~ values also depend on the temperature of preliminary treating the inorganic oxide surface. This temperature determines concentration of highly polarizable Bronsted and Lewis acid/base sites on the oxide surfaces. For example, when the number of silanol groups diminishes monotonously upon increasing the heat treatment temperature of pyrogenic silica surface from 100 to 700 ~ C, ~,~ changes quite unexpectedly from 75 mJ m -2, going through a maximum at 500 ~ C (y~ - 95 mJ m 2) [56]. This was explained by maximum concentration of three member and highly polarizable siloxane rings generated at about 500 - 600 ~ C under silica dehydroxyllation Mechanisms of these rings formation are considered in [57] and references cited therein. Beyond this temperature, the rings transform into 4 membered, non-polar siloxane rings and ~ value decreases to 80 mJ m -2 at Tp - 700 ~ C. Therefore, depending on the humidity of atmosphere, the y~ values may be varied in the range of 20 mJ m -2 or higher. It should be noted that formation and interaction of solid aerosol particles take place in a humidic atmosphere. Then the hydration of the surface and formation of water droplets or ice patches on the aerosols are universally present in the native atmospheric processes. This effects the optical properties of solid aerosols. The theory of the formation of water droplets on aerosol particles consisting of inorganic and organic compounds was developed in [58]. The theory links interfacial free energies of the surfaces and chemical characteristics of the compounds with the ability of the aerosols to form water droplets. However this theory takes not into account perturbing action of the solid surface on the properties and structure of the interfacial water layers. It is known that the freezing temperature of water near the solid surface of various parent and composite inorganic oxides (of Si, A1, Ti and Zn) and carbonaceous materials (activated carbons, carbon black, carbonized silicas) is lowered from 273 K up to 190 K. From X-ray diffraction studies of adsorbed water molecules it follows that the number of nearest-neighbor molecules of the adsorbed water is less than that of bulk liquid water and that the adsorbed water has a long-range ordered structure compared with bulk liquid water [59]. This freezing temperature depression depends on the surface free energy of solid and on the distance of adsorbed water molecules from solid surface [60-66]. The higher value of interfacial free energy, the thicker is water layer susceptible to the action of the surface forces. The thickness of the unfrozen water layer reflecting the perturbing solid surface effect on the structure of the interfacial water varies from 3.0 nm (pyrogenic silica) up to 20 nm in case of some carbonized silicas with minor contain of carbon patches on the silica surface and composite titania/alumina/silica (4 wt% A1 and 2 wt% Si on the rutile surface). The dependence of water free energy versus thickness of the unfrozen water layer near the surface of rutile and mixed Ti, A1, Si oxide is shown in Figure 2. One of the main reasons for the long-range surface forces is the polarization property of

589

3.0 2.5 ~

2.0

<~

i

1.5

I

TiOffA12OffSiO 2

1.0 0.5 0

I

I

7.5

I

I

15.0

d (nm)

Figure 2. Dependencies of water free energy change on the thickness of unfrozen water layer near the surface of rutile and titania/alumina/silica. Figure taken from [65].

solid surface. The polarizability of the adsorption sites on solid surface gives contribution to the total polarizability of the particles. Such tendency in the change of polarizability of the metal oxides series can be derived from the comparison of the dispersive components of the oxides surface free energy. Similar correlation between 7~ values and Hamaker' constants for solid surfaces is well known [67]. The relationship between logarithm of the average water layer thickness on the surface of parent and mixed Si, Ti and A1 oxides, evaluated from data of 1H NMR spectroscopy with bulk freezing and the 7~ values for these oxides is given in Figure 3. The increase in the dispersive component of metal oxides surface free energy leads to enhancement of the thickness of water layer perturbed by solid surface. Hence, presented in Table 5 dispersive components of surface free energies for various solids allow to estimate the contribution of the non-specific interaction to the overall adsorption free energy of organic pollutants on the surface of solid aerosol constituents and to predict the thickness of water layer perturbed by solid surfaces as well as the freezing temperature depression for water droplets on the aerosol surfaces. The dispersive components of surface free energy for organic fraction of solid aerosol surface may be predicted from additive functions, for example, from molar parachor (Pp) [68]. If the group contributions of Pp and molar volume of organic fraction (VM) are known, the y~ value results from the expression

590

_

Ti AO21203//A1203/SiO2zs

E

25

I

I

35

45

55 7d , (mJ m -2)

Figure 3. Relationship between logarithm of average water layer thickness on the surface of inorganic oxides and dispersive components of surface free energy of oxides. Figure taken from [65].

pp = (Td~/4M d

(25)

where M and d are the molecular weight and density of organic fraction, respectively. Also, the 7~ value for the organic fraction can be estimated on the basis of dispersive contribution into its solubility parameter (5 d) [69] 7S = 0.75x

(26)

These approaches were applied with profit to estimation of ),~ values for polystyrene, its co- and tert-polymers below their glass transition temperatures [29]. Calculated data agreed satisfactorily with experimental 7sd values, evaluated from data of inverse gas chromatography for n-alkanes series on the examined polymers. Similar way may be used for prediction of 7sa values for carbonaceous aerosols containing large amount of organic carbon. One of unresolved problems is prediction of non-specific interaction contribution into adsorption free energy for geometrical isomers on the solid surface. For example, cis/trans olefines have the same PD values. However, their

591 adsorption energies on various solids are dissimilar. The adsorption heat of organic compound is determined more by its molecular structure t h a n the bulk as a whole. Hence, the application of graph theory in this connection may be of some merits, though the structures here are approximately represented by graphs without taking into account of the geometrical details [70]. For example, one graph theoretical index (~(q)) is discriminative between geometrical isomers, such as cis-trans alkenes and includes the influences of different hydrogen atoms in molecules [71]. Adsorption heats (AQA) in the Henry region for cis- and transalkenes on the graphitized black surface from [72] are related to these ~(q) indexes as follows AQA = (9.3 + 1.3) F~(q) + (14.3 + 2.3) kJ mo1-1 (c~ = 0.73 kJ mol-1; R = 0.971)

(27)

Also, topological indexes taking into account the differentiation of heteroatoms in the molecular graphs (~ [71] and H' [73]) can be used for description of nonspecific adsorption of alkyl halides and ethers on the graphitized black surface from [72] AQA = (9.1 + 1.4) ~ + (14.2 + 3.2) kJ mo1-1 ((~ = 1.86 kJ mol-1; R = 0.934)

(28)

as well as for nonspecific adsorption of halogen substituted methanes and ethanes on the graphitized black surface from [72] AQA = (1.3 + 0.4) H' + (23.5 + 2.9) kJ mo1-1 ((~ = 3.9 kJ mol-1; R = 0.863)

(29)

Other examples of relationships between adsorption data of various halogenhydrocarbons C1-C5 on a modified silica surface in the Henry region and graph theoretical indexes for these compounds are given in [74]. One of the most important problems in the qualitative description of organic pollutants adsorption on aerosol surfaces is the prediction of adsorption energies of ozone-depleting chlorofluorocarbons on solid surfaces. Unfortunately, there are few studies devoted to this problem. Crude estimation of the adsorption energies may be performed using the PD quantities for adsorbates and 7~ values for solid subsurfaces as well as by using above discussed H' and other graph theoretical indexes for these compounds. More simple relationship between adsorption energies of 17 hydrocarbons, chlorofluorocarbons, ethers and sulfur hexafluoride adsorbed onto porous Carbopack B was obtained in [41]. The authors have found that adsorption energies of these compounds are related to their critical constant ratio - critical temperature (Tc) divided by square root of critical pressure (Pc) as follows

592

nsp

/

819.27

(30)

(N = 17; R = 0.984) where Ensp and Tc are in K, and Pc is in bars. Other important problem in this field is related to prediction of adsorption energies for low-volatile polycyclic aromatic hydrocarbons (PAH), which are main part of the organic carbon in the solid aerosols .Their sublimation heats or PD values may be chosen to estimate the lower limit of the adsorption energies on the solid surfaces. PAH possess highly polarizable n-electron system, low ionization potential (IP) and they interact with OH groups of inorganic oxide surface forming H-complexes or through complete H § transfer from these groups to PAH molecule, while with Lewis acceptor sites of the surfaces they interact with formation of charge transfer complexes with these sites or products of complete electron transfer from PAH molecules to these sites, radical-cations [75,76]. Some adsorption characteristics for PAH on the parent and modified silica gel surfaces are found from data of inverse liquid chromatography [77, 78]. The logarithms of adsorption distribution coefficients of 13 PAHs from n-hexane onto parent silica gel, and silica gels with grafted CN and NH2 groups are related to the molecular descriptors of the compounds as

fn{n,

(31)

where n~ is the number of n-electrons in the PAH molecule, 1 is the length of the molecule, no(in) is a number of "intrinsic" carbon atoms in PAH, while fn and A are empirical coefficients. Then, the terms n~/1 and nc(in) must be taken into consideration in possible relationships between the adsorption heats of PAHs series on the metal oxide surfaces and their PD and IP quantities. The contribution of dipole-dipole interaction between polar organic compound and element-oxygen bonds of inorganic oxide surface to overall adsorption heat must be taken into consideration in the QSARs for adsorption on dehydroxylated inorganic oxide surfaces [24]. For example, the differential adsorption heat of polar organic compounds on the surface of dehydroxylated SiO2 depends on the dipole moment of the adsorbates (~, in Debye) and their molar deformation polarizability as follows AQA = 1.41 PD + 5.36 ~t (N = 12; a = 1.39 kJ mo1-1)

(32)

593

4.2. Donor/acceptor (acid/base) interaction The main contributions to the AG~ and Esp quantities give formation of hydrogen-bonding or donor-acceptor complexes between polar organic adsorbate and polar adsorption sites of solid surface. The Bronsted and Lewis acid-base sites are main types of adsorption sites for polar organic compounds on the surface of metal oxides, salts and various oxidized carbonaceous materials. The quantitative description of the hydrogen-bonding and donor-acceptor interactions of the sites with adsorbed organic compounds is very important for determining of acid~ase characteristics of the surface sites and prediction of the adsorption characteristics. Among possible approaches the E&C equation of Drago and Wayland [79] is applied for description of specific interaction contribution into adsorption heat (AQ~), or adsorption energy Esp in the Henry region sp = EAEB + CACB AQA

(33)

where EA and EB are the electrostatic and CA and CB a r e the covalent contributions of acid (A) and base (B) in the heat formation of donor/acceptor (acid~ase) adduct. Also, more simple Gutmann's approach has been widely used for determining the acid/base characteristics of various solid surfaces [80] AQSP K A x D N K B • = + A 100 100

(34)

where DN and AN are donor and acceptor of a polar adsorbate and KA and KB coefficients denote the acceptor and donor numbers of solid surface. Similar to Eq. 34 the relationship between AG~ and DN, AN quantities for probe polar adsorbates is used for characterization of solid surfaces by means of inverse gas chromatography. The coefficients of Eq. 34 are easily determined using the leastsquares method from the plot of AG~/AN vs. DN/AN for the series of polar adsorbates characterized by different DN and AN values. However, different DN and AN values for the probe adsorbates are usually used for the evaluation of the KA and KB quantities for solid surfaces. Riddle and Fowkes [81] have shown that 31p NMR spectrum of triethylphosphine oxide used in determining the AN values for polar compounds, is appreciably shifted downfield due to van der Waals interactions with the solvents. Hence, AN values were corrected for the van der Waals contribution to the chemical shift. The corrected AN values for polar adsorbates are designated as AN*. The corrected K~ values are found from Eq. 35 by analogy to Eq. 34 AQ~

=

KAXDN K ~ • + 100 100

(35)

594 The tendencies between such empirical quantities of organic compounds as E, C parameters, donor n u m b e r and q u a n t u m chemical indexes, such as m a x i m u m positive and negative charges in the organic molecule as well as with energies of highest occupied molecular orbitals (EHoMO) and lowest unoccupied molecular orbital (ELuMO) for these compounds have been found in [83] C A = -(8.9 + 8.1) x ELUMO + 0.8 + 0.1 ( N - 18; ~ - 4.9; R - 0.263)

(36a)

C B = (2.3 + 0.7) x EHOMO + 29.9 + 7.9 ( N - 14; a - 103.4; R - 0.671)

(36b)

+

E A = (8.2 + 2.3) x qmax + 2.4 + 0.9

(36c)

( N - 18; ~ - 77.9; R - 0.668) m

E B = (0.70 + 0.67) x qmax + 0.8 + 0.2

(36d)

( N - 14; ~ - 0.91; R - 0.289) DN = (44.1 + 35.0) x qmax + 7.3 + 8.9

(36e)

( N - 7; ~ - 17.4; R - 0.510) These relationships allow to predict the E, C, DN quantities for unstudied polar adsorbates. In recent it was shown that solvation equation developed by A b r a h a m [84, 85] may be used with profit for description of polar organic compounds adsorption on the surface of fullerene and various carbonaceous materials [86, 87] log(SP)=c+rxR 2 +sx~H +ax 2a2H +bx2p

H + I x l o g L 16

(37)

where SP is property of a series of solutes in a fixed phase, R2 is an excess molar refraction, ~H is the solute dipolarity/polarizability, 2 a H and 2 p H are the solute hydrogen-bond acidity and basicity, respectively, and log L 16 is derived from the solute/hexadecane partition coefficient at 293 K. As the descriptors in Eq. 37 refer to particular properties of the solutes, the coefficients a and b correspond to the hydrogen-bond basicity and acidity of solid surface, respectively. It should be mentioned that some intrinsic relationships exist between various scales of the acid/base characteristics of polar organic compounds discussed above. For example, acceptor and donor numbers of organic compounds from Table 6 are related to the solvatochromic solute effective or overall hydrogen-bond acidity and hydrogen-bond basicity as follows

595

AN-

(73.6 + 6.6) E aH + 9.39 + 1.54

(38a)

(N = 11; c = 4.17; R = 0.966) DN-

(23.7 + 9.9) E ~ H + 3.53 + 3.89 kcal mo1-1

(38b)

(N = 12; (~ = 6.27; R = 0.603) Also, following t e n d e n c i e s exist b e t w e e n t h e c o v a l e n t a n d e l e c t r o s t a t i c c o n t r i b u t i o n s of acid (A) a n d b a s e (B) to h e a t of a c i d / b a s e i n t e r a c t i o n a n d t h e h y d r o g e n - b o n d a c i d i t y a n d h y d r o g e n - b o n d b a s i c i t y of t h e o r g a n i c c o m p o u n d s from T a b l e 6. Table 6 P o l a r o r g a n i c c o m p o u n d s in t h e E a r t h a t m o s p h e r e , t h e i r h i g h e s t c o n c e n t r a t i o n s in u r b a n a n d forest a t m o s p h e r e a n d t h e i r a c i d / b a s e c h a r a c t e r i s t i c s Group of Highest concentration, Representative Acid/base characteristics compounds ~g m -3 Aliphatic CH3OH (81) [88] CH3OH a = 0.43; 13 = 0.47 AN = 41.5; AN* = 12.0; alcohols C2HsOH (140) [88] DN = 19.0 C1-C4 C4HgOH (1470) [89] Aliphatic HCOOH (39) [ 9 0 ] CH3COOH a = 0.61; 13= 0.45 AN = 52.9; AN* = 21.16; organic acids CH~COOH (16) [91] DN=20 C1-C2 Aliphatic (CH3)2C=O (620) [ 9 2 ] (CH3)2C=O CB = 2.33; EB = 0.937 ketones CH~(C=O)C2H5 (45) a = 0.08; 13 = 0.48 AN=12.5; AN* = 5.0; DN = 17.0 C2-C6 [93] A N = 1 3 . 4 ; D N = 18.9 Aliphatic H2C=O (67) [ 9 7 ] CH~C(=O)H aldehydes CH~C(=O)H (188) [90] C1-C4 Aliphatic (C2H5)20 (> 0.01) [ 9 8 ] (C2H~)20 CB 3.25; EB - - 0 . 9 3 6 = 0; [3 = 0.45 AN= 3.9; ethers AN* = 1.56; DN = 19.2 C2-C4 CB = 1.74; EB = 0.975 Aliphatic C4H802 (> 0.01) [ 9 9 ] CH~C(O)O a = 0; [3 = 0.45 AN = 9.3; esters of acetic C5H1002 (> 0.01) [104] -C2H5 AN* = 1.5; DN = 17.1 acid C1-C5 Aliphatic CH3CN (> 0.01) [100] CH3CN CB = 1.34; EB = 0.886 nitriles c~ = 0.04; 13= 0.33 AN=18.9; AN*= 4.7; DN = 14.1 Aliphatic (CH3)2NH(>0.01) [101] (CH3)2NH Cs = 8.73; EB = 1.09 amines C2HsNH2 (>0.01) [101] a = 0.08; [3 = 0.66 AN = 9.4; DN = 40 C6HsSH AlkylCHzSH (> 0.01) [102] CA = 0.198; EA = 0.99 mercaptanes CH~SCH3 CB =7.46; EB =0.343; IP= Dialkylsulfides CH3SCHz (1.4) [94] 8.69

596 Table 6 P o l a r organic c o m p o u n d s in the E a r t h a t m o s p h e r e , t h e i r h i g h e s t c o n c e n t r a t i o n s in u r b a n a n d forest a t m o s p h e r e a n d t h e i r a c i ~ b a s e c h a r a c t e r i s t i c s Group of Highest concentration, Compounds ~g m ~ Representative Acid/base characteristics Dialkyldisulfides Cyclic ethers

CHaS-SCHa (> 0.01) [103] 1,4-C4HSO2 (> 0.01) [98,99]

C2H~S-SC2H5

IP = 8.27; (CB = 10.9)

1,4-C4Hs02

Alkyl furanes

2-CH3-C4HaO (> 0.01) [104] 3-CH3-C4H30 [105] C6H5OH (> 0.01) [163] o-CH3C6H4OH (> 0.01) [106] (CHa)2N-N=O(0.76) [95] CFCla (6 ppb) [107] CF2C12 (2.4 ppb) [107] CFaCCla (25 ppt) [108] CCI4 (1.4 ppb) [109] CHaC1 (7.7 ppb) [110] CHCla (225 ppt) [110] CH2C12 (12 ppb) [111] CH3CC13 (10 ppb) [18] C2HC13 (4 ppb) [104] C2C14 (3.7 ppb) [108]

C4H4O

CB = 2.38; EB = 1.09 a = 0; 13= 0.37 AN=10.8; AN*=4.32; DN=14.8 CB = 3.16; EB = 1.38 AN = 3.3; DN = 3.3

(C2H5)4Pb (0.7) [ 9 6 ] (CH3)2Hg [112] CH3HgC1 [112] C6H6 (714) [113] CH~C6H5 (642) [113]

AI(CH3)3

CloHs (12) [114]

C10Hs

Aromatic hydroxyderivatives Nitrosoamines Halogenhydrocarbons

Organometalli c compounds Aromatic hydrocarbons

C6H~OH

CsHsN=O CC14

AN = 8.6; AN* = 3.44; DN = 0

CH2C12 CA = 0.01; EA = 1.66; a = 0.10; 13= 0.05 AN= 20.4; AN* = 8.16; DN = 1 CHC13

(CH3)3SnC1 C6H6

(AH) Polycyclic AH

CA = 0.44; EA = 4 . 3 3 (z = 0.60; 13 = 0.31 D N = 11 C B = 4.52; EB = 1.34

CA = 0.16; EA = 3.02 (z = 0.15; 13= 0.02 AN=25.1; AN* = 9.24; DN=4 CA = 1.43; EA = 16.9 0.03; EA = 5 . 7 6 CB =1.4; EB = 0.11;a = 0; = 0.14; IP = 9.25 AN=8.2; AN* =3.28; DN=0.1 I P = 8.12 CA =

CA, CB, EA, EB are the covalent and electrostatic contributions of acid (A) and base B) into heat of acid/base interaction, in (kcal moll) I/2 from [115] cz = E ctn and [3 - E [3H are the solvatochromic solute effective or overall hydrogen-bond acidity and hydrogen-bond basicity parameters in arbitrary units, from [84, 85]. IP is first ionization potential of the compound, or its EHOMOwith opposite sign, in eV, from [116]. AN and DN are acceptor and donor numbers of the compound, in arbitrary units and in kcal mol -l, respectively, from [80]. AN* is acceptor number corrected for the van der Waals contribution, in kcal mol l, from [81].

597 CB = (12.8 + 4.6) E ~ n - 2.2 +_2.0 (kcal mol-1)1/2

(39a)

(N = 7; a = 1.79; R = 0.780) E B - (1.73 + 0.56) E ~ n + 0.15 + 0.25 (kcal mol-1)1/2

(39b)

(N = 7; a = 0.22; R = 0.807) CA- (0.77 + 0.20) E aH - 0.014 + 0.074 (kcal mol-1)u2

(39c)

(N = 3; (~ = 0.08; R = 0.966) E A - (4.4 _+2.0) E aH + 1.76 + 0.75 (kcal mol-1)1/2

(39d)

(N = 3; a = 0.81; R = 0.903) The above acid/base characteristics are known for m a n y polar organic compounds observed in the E a r t h atmosphere. The highest concentrations of these compounds in the u r b a n and forest atmosphere and acid/base characteristics of representative compounds from these groups are given in Table 6. In case of absence of these p a r a m e t e r s for some groups of the compounds, such as for N-nitrosoamines, the characteristics for compounds containing similar functional groups are given in Table 6. Therefore, using the acid/base characteristics for representative compounds from Table 6 and acid/base characteristics for components of solid aerosol surfaces from forthcoming Table 7, one may predict the specific interaction contributions into adsorption free energy or adsorption heat. By adding them to non-specific interaction contribution calculated from PD and y~ data for these pairs, L a n g m u i r constants and relative concentrations of these pollutants on aerosol surface may be obtained. These acid/base characteristics for other members of groups are given in cited papers. In the of u n k n o w n characteristics for a member, they may be approximately equaled to those from Table 6, or estimated from extrapolation or interpolation of the plot for characteristics vs. n u m b e r of carbon atoms in the compound for a given group. Such examples for dependencies of molar transition energy of 2,4,6-triphenylpyridinium-4-(2,6diphenyl)-N-phenoxide (ET) reflecting the acidic characterictics of the adsorbates for series of aliphatic n-alcohols and fatty acids vs. carbon n u m b e r in their alkyl substituents are presented in Figure 4. The ET values are determined for more great n u m b e r of organic liquids t h a n other acid/base characteristics from Table 6. This quantity for 21 organic solvents is related to their acceptor n u m b e r (AN) as follows: AN = 1.29• - 40.52 [82]. The a c i d ~ a s e characteristics of inorganic oxide and carbonaceous material surfaces which may be the components of solid aerosol surfaces are derived from microcalorimetric data, calorimetric and adsorption titrations method, infrared spectroscopy, inverse gas chromatography, and zetametry of polar organic

598

Table 7 A c i d / b a s e c h a r a c t e r i s t i c s of i n o r g a n i c oxide a n d c a r b o n a c e o u s m a t e r i a l s u r f a c e s Donor and Q u a n t u m chemical and Solid acceptor E and C contributions other descriptors n u m b e r s (IGC) -EHoMO -- 11.2; -ELuMO -- 2 . 0 EA = 4.39; CA = 1.14 Silica K A - 11.7; K~ EA(ad)= 0.31 for s-SiOH (MC/IR) [117] = 21.8 [1261 [83,122] EA = 2.30; CA = 2.48 K B - 5.8 [126] (strongest sites) (CalELUMO(est) = - 0 . 1 ; q+max ---0.78 [24] Ad) [121] EA = 8.8; CA = 1.67 (MC) ELUMO(est) ---- 0 . 0 6 ; q+max -- 0.3 [24] [123] Ed = 772 [132] EA = 4.86; CA = 0.28 ET(30) = 58.0 [133] (MC) [124] Ed = 1068 (I); Ka=13; KB = 29 EA = 5.67; CA = 1.02 Titania 789 (II) [132] (MC) [117] [42]; Ka = 12; KB = 22.5 (ZA) [127]; KA = 24; Ks = 19 [127] KA = 23; KB = Titania/silica 29 [42] (20 wt% Ti in silica matrix) K A - 36; K B Titania/alumin 33 [43] a/silica 2 wt% Si and 4 wt% A1 on rutile surface EA = 4.50; CA = 0.8 (MC) ~-Fe203 [118] EA = 0.5 e 1.0; CA = 1.1 ~-Fe203 (MC) [119] E A = 0 . 5 ; C A = 0 . 0 2 ; EB = E glass 0.2; CB=0.39 (MC) [117] K a - 34; K B Untreated 40 [130] glass fiber EA = 6.0; CA = 0.7 (IGC) Glass beads [120] ET(30)- 51.4 [133] EA = 0.62; CA = 1.60 APS t r e a t e d (IGC) [120] glass beads E d - 1249 (I); 1155 (II) Magnesium 6.5; K ; = [132] oxide 54.3 [125] Zinc oxide 7-Alumina

K A - 19.8; K~ = 9.6 [125] K a - 16.8; K ; = 46.7 [126] K s - 12.4 [126]

E d - 1077 (Ib); 940 (Ia); 889 (IIb); 761 (IIa)[132] ET(30) - 65.0 [133]

599 Table 7 Acid/base characteristics of inorganic oxide and carbonaceous material surfaces Solid Donor and E and C contributions Quantum chemical and acceptor other descriptors numbers (IGC) a-Alumina KA = 21; KB = 15 [127] KA = 23; KB = 15 (ZA) [127] Alumina/silica t ~ - 32.9; K (30 wt% A1 in = 34.4 [126] silica matrix) Ks = 9.6 [126] Alumina/silica EA = 22.3; CA- 3.7 (MC) E L U M O ( e s t ) - - - 0.3; (0.23 wt% A1 in q+max= 2.4 [24] [1231 silica matrix) Carbonized KA = 1.0; Ks = carbon fiber 1.4 [128] Stabilized KA = 1.5; KB = carbon fiber 1.3 [128] Carbonized KA = 6.5; K~ = pitch-based 8.0 [129] fiber Graphitized KA = 3.6; KB = pitch-based 5.2. [129] fiber Untreated KA = 1.8; KB = carbon fiber 3.4 [131] Oxidized KA = 9.2; Ks = carbon fiber 10.0 [131] MC is microcalorimetry, IR is infrared spectroscopy, IGC is inverse gas chromatography, CalAd is calorimetric and adsorption titrations method, ZA is zetametry. ET(30) is molar transition energy (in kcal mol 1) of 2,4,6-triphenylpyridinium-4-(2,6-diphenylN-phenoxide [ 133]. Ea is difference between formation energies of surface conjugate base (s-O) and surface OH group (s-OH) (in kJ mol l ) computed by ab initio ECP-ST method [132]. ELUMO(est) is energy of lowest unoccupied molecular orbital of surface site (in eV), estimated from relationship between ELUMO(and CA [24]. q+max is maximum positive charge in the surface site (in atomic units), estimated from + relationship between q maxand EA [24]. EHOMOand ELUMOare energies of highest occupied and lowest unoccupied molecular orbitals of surface cluster for OH group, in eV, in MNDO approximation [6,7]. EA, EB and CA, CB are electrostatic and covalent contributions of acceptor (acid) and donor (base) surface sites into the adsorption heat, in (kcal moll) 1/2. KA (in arbitrary units) and KB (in kcal tool l ) are acceptor and donor numbers of surface sites; K~ is the donor number of surface sites (in arbitrary units) calculated using acceptor numbers for probe adsorbates corrected for van-der Waals contribution (AN*).

600 50 Aliphatic n-alcohols

"•

55 c acids

50

-

5

0

I

I

I

I

I

1

2

3

4

5

6 nC

Figure 4. Dependencies of ET acceptor parameter for series of n-alcohols and fatty acids on the number of carbon atoms in their alkyl substitutients.

compounds on the examined solid surfaces and from quantum chemical calculations. These characteristics are presented in Table 7. It is evident from this Table that carbonized carbon fiber which may be chosen as the model of carbon subsurface of a solid aerosol has minimal donor (KB) and acceptor (KA) properties. Its graphitization leads to decrease of these quantities, whereas its oxidation gives enhancement of the acid/base characteristics. The latter process may be occurred at interaction of the carbonized part of the aerosol surface with different inorganic acids (HNO3, H2S04, HC1) and acidic gases (SO2, NO2 and C12) in the atmosphere. The electron donor ability or basicity of adsorption sites of the inorganic oxides in accordance with the change of their KB and K~ values increases in the order MgO > TiO2/A1203/SiO2 >A1203 > TiO2//SiO2 > TiO2 > ZnO ~ A1203/SIO2 > SiO2 and the electron acceptor ability (acidity) of the inorganic oxides, as it follows from their KA values, varies as follows: TiO2/A1203/Si02 > Ti02/SiO2 > A1203/Si02 > ZnO > A1203 > TiO2 > Si02 > MgO From the comparison of CA and EA contributions for the oxides it follows that they have more high EA than CA values, i.e. they exhibit "hard" acidic properties. This was explained by a large contribution of the electrostatic field effect of the oxide solid matrix to the adsorption heat of test bases [24]. The most high EA

601 quantity was estimated for alumina/silica with 0.23 wt% A1. This may be due to formation of strong Lewis acid sites and enhancement of the acidity of OH groups which are bound to trigonal aluminium cations on the boundaries between patches of A1203 and SiO2 lattice. Modification of inorganic oxide surface by "soft" organic groups, as in the case of grafted 7-aminopropylsilylated glass beads, gives high decrease of the EA contribution, whereas the CA contribution even increases after the surface treatment. Some q u a n t u m chemical indexes of inorganic oxides and acceptor characteristics (ET) are presented in the last column of Table 7. The q:a.~ and ELUMO quantities were estimated in [24] for silica and alumina/silica surfaces from the relationships between these indexes and CA, and EA contributions. The range of these indexes is close to those computed by quantum chemistry. The experimentally derived ET parameter reflecting the acidic properties of solid surface increases with transition from silica to alumina and decreases in the case of 7-aminopropylsilylated silica in comparison with untreated silica. This order coincides with one determined from change of KA and EA for these solids. In addition, the proton affinities of conjugated bases of the oxide surface OH groups computed by q u a n t u m chemistry (Ea) are presented for some oxides. This quantity characterizes the acidity of the surface OH group in gas phase. The acidity of OH groups of the metal oxide surfaces determined by such manner decreases in the following order 7- A1203 > SiOffamorph) > TiO2(rutile) > MgO It may be seen that this order is close to one determined from change of KA quantity for the metal oxide surfaces. Crude estimation for ability of inorganic oxide surface to interact with organic compounds as acid or base can be found from mean electronegativities of these solids [70]. The mean orbital electronegativities of a solid ()~s) can be calculated using following expressions

~:S =

n i x 7~i Z ni

~i = IPi + EAi

(40)

(41)

where ~i is the Mulliken' orbital electronegativity, first ionization potential and electron affinity of i-th atom in an inorganic molecule, while ni is the number of these atoms in the molecule. These quantities for various inorganic oxides, salts and some minerals are presented in Table 8.

602 Table 8 Mean orbital electronegativities (~) of various solids (taken from [70] Solid

Z, eV

Solid

Z, eV

K20 Na20 CaO SrO BaO Ca3SiO5 MnO CaSiO4

4.06 4.34 4.81 4.87 4.96 5.41 5.44 5.58

MgO 5.68 Ca~Si207 5.71 La203 5.75 A1203 5.82 PbO 5.88 CaSiO3 5.89 Mg2SiO4 6.08 CaCO3 6.19

Solid

)~, eV

Solid

ZnO B203 TiO2 Fe203 MgSiO~ ZrO2 V205 A1PO4

6.14 6.18 6.21 6.22 6.24 6.24 6.44 6.50

Talc MgCO3 Mo3PW1204o SiO2 Kaolinite MoO3 WO3

~, eV 6.51 6.54 > 6.56 6.61 6.61 6.69 6.72

The acidity of the solid surface increases as mean solids electronegativity increases, whereas basic properties change in the opposite direction. Actually, it is known that alkaline, alkaline-earth metals, Mg and Zn oxides exhibit mainly basic properties in various catalytic transformations whereas Zr, Ti, Si, Mo oxides, clays and heteropoly acids typically display acidic properties. The conclusions from Table 8 coincide with observed order of KB change at transition from SiO2 to MgO. Then, a strong interaction should be expected between typical organic bases (ethers, amines, nitriles, etc.) and surface sites of last group of inorganic solids from Table 8. It should be mentioned that the behavior of OH groups at oxide surfaces depends strongly on the composition of oxide and on local chemical environment. It is known that OH groups of the alumina surfaces display wide spectrum of Bronsted acidic and basic properties [134]. The single OH groups of the surfaces exhibit mainly typical basic properties, whereas the bridged OH groups bounded to trigonal aluminium atom behave as typical Bronsted acid sites [135]. On partially dehydroxylated surfaces, they have various acid/base character and interact with organic adsorbates according to their own characteristics (IP, EA quantities, proton affinity and partial charges on the atoms in the surface clusters). Therefore, the acidic and basic properties of oxide surfaces depend on the temperature of surface pretreatment, surface coverage of adsorbates and concentration of physically adsorbed water which effects on the stability of adsorption complexes between organic adsorbates and the surface sites [136-137]. So, the Gutmann' and Drago-Weyland' parameters for different inorganic oxides and carbonaceous materials allow to predict the specific interaction contributions to the adsorption energies of organic compounds on solid surfaces. Also, these parameters may be useful for prediction of the reactivity of surface sites towards various reagents from gas phase. In contrast to descriptors of the molecular structure of organic compounds, which can be computed by quantum chemistry methods or taken from various quantitative relationships derived in

603 the processes, taking place in the solutions or gas phase, similar procedures for q u a n t u m chemical computation or evaluation of experimental descriptors for solid surfaces often run to various serious problems and uncertainties. The G u t m a n n ' and Drago-Weyland' parameters may be useful descriptors reflecting the reactivity of surface sites in various surface acid~ase reactions, including catalytic transformations of adsorbed compounds. For example, OH groups on the silica surface manifested as nucleophiles at interaction with most organosilicon compounds. This reaction causes the substitution of proton of this group by organosilicon residue. On the hypothesis that the reaction mechanism (SEi) is invariant in going from SiO2 to parent and mixed Ti and A1 oxides, one would expect decrease of chemisorption barrier at enhancement of the KB for oxides series. Actually, the chemisorption activation energy of an organosiloxane decreases as the KB of the metal oxide surface increases [83].

6

~

5 ZnO

9 ~O

4

"*=

3 -

~ 3 / S i O

2

.a 2 -

A1203

i

5

10

~'TiO2

I

I

I

15

20

25

*

30

35 -1

K a (kcal mol ) Figure 5. Ozone lifetime on the surface of inorganic oxides as a function of acceptor number for these oxides.

Second example is related to ozone lifetime on the metal oxide surfaces and it presents great interest for heterogeneous atmospheric chemistry. As ozone molecule possess weak basic properties [138], the decrease of ozone lifetime on the solid surfaces determined in [139] should be expected with the increase of the metal oxides acidity characterized by KA quantity. As it follows from Figure 5, the ozone lifetime on the parent A1, Zn, Mg, Ti oxide and alumosilicate surfaces decreases in going from basic ZnO and MgO to more acidic AleO3, TiO2 and A1203/SIO2 surfaces. This may be explained by adsorption of ozone molecule onto

604 stronger Lewis acid surface sites causing the ozone dissociation into free dioxygen and a surface oxygen atom, which remains attached to the metal cation in the oxide lattice. The acidYbase characteristics for other metal oxide surfaces which are absent in Table 7 may be estimated crudely from comparison the sequences for change of KA, KB, E and C parameters for series of metal oxides with order of change of the acid/base parameters determined in other scales or from quantitative relationships between these scales for a given series of solid surfaces. The different methods are proposed for the estimation of the order for relative acidity and basicity of inorganic oxide surfaces. The basicity of oxides increases at enhancement of negative charge on lattice oxygen atom in the following order [140] MgO > MgA1204 > A1203 > Zr02 ~ TiO2 Above basicity order coincides with that determined from increase of the positive charge on metal cations in the oxides lattice A1203 > ZrO2 ~ TiO2 Desorption energies determined from TPD experiments or differential adsorption heats of SO2 possessing acidic properties on the oxide surfaces may be chosen as a measure of their basicity MgO>CoO2 >ZrO2 >MgA1204 >A1203 >TiO2 The Lewis basicity of metal oxide sites decreases as [140] MgO > ZrO2 > TiO2 (rutile) > TiO2 (anatase) > MgA1204 > y- A120~ The temperature-programmed desorption of preadsorbed acidic carbon dioxide is frequently used to measure the number and strength of basic sites. The strength of basic sites determined from temperatures for maximum of the CO2 desorption peak in TPD spectra is in the increasing order of [141] BaO > SrO > CaO > MgO whereas the metal oxide basicity evaluated from differential adsorption heat distributions of C02 varies as [142] 7- A1203 > TiO2 (rutile)

Other different spectroscopic (XPS, UV adsorption and luminescence spectroscopies of the solid surfaces, IR spectroscopy of adsorbed CO2 and pyrrole)

605 and other methods (TPD of hydrogen, titration of solids by typical acid]base indicators, 1so exchange between carbon dioxide and surface oxygen) are proposed for characterization of the basic surface sites. They are reviewed in detail in [143]. Different techniques have been developed to determine the solid acidity and they are briefly described in recent review [144]. Among them the temperatureprogrammed desorption, microcalorimetry and IR spectroscopy of preadsorbed base molecules (NH3, CO, acetonitrile, pyridine, n-butylamine, quinoline, etc.) are widely used in determination the strength of Bronsted and Lewis acid surface sites. In accordance with these data, the acidity of Lewis sites on metal oxide surfaces varies as [140] y- A1203 > TiO2 > ZrO2 > CeO2 The acid/base characteristics of the metal oxide surface sites are changed at addition of minor concentration of extraneous ions. This may be a typical situation in the case of solid atmospheric aerosols having complex surface composition (Table 2). From data of microcalorimetry of base (NH3) and acid (SOD probes on 7-alumina, silica and magnesium surfaces at adding small amount (0.1 wt%) of ions (Ca 2§ Li § Nd 3§ Ni 2§ Zr 4§ SO~-) it was found that the modification of 7-A1203 surface properties changed moderately its amphoteric properties [145]. More substantial changes are observed on MgO which consisted in formation of sites of moderate and weak basic strength. The number of acid/base sites on doped SiO2 is strongly affected by the presence of introduced ions. The acidity of modified oxides increases at enhancement of generalized electronegativity of the metal ions, (Xi), expressed as X i = (l + 2Z)• X,,, where X0 is the electronegativity of neutral atom (Z = 0) given by Pauling and Z is the charge of metal ion [146]. This behavior is much more pronounced on silica series than on alumina series oxides. The cation electronegativity is a parameter determining the ionicity percentage of cation-oxygen bond in oxides. Then the ratio of the oxidation degree to the ionic radius of doping ions is also representative of the Lewis acid strength of cations. It is observed that acidity of modified oxides increases while the charge/radius ratio grows. The Lewis basicity of the modified oxides is associated with the electric charge of the oxygen adjacent to cation [147]. The average heat of SO2 adsorption per basic site increases as the partial oxygen charge associated with the cation grows. This tendency is much more evidenced on the doped silica samples which, because of very weak acidic character of this oxide, reflect much more the doping effect of inorganic ions. Most of organic compounds (alcohols, ethers, nitriles, amines, thio- and phosphor-derivatives, organometallic) are irreversibly adsorbed on most metal oxide surfaces at low temperatures after heat pretreatment of surfaces at moderate and high temperatures. Preliminary partially dehydroxylation of nonreactive cristobalite and pyrogenic silica surfaces at 1073 K also gives the

606 irreversible adsorption of methanol and tert-butanol on the surfaces at 303 K, and their initial differential heats of adsorption increase in comparison with that for the solids pretreated at 423 K [148]. The physical adsorption of gas phase reagent onto solid particle surface and formation of surface H- or donor/acceptor adducts is a first stage of its chemisorption and different surface reactions. The subsequent transformations of the adducts are affected by way of temperature or reagent concentration enhancement as well as by presence of third base or acid component (catalyst). Because different inorganic gases possessing high acid/base characteristics (SO2, HC1, C12, NH3) are adsorbed on the surface of atmospheric solid aerosols, the reactivity of the surface Bronsted and Lewis sites grows and formed H- or donor-acceptor complexes between these sites and organic pollutants are close to the transition state of the chemisorption. Various problems which arise in connection between chemisorption kinetics and adsorption equilibrium of organic compounds on metal oxide surfaces are discussed in [83]. The simple expression for apparent chemisorption activation energy (Ec~hem) with consideration of the H- or donor/acceptor complex formation prior to slow chemical reaction on the solid surface may be derived using the steadyconcentrations assumption for such a complex ~ (ESp + E nsp ) Eapp = E chem-

(42)

where Ec~hemis the activation energy of the chemical reaction between interacting gas and solid surface. Therefore, the apparent reaction barrier reduces with the increase of preliminary adsorption complex stability, and overall reaction rate grows. The overall adsorption energy can be estimated by means of methods discussed above. Moreover, the increase of preliminary adsorption complex stability gives increase of the reagent concentration on the surface and effectiveness of its following monomolecular processes or bimolecular LangmuirHinshelwood surface reactions is grower. Most descriptors of organic compound structures used in their relationships with adsorption energies are also available for quantitative description of chemisorption and surface reactions. However, because of cleavage and formation of new chemical bonds during chemical reactions and large differences between the structure of reagents in the initial and transition states, other descriptors, including the bond dissociation energies, relaxation energies of reagents in an transition state, are proposed for estimation of reaction barriers on metal oxide surfaces [24].

Q

C H E M I S O R P T I O N AND S U R F A C E REACTIONS OF ORGANIC C O M P O U N D S ON METAL OXIDES

The reactions of solid aerosols (NaC1, sulfuric acid-coated solid-propellant rocket motor-exhaust particles, a-A1203, ice, soot particles) with atmospheric

607 gases (NO2, N205, C1ONO2, CO2, SO2, HC1, O2, O3, H20) have been studied for many years [149-159]. However, little attention has been paid to the possibility of organic atmospheric species reacting directly with solid aerosols. These reactions are of interest as a possible means of disposing of stockpiles of organic compounds or products of their photochemical and thermal transformations in Earth atmosphere. Main mechanisms of organic compounds interaction with Bronsted and Lewis acid/base sites of metal oxide surfaces, stages of chemisorption reactions, kinetic equations taking into consideration the chemical and structural heterogeneities of the metal oxide surfaces and some quantitative "structure-reactivity" relationships (QSRRs) for these transformations have been discussed in our recent review [137]. However, most of experimental data considered in [137] are concern to the best investigated reactions of various organosilicon compounds with OH groups of parent and mixed Si, Ti and A1 oxide surfaces. Considered regularities for effects of substituents in these compounds, dissociation energy of bonds and structure of oxide surfaces on the chemisorption activation barrier can be extended to the chemisorption of organic pollutants on solid aerosol surface. But in this case the difference in the electronegativities of Si in the organosilicon compounds and C atoms in most organic pollutant molecules is to be taken into account, as well as the fact that last reactions are more complex and their kinetics is far less studied than that for former processes. Most of quantitative data for products of C, N, O, S, C1, F-containing organic compounds interacting with various metal oxides are derived from FTIR, and X-ray photoelectron spectroscopic, mass-spectrometric and chromatographic studies on the surface and in the gas phase in course of these transformations, whereas main body of kinetic results in this field is derived from temperature-programmed desorption experiments. We will make an attempt to consider here the main mechanisms of possible reactions of organic pollutants with active sites of metal oxide surfaces, experimental kinetic data derived from TPD measurements, kinetic equations, and possible QSRRs in these transformations. 5.1. M a i n m e c h a n i s m s o f c h e m i s o r p t i o n on t h e m e t a l o x i d e s u r f a c e s

The Bronsted acid (various types of OH groups) and base sites (oxygen atom in these groups and M-O bonds of lattice) as well as Lewis acid (metal cations M n§ surface free radical, e. g. = Si ~ and base sites (ones, similar to Bronsted base sites and O 2-, surface radicals - S i O ~ etc. ) are the main reaction sites of the inorganic oxide surfaces. The organic compounds (RX, RH, where R is an hydrocarbon moiety and X is a polar functional group) interact with these sites in accordance with following mechanisms [137, 143-144, 160]. 5.1.1. R e a c t i o n s w i t h B r o n s t e d a c i d / b a s e s i t e s

i. Electrophilic substitution of the surface OH group proton by an organic residue

(SEi)

608 MO-H + R-X --+ MO-R + H-X ii. Nucleophilic substitution of OH group by an organic residue (SNi). W h e n RX is an alcohol, the grafted alkoxy groups m a y be intermediates in its subsequent acid-catalyzed reactions M-OH

+

R-X -~ M-R

+

X-OH

iii. Dissociative electrophilic or nucleophilic addition to a metal-oxygen bond (AdEi or AdNi) M-O-M + R-X -~ M-OR + M-X iv. Proton transfer from acid site to organic compound (AHi). The formed onium cations m a y be i n t e r m e d i a t e s in subsequent acid-catalyzed reactions, e. g. dehydration of alcohols and hydration of olefins, etc. M-OH + R-X -+ Mn+O2- + [RXH] § v. Proton transfer from organic compound to base sites of oxides (EHi) and subsequent base-catalyzed reactions, e. g., double bond migration in olefines CH2=CH-CH2-CH2X + Mn+O 2- ---> [ C H 2 - C H - C H - C H 2 X ] - + Mn+O - H + --+ CH3-CH=CH-CH2X + Mn+O 2vi. Reactions of u n s a t u r a t e d compound with paired Bronsted acid sites M(OH)-OH+ >C=C< --+ >(HO)C-CH< + Mn+O 25.1.2. Reactions

with

Lewis

acid/base

sites

i. Redox reactions of Lewis sites with formation of organic radical-cations, radical-anions or addition of lattice oxigen to an organic compound R + Mn+O2- --> R X +" + M(n+l)+O2-

R + Mn+O2- - , R X " + Mn+O RH + Mn+O 2- --> RHO + M(n-1)+Dii. Dissociative adsorption on the Lewis acid/base pair with formation of Bronsted acid site (Adi) R-H + Mn+O2" -> R-M-OH Formed surface complexes and grafted organic groups undergo to subsequent monomolecular and bimolecular transformations. For example, n-complexes

609 formed between olefins or aromatic hydrocarbons and Lewis acid sites are turned to c-complexes. These complexes may be oxidized by Lewis acid surface sites up to products of their complete oxidation, carbon dioxide and water. Mutual surface diffusion of adsorbed organic free radicals gives their dimerization, polymerization or addition to adsorbed initial organic compounds or to reactive surface sites. It should be mentioned that surface of solid aerosols contains various metal cations, inorganic anions and inorganic salts. Because small size of aerosol particles, strong electric fields arise on their surfaces nearby charged sites. Electronic and spatial structure of organic molecules and products of their interfacial reactions: ions, free radicals, adsorbed on solid surface differ from the structure of these species in gas phase. As a result, the activation barriers of their reactions with other gases or adsorbed components will change and composition of reaction products will differ from those of similar processes in gas phase or in solutions. The charged surface sites induce orientation of reagent molecules. Conformation transitions being possible in molecules, the conformations with higher dipole moment are mainly formed on solid surface. Regioselectivity of the further reactions of adsorbed products is determined by their conformations on solid surface. The influence of electric field on the structure and reactivity of molecules and ions localized nearby solid surface, and the mechanism of interface reactions was exhibited theoretically and confirmed by some experiments on the surface of various heterogeneous oxide catalysts, especially in the zeolites cavity ([161] and references therein). These effects clearly can be manifested by the action of strong electric field with intensity varied from 0.005 to 0.06 a. u. Also, the intensity of non-uniform electric field near solid rough surface of ultrafine particle exceeds the value for flat surface, as F = U/(A r), where U is the outer potential difference and r is the radius of divided crystallites or roughness of the surface. The effects of intensity and vector direction of the electric field on activation barriers of unimolecular interfacial reactions of simple organic compounds were studied by semiempirical quantum chemistry method for orientation of molecules and radical-anions, heterogeneous electron transfer to adsorbed molecules, dissociation of radical-anions, cations and molecules, inversion of anions and molecules, cis-trans isomerization in olefines and chair-boat conformational transition of cyclic molecules in [161]. It was found t h a t electron and spatial structure of neutral particles and ions, height of activation barriers of their reorientation, dissociation, isomerization and conformation transition, the configuration stability to change as affected by the field in the range 0.01 + 0.05 a. u. These effects are to be taken into account at explanation of the surface reaction kinetics and the products composition on the surface of atmospheric solid aerosols. As is evident from Table 2, the surface of most solid atmospheric aerosols contains marked amount of such acidic components, as adsorbed gases NO2, N205, C1ONO2, SO2, and HC1, and their corresponding acids or salts. It is known that sulfate t r e a t m e n t of oxides such as Fe203, A1202, SnO2, TiO2, ZrO2, etc.,

610 gives high increase in the surface acidity and in the catalytic activity, for example, in dehydration of alcohols, olefins hydration, esterification, acylation and isomerization reactions [162]. These solids were claimed to present superacid sites with acid strengths up to Ho < - 16.04. These sulfated metal oxides usually are prepared from their hydroxides by treatment with H2SO4, SO2, SO3 H2S or (NH4)2304 and following calcination [163]. The nature of the high acidity of the sulfated metal oxides is still controversial. It was proposed that the very strong acidity is due to the increase in number and strength of Lewis acid sites [164]. More realistic structure for active site implies the formation of Bronsted acid sites which would allow catalytic reactions to occur at much lower temperatures with the same rate as they occur at higher temperatures without the presence of the Bronsted sites [165]. Also, the addition of chlorine to alumina enhances the activity of solid for skeletal transformations of hydrocarbons [166]. The role of halogen promoter at the surface is to enhance the Lewis or Bronsted acid environments, respectively, through an electroattractive effect of the halogen atom from an adjacent OH group [167]. Quantum chemical studies indicate that the stabilization of LUMO energy occurs with a narrowing of the energy gap, by the substitution of chlorine for oxygen or OH groups in T-alumina cluster models [168]. The Bronsted acid character of the cluster also increases after substitution. Similar enhancement of surface acidity and reactivity in the chemisorption of various organic bases has been observed after substitution of OH groups on silica surfaces by chlorine [169]. Therefore, the oxide surfaces of solid aerosols in the presence of acid additions can exhibit much more high acidity and catalytic activity in the reactions of adsorbed organic compounds than their one-off surface acidity and reactivity suggests. 5.2. K i n e t i c s of c h e m i s o r p t i o n a n d s u r f a c e r e a c t i o n s The rate of adsorption of gas-phase species, Va can be written as [170]

Va : s r ( o ) x F

(43)

where F is the flux of gas reagent and Sr(O) is the sticking coefficient. The superscript r is 1 for non-dissociative and 2 for dissociative adsorption (chemisorption). For direct adsorption (chemisorption) from gas phase, s r ( o ) = sr (0)x (1- O)r

(44)

where Sr(0) is the sticking coefficient at zero coverage O. In accordance with Arrhenius equation

(Ea)

sr(o)= S~ xexp - ~ -

(45)

611 where

S~ and Ea are the pre-exponential factor and activation energy for

adsorption. The normal pre-exponential factors for these reactions according to e s t i m a t e d from transition state theory are varied from 10 -lo to 1 0 1 7 c m 3 s -1 at r = 1 and 2 [171]. The overall non-dissociative adsorption (chemisorption) rate of gas-phase species onto solid aerosol surface (Va(over)) c a n be described by using the simple model for discrete heterogeneous surface [172]

Va(over )

=

F x S Ox

Ea(i) Oti exp - Ea(i) kT x exp - S O x x x exp - kT

i•1 .

(46)

where Oti i s the relative concentration of i-th subsurface, Ea(i) is the chemisorption activation energy on i-th subsurface and x is the reaction time. Because, in the case of dissociative and associative surface reactions, the consideration m u s t be given to correlation effects between E(i) on different sites, the later kinetic equations will be given only for surface reactions t a k i n g place on uniform surface. W h e n chemisorption is a s s u m e d to proceed via the precursor state (H- or donoracceptor surface complexes) discussed above, then rates of non-dissociative and dissociative chemisorption can be written as [173] V a = qxFA x k a ( 1 - O A ) kd* + k a*(1 - | A ) Va

=

g•

• ka(1-|

(47)

(48)

k~ + k a ( 1 - O A ) 2 where ~ is the t r a p p i n g probability from the gas phase into the precursor state, k a is the rate constant for transition from the precursor state into the chemisorbed state, and k d is the rate constant for desorption of the reagent from precursor state. The rate of desorption for a randomly distributed adsorbate can be written as

r@r Vd = kd

(49)

where k~ is the r a t e coefficient for desorption, obeying to Arrhenius equation r

kd = ~d xexp -~-~

(50)

612 where v~ and Ed are the pre-exponential factor and the activation energy for desorption, respectively. The desorption pre-exponential factors range from 1013 to 1019 s 1 for r = 1 and from 1011 to 1019 s -1 for r = 2. Surface bimolecular reactions are classified into L a n g m u i r - H i n s h e l w o o d a n d Eley-Rideal processes. The first includes reactions between two adsorbed species (A and B) or an adsorbed species and a vacant site (V) V 1 - k l A B , or V! = k 1 A V

(51)

where kl is the rate coefficient, 0i is the surface coverage of species i a n d 0v is the fraction of v a c a n t sites. The kl value is obeyed to A r r h e n i u s equation k 1 = v 1 x exp - ~

(52)

The pre-exponential factor for this reaction varies from 10 -n to 104 cm 2 s -1. The second type of the reactions includes the direct interaction of gas-phase species with adsorbed species and the reaction rate can be w r i t t e n as Ve = ke0APB

(53)

where pB is the partial pressure of r e a g e n t B. The ke value can be r e p r e s e n t e d t h r o u g h t reactive sticking coefficient, So

/ Ee)

S~ )1/2 exp k e - (2~mBk T

(54)

where as is the area per reaction site, and mB is the molecular weight of species B. The pre-exponential factor for these reaction varies from 10 .6 to 10 -17 cm 3 s -1.

5.3. Quantitative "Structure-Reactivity" relationships in surface reactions of organic compounds The Evans-Polanyi relation between the activation b a r r i e r (Ea) and the reaction e n t h a l p y (AH) is most often used for description of reaction kinetics on the solid surface Ea = (~ AHr + ~

(55)

where (z is the t r a n s f e r coefficient (0 _< a < 1) and ~ is close to the intrinsic reaction barrier. As is shown [174], this relation may be useful for prediction of Ea values for chemisorption of some organic compounds on hydroxylated silica surface. Also, the bond-order-conservation-Morse-potential (BOC-MP) method [175] t a k i n g into account reaction enthalpy, adsorption h e a t s and bond dissociation energies in

613 reagents and products may be used for prediction of activation barriers for various surface reactions. The Ea value for dissociative adsorption from the gas phase can be written as QAQB E a =0.5 I DAB-(QA +QB)+ QA +QB _QABI

(56)

where QA, QB and QAB are the adsorption heats of species A, B and AB, respectively. For non-dissociative desorption, Ed is given by Ed = QA or Ed = QAB

(57)

whereas for associative desorption, Ed is QAQB +AH) E d = 0.5 QA + QB

(58)

where the enthalpy change for the surface reaction is given by AH = DAB + QAB- QA- QB

(59)

In the case of disproportionation reaction A + BC -~ AB + C, the activation barrier is Ed=0"5/QABQABQC + + Q c AH)

(60)

where the enthalpy change in the reaction is AH = DBC- DAB- QAB + QA- Qc + QBC

(6~)

However, the attempts to estimate the activation energies for the interaction of organosilanes with hydroxylated silica surface using Q values computed by quantum chemistry (Eq. 60) lead to Ea values rather than on 50 +: 100 kJ mol 1 higher than experimental values. Above expressions apparently are suited for surface reactions, which occur through only nonpolar or weakly polar transition state. The static quantum indexes of reactivity may be used for description of the double exchange reactions, which occur through the transition state with high charge separation. If barrier height is controlled by frontier orbital interactions (interaction of a soft acid with a soft base) then the following relationships may be written for SEi reaction Ea = al/(EHOMO(1) - ELUMO(2))+ a2

(62)

614 where a l and ae are coefficients for given series of organic reagents. For the reactions discussed compound 1 is the MOH group of inorganic oxide surface and compound 2 is an organic compound. Actually, the most relationships are observed between the reactivity of organosilicon compounds toward OH groups of hydroxylated silica surface and ELUMO (electron affinity (EA) with an opposite sign) or p a r a m e t e r s which include the EA value for these compounds [137]. For example, the following tendency is observed between activation energy of organosilicon compounds interaction with OH groups of hydroxylated silica surface corrected for preliminary adsorption complex formation, and their electron affinity (EARsix) Ea = - (1.5 + 0.9) 105/(IPsioH - EARsix) + 251 + 84 kJ tool -1

(63)

where IPsioH is the ionization potential of silica surface cluster of OH group (IPsioH = 9.4 eV). As the Ea value is controlled by charge transfer, the choice between SNi and SEi mechanisms can be done by comparison of reagents electronegativity (~ = IP + EA). If inequality ~M-OH> ~RX is satisfied, t h a n the organic compound is a nucleophil, but at ;~M-OH< ~RX, this compound is an electrophil. With allowance for these inequalities and the ionization potential (IP) and EA values of organosilicon compounds, and C, B, Ti, Ge, Sn halides ()~RX = 11 - 14 eV), and clusters of OH groups of silica surface ()~M-OH= 8.8 + 11.7 eV), it follows t h a t these compounds are electrophils in the reactions with OH groups. The ~RX values for ammonia, aliphatic amines, aliphatic alcohol's, phenol and water fall within the range 5.9 + 8.6 eV. According to the first inequality these compounds are nucleophiles in substitution reactions with the SiOH groups. The crude estimation for ability of inorganic oxide to interact with an organic compound as acid or base can be found also from comparison of mean electronegativities of these solids (Table 8) and gas-phase reagents. It was found t h a t the activation energy of organosilanes ((CH3)3SiX) interaction with silica OH groups decreases with lowering the Si-X bond energy (Esi-x) and with enhancement of the inductive coefficient of the substituent X (~I) Ea = 0.29 Esi-x - 274.5 ~i + 62.9 kJ mo1-1

(64)

Then above conclusion about SEi mechanism for the reactions agreed with sign of coefficients in Eq. 64. The main p a r a m e t e r s determining the organic compounds ability to interact with OH groups of inorganic oxide surfaces are bond dissociation energies of reagents and products as well as the energies of frontier orbitals of the reagents. So, simplified Marcus-like equation for the activation energy of the SEi reaction at varying the substituent X in (CH3)3SiX compounds was found [24] Ea = 0.08 (IPHx - EARsix) + 0.025 (DRsi-x - DHX) - 24.7 kJ mo1-1

(65)

615 where DRSi-X and DHX a r e the Si-X and H-X bond dissociation energies in the gaseous reagent and product. First term in this equation takes into account the relative redistribution of electron density between gas-phase species and surface groups for series of species. Several linear relationships for change of Ea in SEi, SNi and AdEi reactions by varying the structure of organic reagents RX and sites of inorganic oxide surfaces (M-OH, M-O-M and M-Y) can be derived using simple thermodynamic cycle and taking into account the difference in the electronegativities of reagents and products [24]. These equations present the combination of kinetic (first term in right side of Eq. 65) and thermodynamic (second term in this Eq.) contributions to the height of the activation barrier for surface reactions. For example, variation of substituents R in RX changes the activation energy for SEi reaction as follows AEa =- a3(EARx + EAROM) + a4(DR-x - DR-OM)

(66)

The change in influence of metal M on the reactivity of M-OH group varies as AEa = a5(IPMoH - EARoM) + a6(DMO-H - DR-OM)

(67)

In the case of interaction RX compound through SNi mechanism with surface MOH group the activation energy changes at varying the substituent X as AEa = aT(IPRx - IPMR) + as(DR-x - DM-R)

(68)

and effect of metal M on the reactivity of M-OH group in this reaction leads to following variations of activation energy AEa = ag(IPRx - EAMoH) + al0(DM-oH - DR-X)

(69)

The reactivity of oxide surfaces grows after replacement of surface OH by more active Y group, where Y = C1, SOn, NH2, NR'R", etc. The change of activation energy for next SNi reaction at varying of this group Y is AEa = al I(EAMY + EAHY) +

a12(DM_Y -

DHY)

(70)

As it seen, the reactivity of surface groups M-Y is determined by electron acceptor ability of substituent Y and difference in strengths of M-Y and H-Y bonds. These factors may be reasons for weak reactivity of Si-F groups on the fluorinated silica surface towards typical organic nucleophiles (alcohols, amines) in comparison with more reactive Si-C1 groups of chlorinated silica. Difference between the activation energies of SEi and AdEi routes at varying substituent X in RX is

616 AEa = a13(IPMx - EARx) + a14(DR-x - DM-X)

(71)

As it follows from [174], the AEa values for AdEi route to the Si-O bond of SiO2 surface at the interaction with (CH~)3SiX (X = N3, NCS and NCO compounds, are 10 + 20 kJ mol 1 lower than the activation energies of SEi reaction of these reagents. From comparison of relations (65) and (71) it is obvious that the maximum AEa value can be observed in the case of lowest reactivity in the SEi reaction. This explains the observed maximum of AEa at the interactions (CHs)sSiNCO and (CHs)sSiN3 as compared with (CH3)3SiNCS. Because most of inorganic oxides-constituents of industrial solid aerosols are active catalysts in various processes of partial or complete catalytic oxidation of organic and inorganic compounds, the activation energy of these transformations is related to bond energy of oxygen in these oxides (Ebo) in accordance with known relationship [176] Ec - a15 Ebo + a16

(72)

As it follows from comparison of oxygen bond energies and activation energies of methane, carbon oxide and hydrogen complete catalytic oxidation, the catalytic activity of metal oxides in these transformations decreases in the following order Co304 > MnO2> NiO > CuO > Cr203> Fe2Os> ZnO > V20~> MgO > A12Os>> SiO2

The expressions presented here may be used as a basis for possible QSRRs between change of the activation barriers for reactions on the solid aerosol surfaces and descriptors of surface sites as well as gas-phase organic species.

5.4. C h e m i s o r p t i o n of o r g a n i c c o m p o u n d s on the i n o r g a n i c o x i d e s u r f a c e s and t h e r m a l d e c o m p o s i t i o n of the surface s p e c i e s Experimental aspects of temperature-programmed desorption with mass spectrometric registration of volatile reaction products as a tool of investigation of disperse oxide surface and reaction mechanisms has been discussed elsewhere [176-180]. Let us consider some experimental results for chemisorption of organic compounds on the inorganic oxide surfaces and for thermal decomposition of surface species. The type of oxide, chemisorbed substance, desorption product, heating rate (~) and temperatures of peaks maxima (Tin) in TPD spectra are presented in Table 9. The position and relative intensity of the peaks depend on the heating rate. In order to compare these spectra, desorption activation energies (Ed) (given in last column in Table 9) have been calculated using Tm and parameters by simple Redhead's peak maximum method [181] at normal preexponental desorption factor Vd = 1013 s 1.

617 Table 9 Products of organic compounds decomposition preadsorbed or bound to inorganic oxide surfaces, temperatures of peak maximum in their TPD spectrum and the apparent desorption energies estimated using the Readhead's method Product of ~(K s -1) decomposition

Substance

Oxide

CH3OH

SiO2 [186] 0.1

n-CnHgOH SiO2 [187] 0.067 C6Hs(CH2)2OH SiO2 [188] 0.1

CH3OH

TiO2 [185] 0.2

C2HsOH

0.2

n-C3H7OH

0.2

i-C3H7OH

TiOe[185]

0.2

CH3OH C=O H2C=O CH3OCH3 CH3C(O)H CH4 CH2CHCH2CH3 C6Hs(CH2)2OH C6HsCH=CH2 C6H6 C6H5 C6H5 C6HsCH3 CH3OH CH3OCH3 H2C=O C=O H2 CH4 C2HsOH C2HsOC2H5 CH3C(O)H H2 C4Hs CO2 n-C~HTOH C3H7OC3H7 C3H7C(O)H C~H6 H2 C=O i-C3H70H CH3COCH3 C3H6 He C=O COe

Tm (K)

Desorption energies (kJ tool -1)

763 723 473, 863* 323, 583* 323, 600* 923 823 573 743 823 823 573 390,645* 635 675 675 675 675 390, 590* 590 600 630 640 710 390, 575* 590 620 625 600 620 390, 490* 540 560 550 560 710

224 212 137, 254** 93, 170"* 93, 175"* 272 245 167 218 242 242 167 110, 185"* 182 194 194 194 194 110, 169"* 169 172 180 183 204 110, 164"* 169 177 179 172 177 110, 139"* 154 160 157 160 204

618 Table 9 Products of organic compounds decomposition preadsorbed or bound to inorganic oxide surfaces, temperatures of peak maximum in their TPD spectrum and the apparent desorption energies estimated using the Readhead's method Product of ~(K s -1) decomposition

Substance

Oxide

HCOOH HCOOH

ZnO [189] 5 ZnO [189] 5

C6HsOH

ZnO [136] 10 FeO[136] 10 SiO2 [191] 0.1

(CH3)eCHNHe

Fe203

0.1

MgO

0.1

CaO

0.1

A1203

0.1

cyclo-C6H11NH2 A1203[191] 0.1

Fe203[191]

0.I

HCOOH CO2 C=O H2 C6HsOH C6HsOH (CH3)2CHNH2 CH3CH=CH2 CH3CN CH4 CH3CN CH4 (CH3)eCHNH2 CH3CH=CH2 NH3 CH3COCH3 COz, H20 (CH3)2CHNH2 CH3CN CH4 (CH3)zCHNH2 HCN (CH3)2CHNH2 CH3COCH3 CH3CN CH4 C6H11NH2 NH3 C6H12 C6HsNH2 1,3-C6H12 C6HIINH2 NH3 c6H12 C6H10=O CH3CH=CH2 C6HsNH2

Desorption energies (kJ mol 1) 180 45 550 142 550 142 500 129 520 132 556 141 403,703* 116,206"* 803 236 843 248 843 248 698 204 698 204 423 122 523 152 523 152 523 152 638, 708* 186, 207** 413 119 548 159 548 159 413 119 883 260 433 126 588 171 548 159 548 159 433 125 598 174 598 174 673 197 653 191 423 122 513 149 513 149 523 152 513 149 533 155

Tm (K)

619 Table 9 Products of organic compounds decomposition preadsorbed or bound to inorganic oxide surfaces, temperatures of peak maximum in their TPD spectrum and the apparent desorption energies estimated using the Readhead's method Product of Substance

Oxide

[~(K s -1)

decomposition

Tm (K)

Desorption energies (kJ mo1-1)

718 210 COz HzO 533,698* 155,204"* C6HsNHz SiOz [192] 0.1 C6H5NH2 413,638" 119,186"* 443,603* 128,176"* A1203 0.1 C6H5NH2 423 122 CaO 0.1 C6H5NH2 423 122 MgO 0.1 C6H5NH2 433,553* 125,161"* Fe203 0.1 C6HsNHz 718 210 CO2 HzO 693 203 403,703* 116,206"* (CH3)3CNH2 SiO2 0.1 (CH3)3CNH2 843 248 (CH3)2CH=CHe NH3 843 248 443 128 A1203 0.1 (CH~)3CNH2 598 174 (CH3)2CH=CH2 NH3 598 174 Fe203 0.1 (CH3)~CNH2 433 125 573 167 (CH3)zCH=CH2 NH3 573 167 638,688* 186,201"* CO2 H20 638,688* 186,201"* CC14 AlzO3[197] 5 CC1 155 39 300 76 CO2 CFC13 5 CC1 130 32 130 32 CFzCI2 5 CC1 360 92 CO2 CF~C1 5 CC1 120 30 390 100 CO2 *Temperature of the second peak maximum in the TPD MS spectrum of this product. **Apparent desorption energy related to the second peak in the TPD MS spectrum of this product.

5.4.1. Oxygen-containing compounds The decomposition of alcohols has been widely used as the probe of the acid~ase properties of metal oxide catalysts [182]. While it is generally assumed that acidic oxides catalyze dehydration and basic oxides catalyze

620 dehydrogenation, recent studies have shown that the selectivity of alcohol decomposition on metal oxides is a function of reagent and surface structures. Oxides which act as solid bases to abstract protons from alcohols can also produce net dehydration products via alkoxide decomposition [183]. It was found that the decomposition kinetics and selectivity of alkoxides on the TiO2 (anatase) surface are dependent upon the alcohol structure [184]. The decomposition temperatures of surface alkoxides are in the order MeO > EtO > n-PrO > i-PrO. The desorption sequence for the decomposition products from ethanol and l-propanol TPD exhibited a common pattern: dialkyl ether (bimolecular interaction) - aldehyde (a-H abstraction) - olefin (~-H abstraction and oxygen deposition). For methanol TPD, aldehyde was followed by methane due to the absence of ~-H. The absence of dialkyl ether formation from 2-propanol TPD again suggested steric constraints on secondary alcohol adsorption and decomposition. The dehydration selectivity increased in the order MeO < EtO < n-PrO < i-PrO. The TPD MS and IR studies on TiO2 (rutile) single-crystal surfaces have suggested that product distributions in the reactions of primary alcohols and carboxylic acids are governed primarily by the coordination environment of individual surface cations. This model would suggest that the reactivity of adsorbed alcohols should be insensitive to bulk structure. In order to test this hypothesis, methanol, ethanol, and 2-propanol were adsorbed at room temperature on anatase and rutile powders [185]. The alcohols were dissociatively adsorbed to form alkoxides and surface hydroxyls. The alkoxide species were removed via two channels, recombination with surface OH groups at 400 K and decomposition at higher temperatures. Dehydration and dehydrogenation pathways are observed for all of the alcohols, with only the primary alcohols yielding bimolecular reaction products. The similarities in product distribution and peak temperatures from aliphatic alcohols on anatase and rutile, particularly with regard to the selectivity for diethyl ether formation from ethanol, indicate that the bulk crystal structure of the oxide does not have a significant influence on the reactions of these molecules. The products of methanol chemisorption on SiO2 surface pretreated at 1020 K were studied by TPD MS method in [186]. Dimethyl ether, acetaldehyde and carbon oxide were observed in the temperature range usual for the effective alcohol chemisorption on the silica surface (~650 K). Their formation may be explained by following reactions -SiOCH~ + CH3OH --+ -SiOH + CH3OCH3 2 -SiOCH3 -+ -SiOH + - S i l l + CH3CHO Thermal decomposition of the grafted methoxy groups in the range from 820 to 1070 K gives formaldehyde, methane, hydrogen and carbon oxide in accordance with supposed reactions

621 -SiOCH3-->-Sill + H2C=O 2 -SiOCH3 --> -SiOH + -Sill + CH4 + CO -SiOCH3 + H -~ -SiO + CH4 H2C=O ~ CO + H2 The decomposition reactions of alkoxide groups -SiO(CH2)3CH3 on dispersed silica surface were examined by TPD MS and quantum chemistry methods in [187]. Main desorption channel corresponds to the elimination of l-butene above 600 K. The TPD data suggest that this reaction is of first order and that it can be viewed as "unimolecular". Chemical transformations of phenylethanol bound to silica gel surface were studied by TPD MS method after the previous partial carbonization of the grafted groups [188]. The main decomposition products of the groups are C6H5(CH2)2OH, C6HsCH=CH2, C8H6, C6HsC6H5 and C6H5CH3. The formation of phenylethylene is due to the unimolecular decomposition of grafted groups with H transfer from CH2 group (nearest to aromatic ring) to O from SiOR group. The formation of benzene and biphenyl was explained by surface migration of intermediate phenyl radicals and H abstraction from surface groups and associative desorption, respectively. The adsorption and reactions of formic acid on the oxygen-terminated ZnO(001)-O surface have been studied by TPD MS and XPS method in [189]. Small amounts of formic acid dissociate at defect surface sites to yield surface formate (HCOO). The surface HCOO decomposes to yield nearly simultaneous CO2 (37 %), CO (63 %) and H2 TPD peaks at 560 K.

5.4.2. Nitrogen-containing compounds The chemical transformations of various amines adsorbed at oxide/vapour interfaces (SIO2, A1203, Fe203, MgO and CaO) were studied by IR spectroscopy and TPD MS methods in [190-192]. It was found that with primary aliphatic amines the main high-temperature reaction on oxides possessing Lewis-acid sites is nitrile formation, but secondary aliphatic amines additionally show CN bond breakages causing desorption of NH3 and propylene (isopropylamine), cyclohexene (cyclohexylamine) as dehydrogenation (cyclohexylamine -~ aniline; isopropylamine ~ adsorbed imine species), and C-C bond breakage (isopropylamine: adsorbed imine species --> desorption of CH4 and acetonitrile). At beam temperature surface complexes formed between aniline and t-butylamine on the oxide surfaces are similar to those in the case of other amines: there is hydrogen bonding and dissociative adsorption on SiOe, and formation of coordination bonds between amine molecules and Lewis sites on the other oxides. With increasing temperature, in the case of aniline only aniline itself desorbs, whereas when t-butylamine is used, in addition to the unchanged amine, isobutene and NH3 are detected as desorption products, indicating the occurrence of CN bond breakage. With amines examined oxidation reactions take place on the surface of Fee03.

622

5.4.3. A r o m a t i c h y d r o c a r b o n s Spectroscopic data about structure of surface complexes formed after adsorption of aromatic hydrocarbons on the metal oxide surfaces are absent. The detailed scheme of toluene transformations on oxide surfaces was proposed in [193] on the basis of reaction products of their partial and complete catalytic oxidation C6H5CH3 + 02 --->C6H5CHO + 02 ~ C6H5COOH --> C6H6 + CO2 C6H6 + H20 --->C6H50H + 02 ~ 0=C6H4-0 + 02 ~ C4H203 + 02 --> CO, CO2, H20 Also, benzophenone is formed in the oxidation transformations with participating two toluene molecules. 5.4.4. H a l o g e n - c o n t a i n i n g c o m p o u n d s The reactions of several atmospherically relevant halomethane compounds (CF3C1, CF2C12, CFC13, and CC14) with heat-treated 7-alumina powders at 1000 K have been investigated in situ FTIR spectroscopy in an attempt to assess the impact of alumina exhaust particles from solid-propellant rocket motors on stratospheric chemistry [194]. The powders were dosed at 100 K with halomethanes and then gradually heated to promote reaction; infrared spectra were recorded as a function of temperature. The spectral features, which appear at temperatures as low as 120 K, are attributed to adsorbed carbonate, bicarbonate, and/or formate species (COn, n = 2, 3), indicating that halomethanes dissociatively chemisorbs on heat-treated 7-alumina at temperatures below that of the lower midaltitude stratosphere. These COn species are stable in the range 250 - 400 K, depending on the degree of chlorination of the halomethane. The decomposition ~eactions apparently proceed through C-X (X = F, C1) bond rupture and AI-X and C-O bond formation and involve the participation of surface active sites such as A13§ ions and AlxOy clusters that are produced during dehydroxylation. The IR adsorption experiments provided no information about absorption features associated with AI-X or O-X bonds, and reaction products that are released into gas phase. Such information is crucial, however, to assessing the impact of surface-mediated decomposition of halomethanes on solid-propellant rocket motor-alumina particles on the stratospheric ozone cycle. The XPS studies suggested that under certain conditions, halogen-containing species are evolved from 7-alumina surfaces exposed to halomethane compounds at stratospheric temperatures [195, 196]. The reactions of CF3C1, CF2C12, CFC13, and CC14 with heat-treated y-alumina powders were studied using TPD MS and XPS methods in [197]. Desorbing species were monitored as a function of substrate temperature using a line-of-sight quadrupole mass spectrometer. Hydrogen chloride and halomethyl fragments, which are indicative of halomethane dissociative chemisorption were observed to desorb below 150 K. Carbon dioxide began to desorb between 240 and 320 K. The CO2 most likely arises from COn (carbonate and/or formate) species which are formed via the low-temperature dissociative

623 chemisorption of the halomethanes. In situ XPS analysis of heat-treated powders t h a t had been dosed at 150 K with halomethanes revealed the presence of both organic and inorganic forms of fluorine. Halogen uptake probabilities, which are estimated to be - 10 .5 from the data, increased as the degree of chlorination of the halomethane increased. These results indicate that halomethanes will probably decompose on solid-propellant rocket motor-alumina particles in the stratosphere, forming adsorbed A1-X (X = C1, F) and COn species and releasing gas phase HC1 and CFxCly fragments. However, the impact of these processes on global stratospheric halomethane and ozone concentrations is likely to be minimal. The localized depletion of halomethanes may occur in the vicinity of the exhaust plume of a booster rocket where particle loading is much larger. The adsorption and thermal decomposition of the polyperfluorinated ether (C2F5)20 on 7-A1203 has been studied by IR spectroscopy in [198]. This ether interacts with the isolated surface OH groups, forming an increasing number of associated OH groups from 150 to 600 K. Surface fluoroacetate and surface fluoroformate species are also formed from the thermal decomposition of the (C2F5)20 layer. At T > 300 K, the surface fluoroacetate converts to surface fluoroformate in accordance with following scheme (C2Fs)20~ds (150-200 K) -~ CF3COOads + FCOOads + HF + A1-F + OH (assoc) (C2F5)2Oads (200-300 K) -~ FCOOads + HF + AI-F + OH (assoc) CF3COOads (300-600 K) ~ FCOOads Oxidation of the ether also occurs on surface preadsorbed with pyridine, indicating t h a t Lewis acid A13+ sites, blocked by pyridine adsorption, are not involved in fluoroacetate and fluoroformate formation.

5.4.5. Organometallics XPS, FTIR spectroscopy and TPD MS methods were used to study the chemisorption and decomposition of trimethylaluminium (TMA) on silica under high vacuum [199]. By annealing series of silicas from 425 to 1573 K prior to TMS exposures at 300 K, the distribution of chemisorption products as a function of the relative concentration of various OH groups types was examined. It was proposed t h a t the monomethylaluminium surface complex and methyl groups bonded to silicon are the majority species on the surface at 300 K. Decomposition of the monomethylaluminium complex begins above 373 K and increases the population of methyl groups bonded to silicon on the surface. The methyl groups react to form methane, ethane and adsorbed hydrocarbon fragments. In addition, methyl groups also react further with the surface to form tetramethylsilane. Small amounts of gas-phase TMA and carbon-contaminated aluminosilicate surface are observed. The mentioned above organic compounds, products of thermal decomposition of the species preadsorbed or bound to inorganic oxide surfaces, temperatures of

624 peak maximum in their TPD spectrum and the apparent desorption energies estimated using the Readhead's method are listed in Table 9. From comparison of Eo data for products of halomethanes chemisorption on y-A12Oz in Table 9 it follows that Eo value for CO2 desorption from A12Oz surface increases as the C-C1 bond dissociation energy in the chemisorbed halomethane grows. This dependence is presented in Fig. 6. Also, the Ed value for desorption of primary aliphatic and aromatic amines from the AlaOz and Fe2Oz surfaces increases with reducing the ionization potential of the amine (Figure 7). This observation agrees with proposed mechanism of the amines adsorption on the oxide surface, including the adsorbate interaction with Lewis acid sites of the oxide surfaces. The amine desorption energy decreases as the acceptor number of oxide surface (KA from Table 7) reduces at transition from A12Oa to SiO2.

Ea (kJ mo1-1) 110

Ed (kJ mo1-1) 130 A1203 O

100 CF2C12 90

80

125

[ Cfl4o

70 300

319

338

357

Ec_cl (kJ mol-l) Figure 6. Plot of desorption energy of carbon dioxide from alumina surface vs C-C1 bond dissosiation energy in halometanes.

120 7

7.5

8

8.5

9 IP (eV)

Figure 7. Desorption energy of aromatic and aliphatic primary amines from oxide surface as a function of amine ionization potential.

The quantum chemical computations of energies for surface reactions taking place on the model clusters of active sites and grafted groups on inorganic oxide surfaces may be useful tool for prediction of activation energies for the examined decomposition reactions. The heat of unimolecular decomposition of n-butyloxy group grafted to the silica surface yielding to the surface OH group and 1-butene (260 kJ moll), computed by AM1 method is close to desorption activation energy

625 for this reaction from Table 9 (245 kJ mol-1). Also, the desorption activation energy of various gas-phase species in thermal decomposition of grafted phenylethoxy groups on the silica surface from Table 9 increases as the enthalpy of the reactions, computed by this method grows [188]. This agrees with expressions 57 and 58, derived for non-dissociative and associative desorption by using BOC-MP method. 5.5. P h o t o c h e m i c a l t r a n s f o r m a t i o n s w i t h p a r t i c i p a t i n g t h e o r g a n i c s on solid a e r o s o l s u r f a c e s

The photochemical reactions of organic pollutants adsorbed on solid aerosol surface with various small atmospheric species, as gases and free radicals in their ground or excited state are of great importance for heterogeneous atmospheric chemistry. The solid surfaces are classified into two categories: (i) nonreactive surfaces, such as silica or alumina which provide an ordered two-dimensional environment for effecting and controlling photochemical processes more efficiently than can be attained in homogeneous phase; (ii) reactive surfaces, such as titania or metal chalcogenides, which directly participate in photochemical reactions by absorbing the incident photon and transferring charge to an adsorbed molecule or by quenching the excited state of this molecule [200]. It is known that aromatic compounds adsorbed on nonreactive solids exhibit a longlived excited state which is beneficial in enhancing the efficiency of such photochemical reactions, as .their photoionization [201], bimolecular electron transfer between adsorbed molecules [202], hydrogen abstraction, etc. The dispersed quartz surface covered by aromatic compounds, e. g. anthracene or organic dyes, offers a scavenger of singlete oxygen 102 (lAg) formed after illumination of the surface by visible light during 2 - 10 minutes at low temperature (273 K or low) [203]. The desorption activation energy of the 102 species from such surfaces varies from 90 to 130 kJ mo1-1. The freezing out of the adsorbent at temperatures from 273 to 323 K gives release of the 102 species into gas phase. Because most of industrial solid aerosols consists of inorganic nuclei, covered by layer of polycyclic aromatic hydrocarbons mixture, the observed effect of the singlete oxygen conservation by dispersed particles may occur in the atmospheric heterogeneous processes, e. g. in depletion of the ozone layer. The presence of reactive subsurfaces is possible on the aerosol surfaces consisting such semiconductor nanoparticles as TiO2, ZnO, ZnS, CdS and other metal chalcogenides. Their role in initiating and controlling such surface transformations of organic compounds as oxidation of olefins, arenes, alkanes, amines, alcohols and the degradation of chlorinated organics, phenols, and haloaromatics has been reviewed recently [204-206]. The reactive oxygen species, such as free radicals HO2, RO2, OH, may be formed on the illuminated semiconductor surfaces in the presence of air via dioxygen reduction by a conduction-band electron in the presence of a suitable adsorbed species possessing electron donor properties. Also, by virtue of producing these radicals via photochemical processes and high concentration of reactive ozone in the

626 stratosphere, they may interact in ground or excited state with organics adsorbed on the solid aerosols via Eley-Rideal or Langmuir-Hinshelwood mechanism. Since ozone transformations have great importance in environmental atmospheric chemistry, let us consider its possible reactions on the solid surfaces. Despite the great importance in environmental chemistry, the ozone adsorption has been studied by IR spectroscopy at low temperature (70 K) on the silica, titania and alumina surfaces only in recent years [138,157,207]. Ozone molecules are shown to form weak hydrogen bonds with OH groups of silica and titania surfaces. Adsorbed ozone reveals comparatively high basicity, close to t h a t of CO and is bound to silica OH group r a t h e r via one of terminal oxygen atoms, t h a n via the central one. The isotope substitution experiments provide evidence for strong deformation of ozone molecules adsorbed on the titania surface, which are bound to titanium ions also via one of terminal oxygen atoms. Although alumina is more acidic t h a n TiO2, IR spectroscopy did not detect 03 species adsorbed on strong Lewis A13+ sites. However, 03 decomposition occurred and pyridine specific poisoning evidenced that such sites are involved in the decomposition. Quantum chemistry calculations confirmed such a result and specified that the remaining O specie resulting from the decomposition, in addition to physisorbed 02, was attached to the aluminum ion constituting the Lewis sites. The oxygen molecules would then gradually be desorbed. Whereas the 102 (lAg) species pruduced via ozone photodecomposition (~ < 1180 nm) possess slow reactivity toward organic molecules, the formed via this process more active oxygen atom O(1D) may abstract hydrogen atom from the organics, surface OH groups or adsorbed water yielding to reactive hydroxyl radicals. The subsequent interaction of these radicals with ozone, organics, carbon oxide or nitrogen oxide gives active species HO2, R, H, and HONO, respectively [208]. These species are responsible for main chemical transformations of organic compounds in the atmosphere. In addition, the ozone decomposition on the solid aerosol surface is possible via its basecatalyzed route (OH-) over basic surface sites: 03 + OH- -> O~- + HO2, with the rate coefficients varying from 70 to 370 1 mo1-1 s 1 [209, 210]. 6.

CONCLUSIONS

The considered here problems in heterogeneous atmospheric organic chemistry are far from such traditional fields as surface science which deals with surface processes on well-characterized solids, as single crystals of metals and their oxides, graphite, etc., industrial adsorption, where main interest is directed to various microporous and mesoporous solids, such as zeolites, activated carbons, silica gels, and industrial heterogeneous catalysis dealing mainly with supported catalysts containing noble metals, zeolites, etc. In contrast to these fields, the study of organic pollutants interaction with solid aerosol surfaces is of no interest for industry. However, the importance of this field for the health and environmental sciences is beyond question. The studies in the heterogeneous

627 atmospheric organic chemistry are initiated only in recent years and methods developed in the surface science and catalysis have given impetus to understanding those complex processes. It is clear that the reactivity of organic pollutants toward the solid aerosol surfaces is controlled by the surface chemistry of the individual constituents of the particles. Significant advances have been made recently in the field of the powders surface characterization due to the development of theory and techniques in inverse adsorption gas chromatography. The dispersive components of the surface free energy and acid/base components of specific interaction' contribution to the adsorption energy of test polar adsorbates on various parent and modified inorganic oxides and carbonaceous materials, determined by this method make possible to estimate the adsorption energies and relative concentrations of nonpolar and polar organic compounds on the complex aerosol surfaces. The quantitative "structure-activity" and "structure-reactivity" relationships between the descriptors of organic compounds, solid surfaces and their activity in adsorption equilibria and reactivity in surface reactions presented here may be used to predict the direction and relative rate for these transformations on solid aerosols. The modeling of industrial aerosols, such as fly ash microparticles through the use of chemically modified pyrogenic metal oxide microparticles of controlled composition and surface structure offers a clearer view of detailed mechanisms for the formation of these aerosols and their interaction with atmospheric organic species. This interaction is the main reason for appearance of secondary organic carbon in the solid aerosols. The study of thermal decomposition of adsorbed organic species or the grafted groups on the model microparticles by temperatureprogrammed desorption mass spectrometry provides the comprehensive information on kinetics and mechanism of the surface reactions and gives insight to the mechanism of formation of fly ash particles.

Acknowledgements Authors wish to thank Professor P. F. Gozhyk (Ukrainian Antarctic Center) for fruitful discussions. One of the authors (V.A.P.) is indebted to Swiss National Science Foundation (Grant 7UKPJ048657) for financial support.

REFERENCES 1.

2. 3.

Environmental Action Programme for Central and Eastern Europe. Approved by Conference of Environment Protection Ministers. Lucerne, Switzerland, April, 1993, Kingfish Media Bt (ed.), Budapest, 1995. D.W. Dockery, C. A. Pope and X. Xu, N. Eng. J. Med., 329 (1993) 1753. V.M. Bogatyr'ov, M. O. Rogava and N. P. Galagan, in: Silica in Medicine and Biology, A. A. Chuiko, (ed.), Kiev- Stavropol, 1993.

628 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

C. Giordano (ed.), Sorbents and Their Clinical Applications, Academic Press, New York-London, 1980. N.A. Belyakov (ed.), Enterosorption, Center of the Sorption Technology, Leningrad, 1991. L.S. Ivlev, Chemical Composition and Structure of Atmospheric Aerosols, Leningrad University, Leningrad, 1982. V.A. Isidorov, Organic Chemistry of Atmosphere, Khimia, Leningrad, 1985. G.M.-A. Aliyev, Handbook "Dust separation and industrial gas cleaning", Moscow, Metallurgia, 1986. P. Neuling, R. Neeb and R. Eichman, Z. Anal. Chem., 302 (1980) 375. B . R . T . Simonett, Mar. Chem., 5 (1977) 433. G. Broddin, W. Cautreels and K. Von Cauwenberghe, Atm. Environ., 14 (1980) 895. M. Katz and C. Chan, Environ. Sci. Technol., 14 (1980) 838. B.R.T. Simonett, R. Chester and G. Eglinton, Nature, 267 (1977) 682. B.R.T. Simonett, M. A. Mazurek and T. A. Cahill, J. Air Pollut. Contr. Assoc., 30 (1980) 387. B.R.T. Simonett and M. A. Mazurek, Atm. Environ., 16 (1982) 2139. L. Van Vaeck and K. Van Cauwenberghe, Atm. Environ., 12 (1978) 2229. F.W. Karasek, D. W. Denney and K. W. Chan, Anal. Chem., 50 (1978) 82. W.W. Bunn, E. R. Deane and D. W. Klein, Water, Air Soil Pollut., 4 (1975) 367. M. Jaroniec and R. Madey, Physical Adsorption on Heterogeneous Solids, Elsevier, Amsterdam, 1988. I. Langmuir, J. Am. Chem. Soc., 40 (1918) 1361. A. Clark, The Theory of Adsorption and Catalysis, Academic Press, New York, 1970 M. Jaroniec, Surface Sci., 50 (1975) 553. E. C. Markham and A. F. Benton, J. Am. Chem. Soc., 53 (1931) 497. A. Dabrowski, V. I. Bogillo and V. P. Shkilev, Langmuir, 13 (1997) 936. A. Voelkel, Crit. Rev. in Anal. Chem., 22 (1991) 411. P. Mukhopadhyay and H. P. Schreiber, Colloids and Surfaces, A, 100 (1995) 47. M.M. Chehimi, in: Handbook of Advanced Materials Testing, N. P. Chesemisinoff (ed.), Marcel Dekker, New York, 1994, Ch. 33. G.M. Dorris and D. G. Gray, J. Colloid Interface Sci., 77 (1980) 353. V.I. Bogillo and A. Voelkel, Polymer, 36 (1995) 3503. F. London, Trans. Faraday Soc., 33 (1937) 8. P. C. Hiemenz, Principles of Colloid and Surface Chemistry, 2nd ed., Marcel Dekker, New York, 1986. C. Saint Flour and E. Papirer, Ins. Eng. Chem. Prod. Res. Dev., 21 (1982) 666. J . H . De Boer and S. Druyer, Proc. K. Ned. Akad. Wet., B, 55 (1952) 451. C. Kemball and E. K. Rideal, Proc. R. Soc. Ser. A, 187 (1946) 53.

629 35. V. K. Pogorelyi, V. I. Bogillo and T. F. Divnich, Teoret. Exper. Khim., 13 (1977) 677. 36. V.I. Bogillo and G. V. Filonenko, in: Modern Problems in Physical Chemistry of Solutions, N. A. Smirnova (ed.), Leningrad University, Leningrad, 1981, 89. 37. S.A. Reznikov and L. A. Belyakova, Zh. Fiz. Khim., 67 (1993) 2047. 38. V. Ya. Davydov, T. M. Roshina, G. N. Filatova and N. M. Khrustaleva, Vestn. Moscow Univ., ser. 2, 35 (1994) 411. 39. T.M. Roshina, K. B. Gurevitch, V. Ya. Davydov and A. A. Mandrugin, Vestn. Moscow Univ., ser. 2, 36 (1996) 42. 40. T.M. Roshina, A. V. Goncharov and N. V. Polyakova, Vestn. Moscow Univ., ser. 2, 33 (1992) 335. 41. T.R. Rybolt, M. T. Epperson, H. W. Weaver, H. E. Thomas, S. E. Clare, B. M. Manning and J. T. McClung, J. Colloid Interface Sci., 173 (1995) 202. 42. V.I. Bogillo, V. P. Shkilev and A. Voelkel, Adsorpt. Sci. & Technol., 14 (1996) 189. 43. V.I. Bogillo and A. Voelkel, J. Adhesion Sci. & Technol., (1997), to be published. 44. L. Lavielle and J. Schultz, Langmuir, 7 (1991) 978. 45. H.P. Schreiber and F. St. Germain, J. Adhesion Sci. & Technol., 4 (1990) 319. 46. J.B. Donnet, S. J. Park and H. Balard, Chromatographia, 31 (1991) 434. 47. H.P. Schreiber, J. M. Viau, A. Fetoui and Z. Deng, Polym. Eng. Sci., 30 (1990) 263. 48. J. Kuczynski and E. Papirer, Eur. Polym. J., 27 (1991) 653. 49. T.J. Bandosz, J. Jagiello, B. Andersen and J. A. Schwartz, Clays & Clay Minerals, 40 (1992) 306. 50. L.I. Dernovaya, A. E. Chalykh and Yu. A. Eltekov, Zh. Fiz. Khim., 67 (1993) 1996. 51. E. Papirer, G. Ligner, H. Balard, A. Vidal and F. Mauss, in: Proc. Chemically Modified Surface Symp., Midland, Michigan, June, 1989, D. E. Leyden and W. T. Collins (eds.), Sci. Publisher, New York, 1989, 15. 52. E. Papirer, A. Vidal and H. Balard, in: Inverse Gas Chromatography: Characterization of Polymers and other Materials, D. R. Lloyd, T. C. Ward and H. P. Schreiber (eds.), ACS Symp. Ser. 391, Am. Chem. Soc., Washington, DC, 1989, Ch. 18, 248. 53. W.J. Wang, S. Wolff and J. B. Donnet, Rubber Chem. Technol., 64 (1991) 559. 54. E. Papirer, J. M. Perrin, B. Siffert, G. Philipponneau and J. M. Lamerant, Progr. Colloid Polym. Sci., 84 (1991) 252. 55. F . J . Lopez-Garzon and M. Domingo Garcia, in: Adsorption on New and Modified Inorganic Sorbents, A. Dabrowski, V. A. Tertykh (eds.), Elsevier, Amsterdam, 1996, Ch. 2.8, 517.

630 56. E. Papirer and H. Balard, in: Adsorption on New and Modified Inorganic Sorbents, A. Dabrowski, V. A. Tertykh (eds.), Elsevier, Amsterdam, 1996, Ch. 2.6, 479. 57. V.I. Bogillo, L. S. Pirnach and A. Dabrowski, Langmuir, 13, (1997) 928. 58. B. Gorbunov and R. Hamilton, J. Aerosol Sci., 28 (1997) 239. 59. T. Iiyama, K. Nishikawa, T. Otowa and K. Kaneko, J. Phys. Chem., 99 (1995) 1075. 60. V.V. Turov, R. Leboda, V. I. Bogillo and J. Skubiszewska-Zieba, Langmuir, 11 (1995) 931. 61. V.V. Turov, R. Leboda, V. I. Bogillo and J. Skubiszewska-Zieba, Adsorpt. Sci. & Technol., 14 (1996) 319. 62. V.V. Turov, V. I. Bogillo and R. Leboda, in: Ext. Abstr. "EUROFILLERS '95", September, 1995, Mulhouse, France, 131. 63. V.V. Turov, R. Leboda, V. I. Bogillo and J. Skubiszewska-Zieba, Langmuir, 13 (1997) 1237. 64. V.V. Turov, R. Leboda, V. I. Bogillo and J. Skubiszewska-Zieba, J. Chem. Soc. Faraday Trans., 93 (1997) 4047. 65. V.I. Bogillo, V. V. Turov and A. Voelkel, J. Adhesion Sci.& Technol., 1997, to be published. 66. V.V. Turov, R. Leboda, V. I. Bogillo and J. Skubiszewska-Zieba, Comp. Interfaces, in press. 67. F.M. Fowkes, Ind. Eng. Chem., 56 (1964) 40. 68. S. Sugden, The Parachor and Valency, S. Routledge, London, 1930. 69. R. F. Fedors, Polym. Eng. Sci., 14 (1974) 147. 70. V.I. Bogillo, Comp. Interfaces, in press. 71. E. Estrada, J. Chem. Inf. Comput. Sci., 35 (1995) 708. 72. N. N. Avgul', A. V. Kiselev and D. P. Poshkus, Adsorption of Gases and Vapours on the Homogeneous Surfaces, Khimia, Moscow, 1975. 73. M. Randic, J. Am. Chem. Soc., 97 (1975) 6609. 74. B.M. Katz and V. V. Kutarov, Zh. Fiz. Khim., 68 (1994) 2057. 75. J.K. Thomas, Chem. Rev., 93 (1993) 301. 76. K.B. Yoon, Chem. Rev., 93 (1993) 321. 77. Yu. S. Nikitin and A. A. Pyatugin, in: Thesis of III All Union Symp. Molecular Liquid Chromatography, Riga, 1984, 28. 78. A.V. Kiselev, Intermolecular Interactions in Adsorption and Chromatography, Vyusshaya Shkola, Moscow, 1986. 79. R. S. Drago, Coord. Chem. Rev., 33 (1980) 251. 80. V. Gutmann, The Donor-Acceptor Approach to Molecular Interactions, Plenum, New York, 1978, Ch. 2. 81. F.L. Riddle, Jr. and F. M. Fowkes, J. Am. Chem. Soc., 112 (1990) 3259. 82. R. Schmidt and V. N. Sapunov, Non-Formal Kinetics, Verlag Chemie, Weinheim, 1982. 83. V.I. Bogillo and V. M. Gun'ko, Langmuir, 12 (1996) 115. 84. M. H. Abraham, Chem. Soc. Rev., 22 (1993) 73.

631 85. M. H. Abraham, J. A. Haftvan, G. S. Whiting, A. Leo and R. S. Taft, J. Chem. Soc. Perkin Trans., 2 (1994) 1777. 86. M. H. Abraham, C. M. Du, J. W. Grate, R. A. McGill and W. J. Shuely, J. Chem. Soc., Chem. Commun., (1993) 1863. 87. M. H. Abraham and D. P. Walsh, J. Chromatogr., 627 (1992) 294. 88. E. Lahman, B. Seifert and W. Dulson, Bundesgesundheitsbl., 21 (1978) 75. 89. L.A. Cavanagh, C. F. Schadt and E. Robinson, Environ. Sci. Technol., 3 (I 969) 251. 90. A. P. Altshuller and S. P. McPherson, J. Air Poll. Contr. Assoc.,13 (1963) 109. 91. G.A. Dawson, J. C. Farmer and J. L. Moyers, Geophys. Res. Lett., 7 (1980) 725. 92. E. Robinson, R. A. Rasmussen and H. H. Westberg, J. Geophys. Res., 78 (1973) 5345. 93. D. Grosjean, Environ. Sci. Technol., 16 (1982) 254. 94. P . J . Maroulis and A. R. Bandy, Science, 196 (1977) 647. 95. D.H. Fine, D. P. Rounbehler and E. Sawicki, Anal. Lett., 9 (1976) 595. 96. E. Rohbock, H. W. Georg and J. Muller, Atm. Environ., 14 (1980) 89. 97. D. C. Lowe, U. Schmidt and D. H. Ehhalt, Geophys. Res. Lett., 7 (1980) 825. 98. B.V. Ioffe, V. A. Isidorov and I. G. Zenkevich, Environ. Sci. Technol., 13 (1979) 864. 99. M.T. Dmitriev, E. G. Rastyannikov and V. S. Gladkov, Proc. Cent. Hydrometeorol. Obser., 16 (1982) 31. 100. K. H. Becker and A. Ionescu, Geophys. Res. Lett., 9 (1982) 1349. 101. G. Holzer, H. Shanfield and A. Zlatkis, J. Chromatogr., 142 (1977) 755. 102. P. Warneck, Chemistry of the Natural Atmosphere, Academic Press Inc., San Diego, 1988. 103. J. H. Williams, Anal. Chem., 37 (1965) 1723. 104. B. V. Ioffe, V. A. Isidorov and I. G. Zenkevich, J. Chromatogr., 142 (1977) 787. 105. R. A. Saunders, J. R. Griffith and F. E. Saalfield, Biomed. Mass Spectr., 1 (1974) 192. 106. R. Hagemann, H. Virelizer and D. Gaudin, Analysis, 6 (1978) 401. 107. H. B. Singh, L. J. Salas and R. E. Stiles, J. Geophys. Res., 88C (1983) 3675. 108. Y. Makide, T. Tominaga and F. S. Rowland, Chem. Lett., 4 (1979) 355. 109. P. J. B. Fraser and G. I. Pearmon, Atm. Environ., 12 (1978) 839. 110. H. B. Singh, L. J. Salas and L. A. Cavanagh, J. Air Poll. Contr. Assoc., 27 (1977) 332. 111. E. P. Grimsrud and R. A. Rasmussen, Atm. Environ., 9 (1975) 1014. 112. B. A. Soldano, P. Bien and P. Kavan, Atm. Environ., 9 (1975) 941. 113. W. Bertsch, R. C. Chang and A. Zlatkis, J. Chromatogr. Sci., 12 (1974) 175. 114. P. F. Nelson and S. M. Qulgley, Environ. Sci. Technol., 16 (1982) 650. l15. R. S. Drago and B. Wayland, J. Am. Chem. Soc., 87 (1965) 3571.

632 116. Handbook on Dissociation Energies of Chemical Bonds. Ionization Potentials and Electron Affinity, V. N. Kondratyev (ed.), Nauka, Moscow, 1974. 117. F. M. Fowkes, D. W. Dwight, D. A. Cole and T. C. Huang, J. Non-Crystalline Solids, 120 (1990) 47. 118. S. T. Joslin and F. M. Fowkes, Ind. Eng. Chem. Prod. Res. Dev., 24 (1985) 369. 119. J. A. Manson, Pure Appl. Chem., 57 (1985) 1667. 120.A.C. Tiburcio and J. A. Manson, J. Appl. Polym. Sci., 42 (1991) 427. 121. C. W. Chronister and R. S. Drago, J. Am. Chem. Soc., 115 (1993) 4793. 122. V. M. Gun'ko, V. I. Bogillo and A. A. Chuiko, ACH-Models Chem., 131 (1994) 561. 123. N. Cardona-Martinez and J. A. Dumesic, J. Catal., 128 (1991) 23. 124. V.I. Bogillo, Ukr. Khim. Zh., 57 (1991) 1274. 125. T. Hamiel, M. Nardin, M. Raqeul-Lesconet, H. Haidara and J. Schultz, Colloids Surf., A, 125 (1997) 155. 126. V. I. Bogillo, V. P. Shkilev and A. Voelkel, J. Material Chem., to be published. 127. B. Siffert, J. Eleli-Letsango, A. Jada and E. Papirer, Colloid. Surf., A, 92 (1994) 107. 128. N. Nardin, H. Balard and E. Papirer, Carbon, 28 (1990) 43. 129. J. B. Donnet and S. J. Park, Carbon, 29 (1991) 955. 130. I. Ulkem and H. P. Schreiber, Comp. Interfaces, 2 (1994) 253. 131. H. J. Jacobasch, K. Grundke, E. Mader, K. H. Freitag and U. Panzer, J. Adhesion Sci. Technol., 6 (1992) 1381. 132. P. Nortier, A. P. Borosy and M. Allavena, J. Phys. Chem., 101 (1997) 1347. 133. S. Spange, A. Reuter, A. Schramm and C. Reichardt, Organic Reactivity, 29 (1995) 91. 134. H. Knozinger and P. Ratnasamy, Catal. Rev.-Sci. Eng., 17 (1978) 31. 135. G. M. Zhidomirov and V. B. Kazansky, Adv. Catal., 34 (1986) 131. 136. M. Nakazawa and G. A. Somorjai, Appl. Surface Sci., 84 (1995) 309. 137. V. I. Bogillo, in: Adsorption on New and Modified Inorganic Sorbents, A. Dabrowski, V. A. Tertykh (eds.), Elsevier, Amsterdam, 1996, Ch. 1.9. 138. K. M. Bulanin, A. V. Alekseev, A. V. Bystrov, D. S. Lavalley and J. C. Tsyganenko, J. Phys. Chem., 98 (1994) 5100. 139. R. D. Cadle, P. Crutzen and D. Ahhalt, J. Geophys. Res., 80 (1975) 3381. 140. M. Ziolek, J. Kujawa, O. Saur and J. C. Lavalley, J. Phys. Chem., 97 (1993) 9761 141. C. Zhang, H. Hattori and K. Tanabe, Appl. Catal., 36 (1988) 189. 142.A. Auroux and A. Gervasini, J. Phys. Chem., 94 (1990) 6371. 143. H. Hattori, Chem. Rev., 95 (1995) 537. 144. A. Corma, Chem. Rev., 95 (1995) 559. 145. A. Gervasini, G. Bellussi, J. Fenyvesi and A. Auroux, J. Phys. Chem., 99 (1995) 5117. 146. K. I. Tanaka and A. Ozaki, J. Catal. 8 (1967) 1.

633 147. R. T. Sanderson, Inorganic Chemistry, Reinhold, New York, 1967, 136. 148. V. Bolis, A. Cavenago and B. Fubini, Langmuir, 13 (1997) 895. 149. R. Robins, R. Cadle and D. M. Eckhardt, Meterol., 16 (1959) 53. 150. R. Vogt and B. J. Finlayson-Pitts, J. Phys. Chem., 98 (1994) 3747. 151. D. A. Otterson and M. D. Davies, NASA Technical Note NASA TN D-2434, National Aeronautics and Space Administration, Washington, DC, 1964. 152. D. J. Dai, S. J. Peters and G. E. Ewing, J. Phys. Chem., 99 (1995) 10299. 153. M. R. Denison, J. J. Lamb, W. D. Bjorndahl, E. Y. Wong and P. D. Lohn, J. Spacecr. Rockets, 31 (1994) 435. 154. M. Y. Danilin, Ann. Geophys., 11 (1993) 828. 155. O. Mohler, K. H. Naumann, H. Saathoff and U. Schurath, Abstr. VI Int. Conf. Carbonaceous Particles in the Atmoshere, September, 1997, Vienna, Austria, 39. 156. M. J. Molina, L.T. Molina, R. Y. Zhang, R. T. Meads and D. D. Spencer, Geophys. Res. Lett., 24 (1997) 1619. 157. K. Thomas, P. E. Hoggan, L. Mariey, J. Lamotte and J. C. Lavalley, Catal. Lett., 46 (1997) 77. 158. D. C. Clary and L. C. Wang, J. Chem. Soc. Faraday Trans., 93 (1997) 2763. 159. H. W. Gao, M. Y. Huang and X. Y. Xu, Sci. China Ser. D., 40 (1997) 518. 160. G. I. Golodets, in: Partial Oxidation of Organic Compounds, ser. Problems in Kinetics and Catalysis, O. V. Krylov (ed.), vol. 19, Nauka, Moscow, 1985, 28. 161.V.V. Lobanov and V. I. Bogillo, Langmuir, 12 (1996) 5171. 162. T. Jin, T. Yamaguchi and K. Tanabe, J. Phys. Chem., 90 (1986) 4794. 163. T. Jin, M. Machida, T. Yamaguchi and K. Tanabe, Inorg. Chem., 23 (1984) 4396. 164. T. Yamaguchi, Appl. Catal., 61 (1990) I. 165. K. Afata and M. Hino, Appl. Catal., 59 (1990) 197. 166. F. G. Ciapetta and D. N. Wallace, Catal. Rev., 5 (1971) 67. 167. J. Thomson, G. Webb and J. M. Winfield, J. Mol. Catal., 68 (1991) 347. 168. J. Thomson, G. Webb, B. C. Webster and J. M. Winfield, J. Chem. Soc., Faraday Trans., 91 (1995) 155. 169. E. F. Vansant, P. Van Der Voort and K. C. Vrancken, Characterization and Chemical Modification of the Silica Surface, Elsevier, Amsterdam, 1995. 170. S. J. Lombardo and A. T. Bell, Surface Sci. Rept., 13 (1991) I. 171. V. P. Zhdanov, Surf. Sci. Rept., 12, 18 (1991) 3. 172. V. I. Bogillo, Ukr. Khim. Zh., 58 (1992) 144. 173. W. H. Weinberg, in: Springer Series in Surface Sciences, vol. 8, M. Grunze, H. J. Kreuzer (eds.), Springer, Berlin, 1987, 94. 174. V. I. Bogillo and A. A. Chuiko, Ukr. Khim. Zh., 59 (1993) 920. 175. E. Shustorovich, Surf. Sci. Rep., 6 (1986) I. 176. G. K. Boreskov, Kinet. Catal., 8 (1967) 1020. 177. V. A. Pokrovskiy, K.B. Yatsimirski and V. A. Nazarenko, Teor. Exp. Khim., 23 (1987) 377. 178. V. A. Pokrovskiy and P. F. Ustyuzhanin, Teor. Eksp. Khim., 24 (1988) 75 I.

634 179. V. A. Pokrovskiy, Rapid. Commun. Mass Spectrom., 8 (1995) 588. 180. V. A. Pokrovskiy, Adsorpt. Sci. & Technol., 14, 5 (1997) 301. 181. P. A. Redhead, Vacuum, 12 (1962) 203. 182. K. Tanabe, Solid Acids and Bases, Academic, New York, 1970. 183. M. Bowker, H. Houghton and K. C. Waugh, J. Chem. Soc., Faraday Trans., I, 78 (1982) 2573. 184. K. S. Kim, M. A. Barteau and W. E. Farneth, Langmuir, 4 (1988) 533. 185. V. S. Lusvardi, M. A. Barteau and W. E. Farneth, J. Catal., 153 (1995) 41. 186. Yu. I. Gorlov, G. N. Mel'nichenko, V. A. Nazarenko and V. A. Pokrovskiy, Fiz. Khim. Tekhnol. Poverkhn., 1 (1993) 123. 187. V. M. Gun'ko and V. A. Pokrovsky, Int. J. Mass Spectr. Ion Phys., 148 (I 995) 45. 188. V. M. Gun'ko and V. A. Pokrovskiy, Langmuir, submitted for publication. 189. A. Ludviksson, R. Zhang, C. T. Campbell and K. Griffiths, Surface Sci., 313 (1994) 64. 190. R. Sokoll, H. Hobert and I. Schmuck, J. Catal., 121 (1990) 153. 191. R. Sokoll, H. Hobert and I. Schmuck, J. Catal., 125 (1990) 276. 192. R. Sokoll, H. Hobert and I. Schmuck, J. Catal., 125 (1990) 285. 193. J. E. Germain, Intra-Sci. Chem. Repts., 6 (1972) I01. 194. Q. Dai, G. N. Robinson and A. Freedman, J. Phys. Chem., B, I01 (1997) 4940. 195. G. N. Robinson, A. Freedman, C. E. Kolb and D. R. Worsnop, Geophys. Res. Lett., 21 (1996) 377. 196. G. N. Robinson, A. Freedman, C. E. Kolb and D. R. Worsnop, Geophys. Res. Lett., 23 (1996) 317. 197. G. N. Robinson, Q. Dai and A. Freedman, J. Phys. Chem., I01 (1997) 4947. 198. L. M. Ng, E. Lyth, M. V. Zeller and D. L. Boyd, Langmuir, 11 (1995) 127. 199. M. E. Bartram, T. A. Michalske and J. W. Rogers, Jr., J. Phys. Chem., 95 (1991) 4453. 200. P. V. Kamat, Chem. Rev., 93 (1993) 267. 201. D. Oelkrug, G. Krabichler, W. Honnen, F. Wilkinson and C. J. Willsher, J. Phys. Chem., 92 (1988) 3589. 202. V. I. Bogillo, V. P. Shkilev and V. V. Osipov, Langmuir, 13 (1997) 945. 203. I. A. Myasnikov, Zh. Fiz. Khim., 69 (I 995) 2109. 204. M. A. Fox and M. T. Dulay, Chem. Rev., 93 (1993) 341. 205.A. Hagfeldt and M. Gratzel, Chem. Rev., 95 (1995) 49. 206. M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnermann, Chem. Rev., 95 (1995) 69. 207. K. M. Bulanin, J. C. Lavalley and A. A. Tsyganenko, J. Phys. Chem., 99 (I 995) 1039. 208. K. Chelkowska, D. Grasse, L. Fabian and G. Gordon, Ozone Sci. Eng., 14 (1992) 33. 209. J. Staehelin and J. Hoigne, Environ. Sci. Technol., 16 (1982) 676. 210. K. Haruta and T. Takeyama, J. Phys. Chem., 85 (1981) 2383.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

635

C o n t r o l of s u p e r c r i t i c a l g a s e s w i t h solid n a n o s p a c e - e n v i r o n m e n t a l aspects K. Kaneko Department of Chemistry, Faculty of Science, Chiba University 1-33 Yayoi,Inage, Chiba 263, J a p a n 1.

SUPERCRITICAL GASES AND THEIR IMPORTANCE

There are many important supercritical gases whose critical temperature is lower than an ambient temperature. 02, N2, NO, CO2, CO, H2 and CH4 are representatives of a supercritical gas near room temperature. 02, N2, CO2, CH4 and NO are intricated with important bioreactions. 02, H2, CH4 and CO2 play an indispensable role in energy technology. The control of CO2, C H 4 and NO is necessary to preserve our earth environment. Thus, these gases are essentially important to h u m a n and earth. Even inert gases such as Ar and Xe play an important role in modern technology. These supercritical gases have been deeply associated with gas separation, gas storage, catalysis, supercritical extraction, supercritical drying, pollution control and life science. The chemistry of supercritical gases should contribute to establish an energy-saving technology of high performance. Accordingly, control of a supercritical gas is strongly requested for establishment of the future technology. Adsorption of supercritical gases by micropores can offer a hopeful control method. Here micropores are the pores whose width is less t h a n 2 nm according to the IUPAC classification, which is shown in Table 1. In this article, we will use the term of nanospaces for micropores whose width is less than about 1 nm. In a nanospace the molecule-pore wall interaction is markedly enhanced, as mentioned later. The adsorption science for a supercritical gas with the nanospace can support the development of fluid chemistry and technology in future. Therefore, we need to understand the nature of adsorption of the supercritical gas by nanospaces. Also we must know molecular properties of these supercritical gases in order to search the best nanospace system for each fluid gas. The above-mentioned molecules have different properties and structures. 02, N2, NO, CO2, CO and H2 are linear molecules and their point groups are C~v and D~h ; only CH4 has Td symmetry. Although their molecular orbitals for valence electrons other t h a n H2 are composed of 2s and 2p atomic orbitals, occupation of

636 Table 1 Classification of pores Micropore

w<2nm

Ultramicropore

w < 0.7 nm

Supermicropore

0.7 n m < w < 2 n m

Mesopore

2 nm < w < 50 nm

Macropore

w > 50 nm

(Nanopore is not recommended by IUPAC, but it is often used for pore whose width is less than 10 nm).

the molecular orbitals by electrons are different from one molecule to another, providing different intermolecular interaction energy, electrostatic s t r u c t u r e and magnetic strucuture, as shown below. Their molecular sizes and interaction energies are different from each other. L e n n a r d - J o n e s p a r a m e t e r s (~ff and ~ff ) give good m e a s u r e s of the size and interaction energy. H2 of ~ff = 0.292 nm is the smallest molecule, while CO2 of aff = 0.376 nm is the greatest of these molecules. However, aff, being the one-center L e n n a r d - J o n e s p a r a m e t e r , is obtained by the a s s u m p t i o n of the spherical shape and it is not the exact m e a s u r e of the molecular size. Each molecule has a different ~ff value; CO2 has the greatest interaction energy. NO and CO have a small p e r m a n e n t dipole moment. O2, N2, CO2 and H2 have the quadrupole moment, while NO and CO have the quadrupole m o m e n t too. Only CH4 has the octapole moment. Although the p r e d o m i n a n t intermolecular interaction stems from the dispersion interaction, these multipole m o m e n t s contribute to the intermolecular interaction, determining their m u t u a l orientation. 02 and NO are paramagnetic, whereas others are diamagnetic. Hence 02 and NO can interact with magnetically. Thus, these molecules have different properties. Table 2 s u m m a r i z e s physical properties of these molecules. We can use the physical property difference. Here Tb, Tc and Pc are the boiling t e m p e r a t u r e , critical t e m p e r a t u r e and critical pressure, di, qu and oc denote dipole, quadrupole and octapole moments. The units of di and qu are Cm and Cm 2, respectively. Although NO and CO have the quadrupole m o m e n t in addition to the dipole moment, their quadrupole m o m e n t s are ommitted, d i a and p a r a denote d i a m a g n e t i s m and p a r a m a g n e t i s m .

637 Table 2 Physical properties of i m p o r t a n t molecules Molecule H2

Tb K

Tc K

Pc MPa

(~ff nm

s~f/kB K

20.3

33.0

1.29

0.292

38.0

Multipole moment

Magnetism

qu

dia

+2.1.10-40 02

90.2

154.6

5.04

0.338

126.3

qu

para

-1.33.10-4o N2

77.3

126.2

3.39

0.363

104.2

qu

dia

-4.90" 10 -40 NO

121.4

180

6.48

0.347

119

di

para

0.158-10 -3o CO

81.6

132.9

3.50

0.359

110

di

dia

0.112.10-3o CO2

194.7

304.2

7.48

0.376

245.3

qu

dia

-14.9.10-40 CH4

2.

111.6

190.5

4.60

0.372

161.3

oc

dia

W H A T IS A S U P E R C R I T I C A L GAS ?

The state of m a t t e r is described in terms of the pressure P, the molar volume Vm and the t e m p e r a t u r e T. Then three phases of gas, liquid and solid are expressed by the P-Vm projection, as shown in Figure 1. The P-Vm projection indicates the presence of the coexistent region of gas and liquid in equilibrium, which is designated by ( 1 + g ). The broken line parallel to the abscissa of the P-Vm projection at a t e m p e r a t u r e T1 denotes the coexistent region. The coexistent region becomes narrower, as the t e m p e r a t u r e is raised. Finally the coexistent region is reduced to a mere point whose t e m p e r a t u r e is called critical t e m p e r a t u r e Tc. The critical t e m p e r a t u r e Tc is the m a x i m u m t e m p e r a t u r e at which a gas can be liquefied and above Tc liquid cannot coexist. Above Tc there is no vaporization curve and no distinction between liquid and gas. Therefore, we must distinguish the gaseous states above and below Tc. The term vapor is used to describe a gaseous substance when its t e m p e r a t u r e is below Tc and the vapor can be condensed to liquid by pressure alone. However, the gas above Tc, which is called a supercritical gas, cannot be liquefied even at quite high pressures. Vapor has own s a t u r a t e d vapor pressure Po. Then we can use the relative pressure P/P0 for description of adsorption of vapor. On the other hand, the supercritical gas

638

P 1

"'criticalpoint V ~ T V, Figure 1. Phase diagram.

has no s a t u r a t e d vapor pressure. The relative pressure expression cannot be used for description of adsorption of a supercritical gas. Recently not only liquid but also supercritical gas has been used as solvents [2]. The solvent power of a supercritical gas increases with density at a given t e m p e r a t u r e and it increases with t e m p e r a t u r e at a given density. A supercritical gas exhibits physicochemical properties of an intermediate between a liquid and a gas. The relatively high liquid-like density at high pressure affords good solvent power and the mass transfer in the supercritical gas is rapid relative to a liquid. Also the extremely low value of surface tension of the supercritical gas allows better penetration into the sample matrix relative to liquid solvents. The critical t e m p e r a t u r e and pressure of CO2 are 304.21 K and 7.477 MPa, respectively and thereby CO2 has been widely served in supercritical fluid technology. However, the fundamental u n d e r s t a n d i n g of the supercritical state is not necessarily sufficient yet. The supercritical state in the bulk phase has been studied with small angle Xoray scattering, showing the presence of inherent clusters around the critical point for CO2 [3]. The researches on the bulk supercritical state should stimulate the adsorption study on supercritical gases. 3.

NANOSPACE SYSTEMS

The most representative microporous solids are zeolites and activated carbons. Both have different pore structures and adsorptive properties each other. Recently new porous solids have been developed to accelerate the progress in nanopore fluid chemistry. There are two types of pores of intraparticle pores and interparticle pores [4]. The intraparticle pore is in the primary particle itself, while the interparticle pore originates from the interparticle void space, although there is an ambiguous

639 distinction between both types of pores in some systems. Zeolites have welldefined intraparticle pores which arise from the intrinsic crystalline structure [5]. The pore geometry and pore connectability are evaluated by their crystal structure. Aluminophosphates [6] also have cylindrical intraparticle pores inherent to the crystal structures. Even the hydrophobic phosphates [7] having the straight pore of triangular column were synthesized. On the other hand, a new family of mesoporous zeolites [8-11] such as MCM (Mobil Composition of Matter) or modified kanemite FSM (Folded Sheets Mesoporous Material), which have straight cylindrical mesopores, were developed recently. Furthermore, the mesopore size can be controlled by the surfactant molecular size. These mesoporous silica has a regular honeycomb structure, although the pore walls are noncrystalline. The mesoporous silica provided a new problem that the adsorption hysteresis depends sensitively on the mesopore size [12,13]. The classical capillary condensation theory cannot explain the dependence of the adsorption hysteresis, but Inoue et al. gave a new approach to this problem with the extended Saam-Cole theory [14]. The carbon nanotube [15] has also the intraparticle pore stemming from the intrinsic crystalline structure; the tube-wall is composed of graphitic structures. However, the so-called carbon nanotube has a both end-closed cylindrical pore. On the other hand, the catalytic method can produce the one end-open pore in which gas adsorption is available [16]. The CVD technique using the template porous solids produces mesocarbon tubes [17]. Activated carbons are the most popular adsorbent [18-20]. Activated carbons are mainly composed of micrographitic units. The edge carbon atoms of the micrographite are more reactive than carbon atoms on the basal plane, developing pores along the basal plane of the micrographite. Activated carbon fibers (ACFs) [21] have only uniform micropores, while the conventional granulated activated carbons have a wide pore size distribution from micropores to macropores. In particular, pitch-based ACFs have less amount of surface functional groups and the pore width can be well controlled. Superhigh surface-area carbons [22,23] obtained by KOH activation have considerably uniform micropores whose pore width is greater than that of ACFs. The activated carbon film from Kapton film has the oriented structure of slit micropores [24]. The carbon aerogel has uniform mesopores and micropores can be donated to the carbon aerogel [25,26]. Also graphite intercalated compounds can offer micropores [27]. In addition to the above porous systems, pillared clay minerals [28,29] and pore-size controlled glasses [30] can be available for controlling a supercritical gas. Microporous BN is also an attractive solid [31]. An organic metal complex is a new hopeful nanopore system of which width can be variable [32]. As these porous solids have different pore geometry and chemical nature (Table 3), we can choose the best fit nanopore system for control of each fluid.

640 Table 3 Nanoporous s y s t e m s Surface component

Compound Zeolite A l u m i n o p h o s p h a t e ALPO

Si, A1, O

Pore shape nm

Pore width nm

cylinder, cage

0.3 - 1

cylinder

0.8-1.3

A1,P, O

[SAPO: Si, A1, P, O; TAPO: Ti, A1, P, O; FAPO: Fe, A1, P, O, etc] Aluminiummethyl phosphonate A1, CH3PO3 cylinder < 1 triangle prism Mesoporous zeolite Si, O cylinder 2 - 10 [A1, Zn, Ti, Zr, W, Pb etc can be doped] activated carbon fiber carbon aerogel activated carbon aerogel pore-oriented carbon film carbon n a n o t u b e carbon mesotube graphite i n t e r c a l a t i o n comp. microporous BN pillared clay porous glass organic m e t a l complex

4.

C C C C C C C, Ketc B, N, H Si, A1, O Si, O Organic group

slit i n t e r g r a n u l a r void slit + void slit cylinder cylinder slit slit slit cylinder void

0.6 - 1.3 5 - 30 1, 5-30 < 1 > 2 30 - 200 < 0.5 < 1 > 0.5 5 - 104 variable

DEEP POTENTIAL WELL OF NANOSPACE

We will discuss the interaction of a molecule with the graphitic slit pore of the micropore model of activated carbon. The interaction b e t w e e n a molecule and a surface atom as a function q)(r) of the distance r between t h e m can be expressed by the L e n n a r d - J o n e s potential, (I)(r) = 4~sf [(asf/r) 12 - ( a s f / r ) 6 ]

(1)

where ~sf and asf are the well depth and effective d i a m e t e r for the moleculegraphitic carbon atom. These cross p a r a m e t e r s are calculated according to the Lorentz-Berthelot rules, ~sf = (~ss ~ff )l/e ; asf = (ass + aff )/2. Here, (ass , ~s) and (aff, ~ff ) are the L e n n a r d - J o n e s p a r a m e t e r s for a surface atom and a molecule, respectively. The interaction potential (I)(z) for a molecule and a single graphite slab is given by the Steele 10-4-3 potential [33],

641

(I)(z) = 2rtPCg sf (~2sfA{(2/5)((~sf/z)10 - ((~sf/z }4 _(~4sf/bA(0"61A + z)3 ] }

(2)

where z is the vertical distance of the molecule above the surface, A is the separation between graphite layers (=0.335 nm), pc is the n u m b e r density of carbon atom in a graphite layer (=1i4 /nma). As the micropores of activated carbon can be approximated by the slit spaces between the p r e d o m i n a n t basal planes of nanographitic units, the whole interaction potential ~(Z)p of a molecule with the micropore of an inter-graphite surface distance H can be given by eq. (3).

|

: |

(a)

Consequently, we can evaluate the potential profile of the molecule adsorbed in the graphitic micropore. Here H is not the effective pore width w determined by the adsorption experiment. The difference between H and w is a function of ~sf and ~ff [34]. (4)

H - w = 0.85~sf - ~ff

5O0

-500 1

s

s'

~" -1500

-2500

'

'

I

,

,

-0.4

-0.2

0 z/rim

0.2

0.4

Figure 2. The potential profile pore width w.

I

of a N2 molecule in the

graphite slit pore as a function of the

In the case of the Ne-graphitic slit pore system, w is equal to H - 0.24 nm. Figure 2 shows potential profiles of N2 in a slit-shaped graphite pore as a function of w using the one-center approximation. Here, the molecular position in Figure 2 is expressed by a distance z from the central plane between two surfaces. The broken line shows the potential profile for the single graphite surface. The

642 potential becomes deeper with decrease in the w value. The potential profile has double minima for w > 0.6 nm, but a narrower pore has a single potential minimum. Thus, micropores have stronger adsorption fields t h a n flat or mesoporous surfaces. The depth of the potential well for vapor is enough great to give the Type I adsorption isotherm, i.e. an enhanced adsorption at a low relative pressure range, being characteristics of micropore filling [35]. Activated carbon has a b u n d a n t micropores and their adsorption field can be approximated by the above mentioned graphite-slit space model. The depth of the potential well of nanospace is not enough for a supercritical gas to be sufficiently adsorbed, but chemical modification of the pore-walls can deepen the potential well for the supercritical gas, leading to a marked micropore filling. Furthermore, supercritical gas molecules tend to be adsorbed in micropores of the deep potential well and their intermolecular interaction is enhanced to form an organized structure, as if molecules were compressed by a high pressure. For example, when adsorbed molecules of supercritical gas form the dimer in a chemically modified micropore, the interaction profile of the dimer with the micropore becomes completely different from that of the monomer. Hence, the adsorption property for the supercritical gas can be changed dramatically by use of a chemically modified micropore.

0

F U N D A M E N T A L P R O B L E M S IN A D S O R P T I O N OF S U P E R C R I T I C A L GAS

The supercritical gas which has no concept of the s a t u r a t e d vapor pressure cannot be sufficiently physisorbed on the flat surface, macropores and even mesopores with physical adsorption. The adsorption of vapor by micropores, which is called micropore filling, is enhanced at a very low pressure region due to overlapping of the molecule/pore-wall interactions. The deep potential well of the micropore gives rise to adsorption of the supercritical gas to some extent. Still the micropore cannot adsorb a b u n d a n t amount of the supercritical gas. That is, the quite narrow micropore whose width is fit for the size of the adsorbate molecule has a very deep molecular potential well and is effective even for adsorption of the supercritical gas. However, the amount of adsorption by such a narrow micropore is limited to much less t h a n the micropore volume due to the diffusion problem; molecules adsorbed near the entrance in the micropore are bound too strongly to migrate to the inner pore position. Accordingly, the supercritical gas is not an objective gas for a predominant micropore filling [36]. In case of micropore filling of vapor, the potential well of the micropore whose width is even more t h a n trilayer thickness of the molecular size is enough deep for vapor molecules to be sufficiently filled at a low pressure region without any obstacle for the intrapore diffusion. The analytical method for physical adsorption of a supercritical gas by micropores is much less advanced compared with physical adsorption of vapor

643 [37-39], although molecular simulation [40-42] based on the Lennard-Jones is effective for description of adsorption of supercritical gases by micropores. However, we need a simple description method like Dubinin-Radushkevich (DR) equation for adsorption of supercritical gas in micropores, as given by eq.5 [43]. W = Wo exp[-(A/~ E) 2 ]

A : RT ln(P/P 0), E = 13E o

(5)

Here, W0 is the pore volume, E the energy constant, Eo the characteristic adsorption energy and 13 the affinity coefficient. The DR equation includes the saturated vapor pressure Po and thereby it cannot be applied to micropore filling of supercritical gas. As the interaction potential at the mid-point of the slit-pore is the deepest in the pore of w < 0.6 nm and it is not seriously different from the double minima even for w > 0.6 nm, molecular potential for a molecule in the micropore can be approximated by the potential at the mid-point of the slit-pore. The molecular potential indicates the presence of the inherent micropore of the volume of WL for each adsorptive. Here the inherent micropore must have a sufficiently strong molecular field in comparison with the thermal energy at a measuring temperature. The inherent micropore volume WL for vapor molecules is almost equal to the micropore volume Wo obtained from N2 adsorption at 77 K (Gurvitch rule). WL for a supercritical gas which depends on the molecule-pore interaction can be evaluated as the saturated amount of adsorption from the Langmuir plot and WL is less than Wo in general. The supercritical gas is transformed into a quasi-vapor in the micropores of which pore volume is WL. Then, the quasi-saturated vapor pressure Poq can be defined for the quasi-vapor. Both of WL and Poq are determined experimentally, as described using the following supercritical DR equation [37]. [ln(WL/W 0 )]1/2 : (RT/~E 0 )0n Poq _ l n P )

(6)

The plot of [ln(WL/Wo)] 1/e vs. In P leads to both of Poq and ~Eo. This supercritical DR plot is quite useful to obtain the important information on adsorption of supercritical gas. This supercritical DR equation can describe the adsorption isotherm of the supercritical gas using the concepts of Poq and WL, which are related to the intermolecular interaction of supercritical gas in the quasi vapor state and the molecule-pore interaction, respectively. Also adsorption isotherms at different t e m p e r a t u r e s can be reduced to a single isotherm with the aid of these reduced quantities of WL/W0 and P/Poq Hence we can predict the adsorption isotherm at an arbitrary temperature by use of the reduced isotherm.

644

D

CONTROL OF MICROPORE FILLING AND NANOSPACE REACTIVITY OF SUPERCRITICAL GASES

As micropore filling is governed by both geometry of pores and chemical nature of the micropore-wall, we can control it with chemical modification of the micropore-walls. The chemical modification with substances having a weak chemisorptive activity for molecules is ~ffective for enhancement of micropore filling of a supercritical gas; this is called chemisorption-assisted micropore filling [44]. A marked enhancement of micropore filling of supercritical NO and CH4 with the chemical modification is described in this article. Micropores have stronger adsorption fields due to the deep potential well than flat or mesoporous graphitic surfaces. We can estimate the effective pressure from the potential profile; molecules confined in the slit-pore of I nm in width are presumed to be exposed to the high pressure of 100 MPa. Therefore, the graphitic micropore can offer the high pressure field from the macroscopic view. The quasihigh pressure effect was evidenced in the disproportionation reaction of the NO dimer and hydrate mediated micropore filling for NO and CH4, as shown later. Hashimoto et al. [45] reported the presence of the quasi high pressure effect even in an electrochemical reaction using ACF.

6.1. Iron o x i d e - d i s p e r s i o n i n d u c e d m i c r o p o r e filling of s u p e r c r i t i c a l NO The critical temperature of NO is 180 K and NO is a supercritical gas at ambient conditions. Almost all microporous adsorbents cannot sufficiently adsorb supercritical NO, although NO of the representative atmospheric pollutant is desired to be removed with a good adsorbent. Table 4 summarizes NO adsorptivities of activated carbon, zeolite and silica gel at 303 K. Although activated carbon can adsorb more NO than representative zeolites and silica gels, the maximum adsorption does not cope with the saturated adsorption which can be estimated from the pore volume. Zeolite is not a good adsorbent for NO regardless of presence of micropores. Mesoporous silica gel is also not fit for adsorption of supercritical NO. Only activated carbon is considerably effective for NO adsorption due to the assistance by the surface functional groups. Hence the control of surface chemistry of activated carbon is quite essential with the relevance to control of the dimerization of NO. An NO molecule has an unpaired electron and gaseous NO shows paramagnetism. It is well-known that NO molecules are dimerized and show diamagnetism at the condensed phase at low temperature [46]. Consequently, we can understand the molecular state of NO molecules adsorbed in micropores of ACF by the magnetic susceptibility measurement. ACF having less surface functional groups shows diamagnetism, giving a reliable result. The magnetic susceptibility ~ of ACF with adsorbed NO was negative near room temperature irrespective of the adsorption of NO which has a large paramagnetic (positive) susceptibility. The ~ value calculated from the c values of both gaseous NO (of the same amount as adsorbed NO) and ACF was positive and decreased with the measuring temperature. Thus, the calculated temperature dependence of ~ was

645 clearly different from the observed one. This is because the adsorbed NO does not exhibit p a r a m a g n e t i s m ; the negative )~ of ACF with adsorbed NO arises from the dimerization of NO in micropores. The fraction of dimers determined from the a m o u n t of NO adsorption, the observed magnetic susceptibility and the diamagnetic susceptibility of the NO dimer in the literature is 0.98 at 298 K and it is still about 0.9 even at 373 K [47-49]. If we can accelerate the dimerization of NO, NO adsorption can be enhanced, because NO dimer is vapor.

Table 4 Amounts of NO adsorption of activated carbons, zeolites and silica gel at 303 K under the equilibrium NO pressures of 13 and 80 kPa Surface NO mg g-1 area adsorbed 80kPa Substance m2g -1 13kPa 1100

15

28

coconut shell-based activated carbon

860

17

47

molecular sieve carbon (pore width: 5/k)

500

28

60

1100

15

28

860

17

47

g r a n u l a t e d coal-based activated carbon

g r a n u l a t e d coal-based activated carbon coconut shell-based activated carbon

500

28

60

cellulose-based activated carbon fiber(CEL-ACF)

1400

36

65

pitch-based activated carbon fiber

1530

10

55

molecular sieve carbon (pore width: 5/k)

polyacrylonitrile-based activated carbon fiber

830

65

115

iron hydroxide dispersed CEL-ACF

820

120

150

1070

260

320

iron oxide dispersed CEL-ACF molecular sieve 3A

-

0.6

molecular sieve 4A

-

0.6

molecular sieve 5A

-

5

24

Na-mordenite

-

4

22

molecular sieve 13 X

-

2

9

iron hydroxide dispersed molecular sieve 13X

-

5

12

Silica gel

680

2 6

3

5

iron hydroxide dispersed silica gel

600

14

20

Iron oxide*

-50

20

50

*) NO is chemisorbed on iron oxide.

646 NO is weakly chemisorbed on transition metal oxides near room temperature. The electrical conductivity of n-type iron oxides decreases very rapidly upon chemisorption of NO near room temperature [50]. Although NO forms the nitrogen oxide with surface oxygens of iron oxides, almost chemisorbed NO molecules are desorbed by evacuation. The weak chemisorption of iron oxide can assist micropore filling of supercritical NO on microporous carbons; the dispersion of iron oxides near the entrance of micropores of ACF should enhance physical adsorption of supercritical NO due to the concentration increase of NO near the dispersed oxide particles. Figure 3 shows adsorption isotherms of NO on iron oxide-dispersed ACF (Fe-ACF) at 303 K. As iron oxyhydroxide decomposes to iron oxide by heating, iron oxide-dispersed ACF was prepared from iron oxyhydroxide-dispersed ACF by heating at different temperatures. The NO adsorption depends on the preheating temperature of dispersed iron oxyhydroxides. Heating of iron oxyhydroxide dispersed ACF at 873 K for 15 h in vacuo gives the greatest amount of adsorption. The dispersed iron oxides was characterized by EXAFS spectroscopy, showing the presence of ultrafine iron oxides. Fe-ACF can adsorb great amount of NO (maximum: 320 mg/g-adsorbent) in the dimer form at 303 K. Thus, dispersion of ultrafine iron oxide particles enhances markedly NO adsorption. The adsorption isotherm shows a remarkable hysteresis; the adsorbed NO cannot be removed by evacuation with a high

300 (

1 : - . O " ~

200

,~ 100

t

t

i

I

I

I

NO pressure / ~a Figure 3. The adsorption isotherms of NO on iron oxide -dispersed ACF at 303 K as a function of heating temperature of dispersed iron oxyhydroxides in vacuo. (A, A) : 573K, (V,V): 673 K, (O,Q) : 773 K and ( ~ ,'00:823 K. Solid and open symbols denote adsorption and desorption branches, respectively.

647 vacuum pump at 303 K, but it can be desorbed with the ultrahigh vacuum system. This irreversibility arises from dimerization of NO in micropores. As NO dimer is vapor at an ambient temperature, supercritical NO can be adsorbed by micropores through the dimerization with the aid of the electronic interaction and magnetic perturbation due to the high spin Fe a§ ions in the dispersed oxide [51-53]. The importance of the magnetic interaction was evidenced by application of the strong external magnetic field. The application of the magnetic field of 1 T to the adsorption system increased instantaneously NO adsorption on ACF by 1% [54]. Not only the magnetic perturbation but also a weak chemisorptive mechanism should be associated with the enhancement of the NO micropore filling.

160

~" 140 0 Z 120 o

lOO I

I

I

I

0

1

2

3

DopedTi IFe 1 %

Figure 4. The saturated amount of NO adsorption per unit micropore volume at 303 K against the amount of Ti dopant.

Doping of Ti 4§ increases the electrical conductivity of n-type iron oxyhydroxide fine crystals or iron oxide thin film due to formation of a quasi-free electron [55,56]. Hence, the mixed valence formation in the dispersed iron oxides with doping of Ti 4§ enhanced the micropore filling of NO [57]. The adsorption isotherm of NO on Ti-doped iron oxyhydroxide dispersed ACF at 303 K is Langmuirian, which is described by the Langmuir equation. The saturated amount of NO adsorption (WL) from the Langmuir plot as a function of the amount of Ti dopant (Ti/Fe: 0 - 3.6 %) is shown in Figure 4. Ti doping increases markedly WL; Ti doping of 3.6 % enhances WL by about 40 %. Ti doping should be associated with NO dimerization. The NO adsorption isotherms are described by the supercritical DR equation (eq.6). the quasi-saturated vapor pressure P0q can be defined for the

648 quasi-vapor. The determined Poq was in the range of 215 to 480 kPa. Ti doping lowers the P0q value so that the NO micropore filling is enhanced. The supercritical DR plot gives the isosteric heat of adsorption at the fractional filling (~ of e -1 (qst,~=l/e). The qst,~=l/e values (23-28 kJ mol 1) are higher t h a n the dissociation enthalpy of the NO dimer (12-16 kJ mo1-1 ) in the condensed bulk phase at a low temperature. Thus NO dimers are stabilized in the micropores of these ACF samples. If the supercritical DR equation is correctly applicable to the description of the micropore filling for supercritical NO, all adsorption data of different doping samples must be expressed by a single reduced adsorption isotherm having the abscissa of the relative pressure P/P0q. Figure 5 shows such a reduced adsorption isotherm for NO vaporized in the micropore. Almost all the observed points form one curve. Thus, the supercritical DR equation can describe well the micropore filling of supercritical NO. The above facts show a typical chemisorption-assisted micropore filling of supercritical NO.

12o

,N.

80 o.,-~

4o

9 Z I

0

I

0.2

I

0.4

P/P0q(NO) Figure 5. Reduced adsorption isotherm of supercritical NO for iron oxyhydroxide dispersed ACF having different amounts of Ti dopants at 303 K. Ti/Fe %: o none, A 0.2, [] 1 and * 3.

6.2. R e d u c t i o n o f NO to N2 in n o b l e m e t a l - t a i l o r e d n a n o s p a c e Oxides of nitrogen NOx is the inevitable by-products of high temperature combustion and the representative atmospheric pollutants. Increasing automobiles have emitted greater amounts of NOx, giving rise to a serious atmospheric pollution problem in urban areas in particular. Although NO2 can be

649 easily absorbed in soils or dissolved in surface water, a diluted NO is kinetically extremely stable in the absence of suitable catalysts in atmosphere. Consequently an ~fficient removal of NO or reduction of NO to N2 has been desired. There are active studies on catalytic reductions of NO to N2 all over the world [58-60]. The most noticeable catalyst for the decomposition of NO to N2 is Cu-ZSM-5 developed by Iwamoto et al. [60]. This Cu-ZSM-5 decomposes NO into N2 and 02 near 750 K, but coexistent SO2 strongly poisons the catalytic activity. Recent efforts are done on the development of the catalyst for selective catalytic reduction with N-free reductants [59]. NO is quite i m p o r t a n t as not only the atmospheric pollutant but also a biological molecule; recently it is elucidated that NO plays an important role in biomolecular organisms of n a n o m e t e r size [61]. Thus reactivity of NO in nanospace has gathered much attention. The nanospace can work as the high pressure field, as mentioned before. The disproportionation reaction of (NO)2 given by Eq. (7), 3(N0)2 = (N02)2 + 2 N 2 0

(7)

is known as the high pressure gas phase reaction above 20 Mpa [62]. NO molecules dimerized in the micropore of ACF at a subatmospheric pressure of NO give rise to the high pressure disproportionation reaction of the NO dimer in micropores [63]. Furthermore, the produced N20 is reduced to N2 at 423 K with the aid of dispersed transition metal oxides [64]. However, the NO reduction reaction is very slow; it takes more t h a n 10 h. Ru fine particles exhibit a high catalytic activity for NO. Ultrafine Ru particles can be dispersed in micropores of ACF. The Ru fine particle-dispersed ACF is designated Ru-ACF (Ru content: 4 wt. %). The surface areas and micropore volumes determined by N2 adsorption are 1130 m2g-1 and 0.51 mlg-' for Ru-ACF and 1100 m2g-1 and 0.52 mlg -1 for ACF, respectively. The average slit-shaped widths of Ru-ACF and ACF are 0.94 and 0.90 nm, respectively. The reaction extent was determined by the compositional changes of the gas phase with FT-IR and Mass spectrometers. Figure 6 shows the FT-IR spectral change of the gas phase over Ru-ACF at 303 K. The intensity of NO at 1876 cm -1 decreases rapidly; it becomes less t h a n 15% of the initial intensity after 9 min. On the other hand, the bands of NO2 and N20 appear at 1618 cm -1 and 2224 c m -1, respectively. As the absorption intensity of NO is noticeably weak compared with that of the NO2 and N20 bands, the concentration of produced NO2 and N20 after 18h corresponds only less than 0.3% of the residual NO; evolution of CO2 was much less t h a n that of NO2 or N20. IR cannot detect N2, then the whole gas after 10 min and 18 h was analyzed by the mass spectroscopy. The mass analysis elucidated the formation of N2 corresponding to the 77 % yield. Hence, NO is rapidly changed into N2 over RuACF at 303 K. Figure 7 shows the change in the NO concentration ratio vs the initial concentration determined from the corrected IR absorption band intensity of NO at 1876 cm -1. NO is rapidly reduced to N2 at 303 K and 323 K and then the

650

I.) o

a O raej

<

2400

2200

2000

I

I

1800

1600

W a v e number(era")

Figure 6. FT-IR spectral change of NO over ultrafine Ru particle-tailored ACF with the reaction time at 303 K. a: before reaction, b" 3 min, c" 9 min, d" 60 min and e" 180 min.

1.0

0.8 O .,..,

.o

0.6

r

o

0.4

0

9 Z

~

0.2

0.0 0

10

20

30

)

i

i

40

50

60

Time/min.

Figure 7. Change of NO concentration ratio with the reaction time for ultrafine Ru particletailored ACF at different temperatures. []" 303 K, ~" 323 K and 9 423 K.

651 NO decomposition, almost finishes within 7 min. Raising the reaction temperature to 423 K lowers the reaction rate. The first order plot for the initial reaction gives a half-life of 3 min. Thus the NO reduction reaction in the Rutailored micropore space is remarkably fast even near ambient conditions. RuACF has much greater rate than ACF. The NO reduction reaction at 323 K is almost similar to that at 303 K. As the amount of NO adsorption at 303 K is greater t h a n that at 323 K, the removal ratio of NO at 323 K (0.92) is slightly smaller t h a n that at 303 K (0.96). The CO2/N2 ratio of Ru-ACF is almost zero below 323 K. Although the NO reduction reaction at 423 K produces not only NO2 and N20, but also CO2, the reduction of NO at the ambient conditions produces negligible amounts of NO2, N20 and CO2. The dimerized NO is in an equilibrium with NO2 and N20 in the micropore of ACF even at ambient conditions, as given by eq. 7. The dispersed Ru fine particles in the micropore should accelerate the decomposition of N20 into N2 and the produced oxygen should be chemisorbed on the carbon pore wall. If we can produce oxygen from this reaction, this reaction should be a hopeful method for NO removal [65].

6.3. Oxide-dispersion induced enhancement of micropore filling of CH4 Methane is the main constituent of natural gas. Adsorption of methane at ambient temperature has been studied with a special relevance to methane storage [66,67]. However, methane is a spherical molecule and the intermolecular interaction is weak. Furthermore, the bulk critical temperature is 191 K. It is quite difficult to adsorb methane sufficiently at ambient conditions. The adsorption conditions of CH4 by activated carbon have been studied with molecular simulations [66,68,69]. The chemisorption-assisted micropore filling concept was applied to adsorption of supercritical CH4. As the intermolecular interaction of CH4 is quite weak compared with that of NO, the CH4 adsorption must be examined at high pressure region. Basic metal oxides such as MgO, CaO, A1203, NiO and Cr203 have a catalytic effect o n CH 4 at high temperature, they and their hydroxides were dispersed on ACF in order to improve adsorptivity for supercritical CH4. The microporosity of these oxide-dispersed ACFs is not seriously changed. The oxide dispersion enhances remarkably the amount of CH4 adsorption at 303 K. The dispersion of NiO fine particles is the most effective for improvement of CH4 adsorptivity. The saturated amount of adsorption of NiO-dispersed ACF is greater t h a n that of ACF by 50 % [70-72]. However, the enhancement of supercritical CH4 adsorption with chemisorption-assisted micropore filling still needs high pressure. We must search another new mechanism for supercritical CH4 adsorption near ambient conditions.

6.4. Hydrate-mediated micropore filling of supercritical gas Recent in situ X-ray diffraction study showed that water molecules confined in the micropores of activated carbon have an organized structure close to the ice structure r a t h e r t h a n the bulk liquid one [73,74]. Recent molecular simulation

652 studies show the hydrogen-bonded network structure of w a t e r molecules in wide carbon micropores [75] and a special bilayer structure of w a t e r molecules formed in the slit space by application of the high pressure [76]. The guest molecules can be intricated in the organized water structure in the micropores without application of the high pressure due to the quasi-high pressure effect. Desorption isotherm of NO with the progress of H20 adsorption in micropores of iron oxidedispersed ACF strongly indicates the formation of NO-H20 clathrate under the subatmospheric pressure [77]. The formation of the bulk NO-H20 clathrate requires more t h a n 100 M P a according to the literature [78]. This fact clearly suggests a new method of controlling the adsorption of supercritical gases with coexistent w a t e r vapor. This idea can be applied to m e t h a n e storage problem [79]. Vast deposits of n a t u r a l gas exist in the form of clathrate h y d r a t e s under the oceanbed and in the permafrost regions. The physicochemical properties of the clathrate h y d r a t e s have gathered much attention due to their special relevance to new energy resources and conservation of the earth's e n v i r o n m e n t [80]. Crystallographic studies on m e t h a n e h y d r a t e s show t h a t the lattice constant of structure I is 1.20 nm [78,80,81]. The stable formation of bulk m e t h a n e h y d r a t e s needs both lower t e m p e r a t u r e s and higher pressures due to the weak bonding between a m e t h a n e molecule and the hydrogen-bonded cage structure of water; a m i n i m u m 80 M P a at 303 K is necessary for the formation of the m e t h a n e hydrate. M e t h a n e h y d r a t e s are formed in mesopores of silica gel of 14 nm in pore width at an equilibrium pressure at 276.1 K of 5.186 MPa, which is higher t h a n t h a t for the bulk hydrates [81]. The role of porous media in the formation of m e t h a n e h y d r a t e under the oceanbed is suggested [82], leading to the necessity of basic chemical research on m e t h a n e hydrate. ACF was used as the microporous adsorbent of the quasi-high pressure effect. The micropore volume and total surface area of ACF d e t e r m i n e d by high resolution nitrogen adsorption isotherm at 77 K are 0.95 ml (g-adsorbent) -1 and 1800 m 2 (g-adsorbent) -1. The average slit-pore width is 1.1 nm. CH4 was adsorbed on ACF having preadsorbed water in the micropore at 303 K under the subatmospheric pressure of CH4. As m e t h a n e is a supercritical gas at 303 K, only a small a m o u n t of m e t h a n e is adsorbed on ACF without application of high pressure above 101.32 kPa. The gas phase content of water is kept constant within + 5 % during the mixed gas adsorption; the adsorption i n c r e m e n t on introducing CH4 to the mixed gas system is attributed to the m e t h a n e adsorption. The volume fractional filling Ow of the pore volume by the preadsorbed w a t e r varied in the range of 0.2 to 0.65. The adsorption rate of CH4 on waterpreadsorbed ACF at 303 K upon introducing CH4 of different pressures was examined for 25 h. The adsorption proceeds rapidly and noticeably at the initial stage in the presence of the preadsorbed water. The adsorption increment reaches a m a x i m u m 1 to 2 h after introduction of CH4 and decreases gradually to reach a steady value after 10 to 25 h. This adsorption uptake is caused by the adsorption of only CH4 because of the above-mentioned gas analysis results.

653

Hence the adsorption increment upon introducing CH4 is expressed as the a m o u n t of adsorbed CH4 in the following descriptions. Figure 8 shows adsorption isotherms of m e t h a n e on the water-preadsorbed ACF (Ow = 0.40) and ACF at 303 K. Here we used the steady values as the a m o u n t of CH4 adsorption. The a m o u n t of CH4 adsorption is almost nil in the absence of the preadsorbed water, while a b u n d a n t CH4 of more t h a n 180 mg (g-adsorbent) -1 is adsorbed on the water-preadsorbed ACF even below 10 kPa. If we presume the liquid densities of m e t h a n e and water molecular assemblies in the micropore, the total filling of micropores with water and m e t h a n e is 0.85 for Ow = 0.40. When water is not preadsorbed, the amount of adsorption even at the CH4 pressure of 10 MPa on the same ACF is only 160 mg (g-adsorbent) -1. Thus, the presence of preadsorbed w a t e r gives rise to a r e m a r k a b l e e n h a n c e m e n t of CH4 adsorption at subatmospheric conditions. The m e t h a n e adsorption enhanced by the preadsorbed water at the subatmospheric pressure briefly corresponds to t h a t under 10 MPa. This noticeable uptake of CH4 in partially water-filled micropore indicates the presence of a strong interaction between w a t e r and CH4 molecules such as the clathrate formation. The cluster of water molecules having the solid-like structure should form a new type of n a n o h y d r a t e of CH4.

200

v~o

150

!00

50

0 0

5

I0

15

CH 4 pressure/kPa Figure 8. The adsorption isotherms of CH4 on ACF having preadsorbed water of the fractional filling ~w - 0.40 and ACF without preadsorbed water at 303 K. e: d~w= 0.40 and A: d~w-0.

7.

F U T U R E P O S S I B I L I T Y OF F L U I D C H E M I S T R Y IN N A N O S P A C E

The both of design of a fit porous system and e s t a b l i s h m e n t of the control method of supercritical gas should develop a new technology of Nanopore

654 Molecular Engineering in the area of gas separation, gas storage, gas purification, supercritical drying, catalysis [83] and adsorption heat pump etc. However, the nanopore molecular engineering needs a fundamental understanding of the interaction of a molecule with the solid nanopore much more and the state of the fluid molecular assembly in the confined space. It is believed that the critical temperature shifts to a lower temperature with the decrease of the pore width until mesopores [84]. Recent structural studies [85-88] show that fluid molecules form an organized structure in nanospaces. How must we consider the critical behavior in the micropore system? Can we apply the phase concept to such nanodimensional molecular assembly? We have no definite idea on this question. Still we have many important unresolved problems relating to fluid chemistry in the nanospace. A new technological application of fluid chemistry should stimulate this field.

REFERENCES 1.

.

3. 4. 5. .

.

8. .

i0.

II. 12. 13. 14. 15. 16.

J. Rouquerol, D. Avnir, C.W. Fairbridge, D.H. Everett, J.H. Haynes, N.Pernicone, J.D.F. Ramsay, K.S.W. Sing and K.K. Unger, Pure Appl.Chem.,

66 (1994) 1739. L.A. Taylor, Supercritical Fluid Extraction, Wiley Interscience, London,1996. K. Nishikawa and M. Takematsu, Chem. Phys. Lett., 226 (1994)359. K. Kaneko, J.Membrane Sci., 96 (1994) 59. R.M. Barrer, Zeolites and Clay Minerals, Academic Press, London,1978, chap.2. M.E. Davis, C. Saldarriage, C. Ontes, J. Garces and C. Crowder, Nature, 331(1988) 6158. K. Maeda, Y. Kiyozumi and F. Mizukami, Angew. Chem., 33 (1994) 2334. S. Inagaki, Y. Fukushima and K. Kuroda, J. Chem. Soc. Chem. Commun., (I 993) 680. C. T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli and J.S. Beck, Nature, 359 (1992) 710. J. S.Beck, J.C.Vartuli, W.J.Roth, M.E. Leonovicz, C.T.Kresge, K.D.Schmitt, C. T.-W.Chue, D.H.Olson, E.W.Sheppard, S.B.MacCullen, J.B.Higgins and J.L.Schelender, J. Am. Chem. Soc., 114 (1992) 10824. Q.Huo, D.I.Margolese and G.D. Stucky, Chem. Mater., 8 (1996) 1147. P.L.Llewellyn, Y.Grillet, F.Schfithe, H.Reichert and K.K.Unger, Microporous Mater., 3 (1994) 345. P.I.Ravikovitch, S.C.O.Domhnaill, A.V.Neimark, F. Schfith and K.K.Unger, Langmuir, II (1995)4765. S. Inoue, Y. Hanzawa and K. Kaneko, Langmuir, to be submitted. S. Iijima, Nature, 354 (1991) 56. S. Inoue, N. Ichikuni, N. Satake, T. Suzuki, T. Uematsu and K. Kaneko, J. Phys. Chem., to be submitted.

655 17. T. Kyoutani, Li-fu Tsai and A. Tomita, Chem. Mater., 7 (1995) 1427. 18. R. C. Bansal, J.-B. Donnet and F. Stoeckli, Active Carbon, Marcel Dekker, New York, 1988. 19. J. W. Patrick, Porosity in Carbons, Edward Arnold, London, 1995. 20. H. Marsh, Introduction to Carbon Science, Butterworths, London, 1989. 21. K. Kaneko, C. Ishii, M. Ruike and M. Kuwabara, Carbon, 30 (1992) 1075. 22. T. Otowa,R. Tanibata and M. Itoh, Gas Sep. Purif., 7 (1993) 241. 23. K. Kaneko and C. Ishii, Colloid Surf., 67 (1992) 203. 24. M. Sato, H. Isobe, H. Yamamoto and K. Kaneko, Carbon, 33 (1995) 1347. 25. R. W. Pekala, C.T. Alviso, F.M. Kong and S.S. Hulsey, J. Noncryst. Solids, 145 (1992) 90. 26. Y. Hanzawa, K. Kaneko, R.W. Pekala and M.S. Dresselhaus, Langmuir, 12 (1996) 6167. 27. H. Sakuno, A. Ogawa, N. Akuzawa and Y. Takahashi, Tanso, (1990) 238. 28. T.J. Pinnavaina and M.F. Thorpe, Access in Nanoporous Materials, Plenum Press, New York, 1995. 29. S. Yamanaka, T. Doi, S. Sako and M. Hattori, Mat. Res. Bull., 19 (1984) 161. 30. M. Chikazawa and T. Takei, Silica Gel and Porous Glass, in: Microporous Soids and Adsorption (Japanese), K. Kaneko (ed.), Tokyo Kagaku Dojin, Tokyo, 1995, chap. 16. 3 I. T.T. Borek, W. Ackerman, D. W. Hua, R. T. Paine and D. M. Smith, Langmuir, 7 (1991) 2844. 32. S. Kawata, S. Kitagawa, M. Kondo and M. Katada, Synth. Metals, 71 (1995) 1917. 33. W.A. Steele, Surface Sci., 36 (1973) 317. 34. K. Kaneko, R. Cracknell and D. Nicholson, Langmuir, I0 (1994) 4606. 35. S.J. Gregg and K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, London, 1982. 36. A.W. Adamson, Physical Chemistry of Surfaces, John Wiley & Sons, 1990, New York, chap. 16. 37. K. Kaneko,K. Shimizu and T. Suzuki, J. Chem. Phys., 97 (1992) 8705. 38. M. Jaroniec and R. Madey, Physical Adsorption on Heterogeneous Solids, Elsevier, Amsterdam, 1988. 39. K. Kaneko, Adsorption on New and Modified Inorganic Sorbents, A. Dabrowski and V.A. Tertykh (eds.), Elsevier, Amsterdam, 1996, chap. 2.10. 40. P. B. Balbuena and K.E. Gubbins, Fluid Phase Equilibria, 76 (1992) 21. 41. D. Nicholson, J. Chem. Soc. Faraday, 92 (1996) I. 42. M.J. Bojan and W. A. Steele, Characterization of Porous Solids IV, T. May and B. McEnaney (eds.), Royal Society, London, in press. 43. M.M. Dubinin and H.F. Stoeckli, J. Colloid Interface Sci., 75 (1980) 34. 44. K. Kaneko, T. Ohta, S. Ozeki, N. Kosugi and H. Kuroda, Appl. Surf. Sci., 33/34 (1988) 365. 45. T. Yamamoto, K. Hashimoto and A. Fujishima, Chem. Lett., to be submitted. 46. J. Billingsley and D.B. Callerar, J. Chem. Soc., 67 (1971) 589.

656 47. K. Kaneko, N. Fukuzaki and S. Ozeki, J. Chem. Phys., 97 (1987) 776. 48. K. Kaneko, N. Fukuzaki, K. Kakei, T. Suzuki and S. Ozeki,Langmuir, 5 (1989) 960. 49. K. Kaneko, A. Kobayashi, T. Suzuki, S. Ozeki, K. Kakei, N. Kosugi and H. Kuroda, J. Chem. Soc. Faraday Trans. I, 84 (1988) 1795. 50. K. Kaneko and A. Matsumoto, J. Phys. Chem., 93 (1989) 8090. 51. K. Kaneko, Langmuir, 3 (1987) 357. 52. K. Kaneko, Colloid Surfaces, 37 (1989) 115. 53. K. Kaneko, in: Characterization of Porous Solids, K.K. Unger, J. Rouquerol, K.S.W. Sing and H. Kral (eds.), Elsevier, Amsterdam, 1988, 183. 54. H. Uchiyama, S. Ozeki and K. Kaneko, Chem. Phys. Chem. , 166 (1990) 531. 55. K. Kaneko, N. Inoue and T. Ishikawa, J. Phys. Chem., 93 (1989)1988. 56. N. Uekawa, M. Watanabe and K. Kaneko, J. Chem. Soc. Faraday Trans., 91 (1995) 2161. 57. Z.M. Wang, T.Suzuki, N.Uekawa, K.Asakura and K.Kaneko, J.Phys.Chem., 96 (1992) 10917 . 58. K. Jellinek, Anorg. Allg. Chem., 49 (1906) 229. 59. M. Shelef, Chem. Rev., 95 (1995) 209. 60. M. Iwamoto, H. Furukawa, Y. Mine, F. Uemura, S. Mikuriya and S. Kagawa, J.Chem. Soc. Chem. Comm., 1986, 1272. 61. J. R. Jr. Lancaster, Am. Sci., 80 (1992) 248. 62. S.F. Agnew, B.I. Swanson, L.H. Jones and R.L. Milles, J. Phys. Chem. , 89 (1985)1678. 63. J. Imai M. Souma, S. Ozeki, T. Suzuki and K.Kaneko, J. Phys. Chem., 95 (1991) 9955. 64. J. Imai and K. Kaneko, Cat. Lett., 20 (1993) 133. 65. Y. Nishi, T. Suzuki and K. Kaneko, J. Phys. Chem., I01 (1997) 1937. 66. K.R. Matrang, A.L. Myers and E.D. Glandt, Chem. Eng. Sci., 47 (1992) 1569. 67. J. Jagiello, T.J. Bandosz, K. Putyera and J. A. Schwarz, J. Chem.Eng. Data, 40 (1995) 1288 . 68. R.F.Cracknell, K.E. Gubbins, M.Maddox and D.Nicholson, D. Account Chem.Res., 28 (1995) 281. 69. B.P.Bojan,E. Cheng, M.W.Cole and W.A.Steele, Adsorption, 2 (1996) 57. 70. K. Kaneko, K. Murata, K. Shimizu, S. Camara and T. Suzuki, Langmuir, 9 (1993) 11651. 7 I. K. Kaneko and K. Murata, Adsorption, 3 (1997) 197. 72. K. Murata and K. Kaneko, J. Chem. Soc. Faraday Trans., to be submitted. 73. T. Iiyama, K. Nishikawa, T. Otowa and K. Kaneko, J. Phys. Chem., 99 (1995) 10075. 74. T. Iiyama, K. Nishikawa, T.Suzuki and K. Kaneko, Chem. Phys. Lett., 274 (1997) 152. 75. E.A.Mtiller, L.F.Rull, L.F.Vega and K.E.Gubbins, J. Phys. Chem., I00 (1996) 1189. 76. K. Koga, X.C. Zeng and H. Tanaka, Phys. Rev. Lett., in press.

657 77. K. Fujie, S. Minagawa, T. Suzuki and K. Kaneko, Chem. Phys. Lett., 236 (1995) 427. 78. K. A. Sparks and J.W. Texter, J. Phys. Chem., 96 (1992) 11022. 79. J. Miyawaki, T. Kanda, T. Okui, M. Saito, T. Suzuki and K. Kaneko, J.Phys.Chem., in press. 80. E.D. Sloan Jr.,, Clathrate Hydrate of Natural Gases, Marcel Dekker, New York, 1990. 81. F.Fleyfel, K. Y.Song, A.Kook, R.Martin and R. Kobayashi, J. Phys. Chem., 97 (1993) 6722. 82. Y.P Handa and D.Stupin, J. Phys. Chem., 96 (1992) 8599. 83. G.-C. Shen, A.-M. Liu, T. Shido and M. Ichikawa, Topics in Catalysis, 2 (1995) 141. 84. M. E. Fisher and H. Nakanishi, J. Chem. Phys., 75 (1983) 5857. 85. Z. M. Wang and K.Kaneko, J. Phys. Chem., 99 (1995) 16714. 86. H. Kanoh and K. Kaneko, J. Phys. Chem., 100 (1996) 755. 87. K. Kaneko, Colloid Surface, 109 (1996) 319. 88. N.Setoyama, K. Kaneko and F. Rodriguez-Reinoso, J. Phys. Chem., 100 (1996) 10331.

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

659

Application of natural adsorbents and adsorption-active materials based thereon in the processes of water purification Yu.I.Tarasevich Institute of Colloid C h e m i s t r y and Chemistry of Water, U k r a i n i a n National Academy of Sciences, 42 Vernadsky avenue, Kyiv (Kiev), 252680, Ukraine 1.

INTRODUCTION

N a t u r a l sorbents are increasingly widely used in w a t e r purification processes due to their frequent occurrence in nature, low cost as compared with synthetic materials, r e m a r k a b l e properties with respect to adsorption, ion exchange and filtration, and also due to the fact t h a t efficient methods are developed to regulate the geometric structure or these materials and chemical n a t u r e of their surface [1,2]. Most i m p o r t a n t steps in w a t e r purification are the removal of disperse impurities and the elimination of substances dissolved as molecules and ions. The present chapter d e m o n s t r a t e s t h a t n a t u r a l sorbents of various types, as well as adsorption-active materials based thereon, can readily be used for the removal of three above-mentioned groups of impurities from water. The structure and properties of a n u m b e r of n a t u r a l and modified sorbents and their application as filtering materials, adagulants, ion exchangers, and the sorbents proper are discussed in this communication. 2.

S T R U C T U R E A N D P R O P E R T I E S OF N A T U R A L S O R B E N T S

Three m a i n types of n a t u r a l sorbents can be distinguished with respect to the features of their structure, chemical composition, porosity and physicochemical properties: silica, layer and layer-ribbon silicates, framework aluminosilicates (zeolites). 2.1. D i s p e r s e s i l i c a These are the m a t e r i a l s of s e d i m e n t a r y origin which comprise 60-95% of amorphous SiO 2. Three types of siliceous rock can be distinguished: diatomites, tripolies and silica clays. They differ from each other in origin and physical and physicochemical characteristics. Diatomite, sometimes still called kieselguhr or infusorial earth, is of phytogenic origin. It consists of fossilised residues of simplest unicellular organisms - shells of diatomaceous algae (diatoms). Their

660 n u m b e r in 1 cm 3 of diatomite varies from 2 to 30 million, depending on particular deposit [3]. Table 1 s u m m a r i s e s the data of our publications [1,4] concerning chemical composition and structural-sorption characteristics of diatomites from various deposits. Specific surface area of samples S was determined by nitrogen t h e r m a l desorption; average effective macropore radius rma and total pore volume V~ were estimated from mercury porosimetry.

Table 1 Chemical composition and structural-sorption characteristics of various diatomites Chemical composition, % Diatomite Inzensk (Russia) Kisatibi (Georgia) Nurnus (Armenia)

SiO 2

A1203 Fe203 MgO

CaO

Adsorption properties Calcination S, V~, rma, lOSS m2/g cm3/g ~m

82.66 4.55

3.21

1.23

0.47

4.78

17.2

1.03

0.49

94.14 1.98

0.28

0.17

0.85

2.93

14.5

1.60

0.66

95.11 0.15

0.23

0.20

0.60

3.52

12.0

2.15

0.70

Djradzor (Armenia)

87.90 6.30

1.64

0.52

1.20

1.64

10.5

2.40

1.00

Batkovitsy (Bulgaria)

81.80 9.40

2.90

1.00

1.20

3.80

11.8

1.70

1.58

Lompok

89.30 4.00

0.70

0.40

0.40

5.00

16.0

-

(USA)

The results of chemical (Table 1) and X-ray analysis show t h a t diatomite is a h y d r a t e d amorphous SiO2 with small admixture of metal oxides. This sorbent is characterised by large n u m b e r of macropores (V• = 1.0-2.4 cm3/g) with reff = 0.5-1.5 ~tm. Mesopores with p r e d o m i n a n t pore radius 20-60 nm also exist in diatomite structure, and contribute ca. 10-15% to the total pore volume [1]; therefore diatomites are mainly macroporous formations. The porosity of the material is calculated from the equation: P(%) = 100(d t _ y) / d t

(1)

where d t and Y are the true density and bulk density, respectively. The density of amorphous silica is d t - 2.2 g/cm3; bulk density of diatomite is y = 0.4-0.5 g/cm 3 in

661 piece and 0.27-0.30 g/cm 3 in powder. For best grades of diatomite the porosity can range up to 90-92%. Tripoly usually exists in the form of weakly cemented, very light and porous rock, quite similar to diatomite in its external appearance, but r a t h e r distinguishable from it under the microscope. Tripoly has predominantly mineral origin and consist of round particles of opal silica 1-2 ~m in diameter. Bulk density of tripoly powder varies in the range 0.5-0.8 g/cm3; its porosity is lower t h a n t h a t of diatomite, and amounts to 60-70%. Table 2 s u m m a r i s e s the data [1,4] concerning structural-sorption characteristics of tripolies taken from selected industrial deposits: specific surface area determined from benzene adsorption (assuming C~H 6 molecular area co - 0.4 nm2), limiting pore volume V s with respect to benzene vapour, total pore volume V~ and effective mesopore rme and macropore rma radius calculated from mercury porosimetry data. Chemical composition of these samples was: SiO 2 65-82%, A1203 - 5 - 1 5 % , Fe203 -2-3.5%, M g O - 0.3-0.7%, C a O - 1-2%, Na20 and K20 - 0.6-2.0%, calcination loss - 6-9.5%.

Table 2 Structural-sorption characteristics of tripolies and silica-clays Adsorbent Tripoly Pionersk (Russia) Kirovograd (Ukraine) K a m e n s k (Moldova) Silica-clay Zikeevka (Russia) K a m e n n o y a r s k (Russia) Saratov (Russia)

S, m2/g

rme, nm

V s, cm3/g

V z, cm3/g

rma, ~m

33 45 75

6.2 5.4 7.2; 15.2

0.14 0.15 0.16

0.74 0.65 0.70

1.0 0.70 1.6

90 80 88

5.8 3.2 2.4; 15.8

0.33 0.25

0.43 0.44 0.44

0.50

The analysis of the results presented in Table 2 shows t h a t macropores with rma~ 1.0 ~m are also present in the structure of tripoly. However, it can be concluded from the value of total pore volume that the n u m b e r of these macropores is lower, as compared with diatomite structure. In fact, average fraction of macropores in the total pore volume of tripoly is 70-75% (cf. with 85-90% in the case of diatomite). The remainder of sorption space volume corresponds to the porosity, for which pores with radius of approximately 5-7 and 15-65 nm [5,6] are mainly responsible. Silica-clays are regarded in the geological literature as siliceous rocks accompanying tripoly and diatomite. Silica-clays are h a r d light porous materials, with colour varying from pale-grey to greenish-black or black. Dark colour of

662 silica-clays is strongly indicative for the presence of impurities, namely iron oxides (3-5%) and organic substances. Usually these materials are comprised mainly of finest particles of opal silica; in some cases they include significant amount of diatom shells which, however, are much more damaged than in diatomites. Persistent components of silica-clays are quartz, mica, feldspar, and glauconite. The advantage of silica-clays over tripolies and diatomites is their greater Mohs' scratch hardness (3-6 according to [3]) and nonsoftening in water. Nevertheless, it is sometimes difficult to distinguish between silica-clay and tripoly; therefore it was agreed that rocks with a bulk density exceeding 1 should be referred to as silica-clays. High density of silica-clays is associated with their lower porosity. The total pore volume of these materials does not exceed 0.45 cm3/g (Table 2). Less than 50% of the total porosity of silica-clays goes for the fraction of mesopores, among which the pores predominate with r = 2-6 nm. Significant role of mesopores in what regards structural-sorption properties is indicated also by the increased (about 100 m2/g) specific surface area inherent in them (Table 2). Table 3 summarises structural-sorption properties of natural silica.

Table 3 Structural-sorption characteristics of natural silica

Vma/VE,%

Silica

S, m2/g

Vz, cm3/g

100

Diatomite Tripoly Silica-clay

10-20 30-70 80-100

2.4-1.0 0.8-0.6 _<0.45

95-90 80-70 _<50

2.2. Layer and layer-ribbon silicates With respect to features of porous structure, layer and layer-ribbon silicates can be divided into three types [7]: layer silicates with expanding structural cell (montmorillonite, hectorite, saponite, vermiculite), layer silicates with rigid structural cell (kaolinite, hydromica, glauconite, pyrophyllite, talc), and layerribbon silicates (palygorskite and sepiolite). The physicochemical characteristics of individual minerals of Ukrainian and Russian (vermiculite) industrial deposits (the capacity of cation exchange E at pH 7; heat of wetting with water for samples dehydrated at 200~ Q; specific surface area S determined from the adsorption of hexane and water vapours assuming their molecular areas to be 0.5 and 0.108 nm 2, respectively; limiting volume of pores V s and total volume of secondary pores V~; particle size d estimated from electron microscopy data) are presented in Table 4.

663 Table 4 Physicochemical characteristics of layer and layer-ribbon silicates Sorbent, deposit Montmorillonite, Pyzhevsk Montmorillonite, Cherkasy Vermiculite, Kovdor Kaolinite, Glukhovets Kaolinite, Glukhov Hydromica, Cherkasy Glauconite, Karachievo Palygorskite, Cherkasy

E, Q, mg-equiv/g J/g

Hexane

Water

S, V s, m2/g cm3/g

S, V s, m2/g cm3/g

VE, cm3/g

d, ~m

1.05

149

36

0.05

428

0.37

0.12

0.05-0.5

0.71

101

60

0.07

311

0.30

0.13

0.05-0.3

1.60

194

14

0.03

471

0.18

-

-

0.01

5.5

10

-

11

0.04

-

1.0-1.5

0.25

40

60

0.17

94

0.23

0.36

0.02-0.5

0.27

59

125

0.25

157

0.25

-

0.05-0.1

0.14

-

54

0.07

112

0.13

0.26

156

153

0.29

302

0.45

0.48

0.01-0.2

Layer silicates with expanding structural cell are characterised by high adsorption ability with respect to water and other polar s u b s t a n c e s - alcohols, amines, nitriles [7] (Table 4). In the process of adsorption of these substances, structural cells of these minerals increase along C axis by 0.3-1.4 nm and their interlayer spaces can be regarded as slit-shaped micropores whose dimensions undergo changes during adsorption process. In addition to primary microporosity inherent to their crystal structure, montmorillonite and vermiculite also possess secondary porosity, mainly due to mesopores formed by spaces between contacting particles. Measurements of adsorption of hexane and other non-polar substances show that the surface area of secondary pores of layer silicates with an expanding structural cell is much lower than that of primary micropores. Layer silicates with rigid structural cell are characterised by the presence of external adsorbing surface only, and their porosity is due to spaces between contacting particles. The specific surface area of minerals for this group of layer silicates is determined mainly by the degree of dispersion of particles (see data in Table 4 for Glukhovets and Glukhov kaolinite), which in t u r n depends on the perfection of mineral crystal structure. Adsorption properties of layered silicates palygorskite and sepiolite are determined both by zeolite channels which exist in mineral structure and possess a size of 0.37x0.64 and 0.37x1.10 nm respectively (primary pores), and by porous

664 space of stacks into which needle-like or fibrous mineral particles are aggregated (secondary porosity). The surface and volume of secondary pores of these minerals are quite large, which results in high adsorption properties with respect to hydrocarbons. Layer silicates also possess the ability for cation exchange. Among them, montmorillonite and vermiculite are characterised by largest cation exchange capacity, E = 1.0-1.6 mg-equiv/g. We have studied the nature and quantity of various types of active centres, which exist at the surface of the main representatives of layer and layer-ribbon silicates [8]. Obtained results are generalised in Table 5.

Table 5 Active centres of layer and layer-ribbon silicates Sorbent,

SiO-Me + Total

mmol/g

A1-OH, Mg-OH, mmol/g

1.45

< 0.17

1.60

0.13

0.03

0.015

0.90

0.15

1.05

0.17

0.07

0.04

0.17

0.10

0.27

0.42

0.18

0.22

0.03

0.25

0.07

0.07

0.06

0.14

0.12

0.26

0.41

0.17

0.15

Exchangeable cations, mg-equiv/g

deposit Vermiculite, Kovdor Montmorillonite, Pyzhevsk Hydromica, Cherkasy Kaolinite, Glukhov Palygorskite, Cherkasy

Isomorphism

Si-OH,

Unsaturated A13+,Mg2+, mmol/g

In Pyzhevsk montmorillonite which contains virtually no isomorphous substitutions of silicon by aluminium according to measurements of the exchange capacity of the Li form of the mineral before and after heating at 300~ [9], we have managed to determine separate contributions of isomorphous substitutions of aluminium by magnesium in octahedral layers and o f - S i O - H § groups which exist at crystal edges, into total cation exchange capacity. For palygorskite, the quantities of the exchangeable cations due to heterovalent isomorphism in the structure and -SiO-H § groups was assessed from conductometric titration curves for NH 4- and Cu- forms of the mineral by alkali [8]. Study of ion-exchange sorption of tetra-ammine-copper ion at NHn-vermiculite enables one to calculate the quantity of cation-exchange centres which exist at external surfaces of this mineral. The quantity o f - SiO-H § groups at vermiculite taking part in cation exchange obviously does not exceed this value. The number of exchangeable -SiO-H § groups at Glukhov kaolinite E ~ 0.03 mg-equiv/g was determined from

665 empirical equation proposed by Brindley [10]. We have observed in [8] that the exchange capacity of Li-saturated Glukhov kaolinite decreased to approximately the same value when the sample was heated to 300~ Acidic - S i O H groups at dry hydromica and palygorskite surfaces were determined using the reaction [11]: -SiOH + CH3MgI -->--Si-O-MgI + CH 4

(2)

The quantities of acidic -SiOH groups existing at other mineral surfaces, and the number of basic =A1OH a n d - M g O H groups present at all surfaces were found by reverse conductometric titration of aqueous suspensions of the sorbents by alkali and acid [12]. The concentration of co-ordination-unsaturated A13§ and Mg 2§ ions at the external surfaces was given by the amounts of pyridine and acetonitrile strongly held by Rb- or K- forms of the mineral when heated to 150200~ The data showh in Table 5 indicates the different roles played by various active centres in determining the surface chemistry of individual types of layer and layer-ribbon silicates. Thus surface properties of palygorskite are largely determined by co-ordination-unsaturated Mg 2§ and A1~+ cations, and also by numerous surface SiOH groups. Studies of infrared spectra [13] indicate that these groups are rapidly exchanged for deuteroxyl groups when exposed to heavy water vapour. Organoderivatives of palygorskite and sepiolite [14,15] can be readily obtained because of the high reactivity of surface hydroxyl groups. The main active adsorption centres of expanding layer silicates montmorillonite and vermiculite are the exchangeable ions. It was shown in [7] that the interaction of adsorbed water and other polar molecules with the exchangeable cations involves the formation of surface co-ordination compounds. In elucidating the structure of such complexes, optical electronic spectroscopy in the ultraviolet, visible and near infrared regions and electronic paramagnetic resonance spectroscopy are informative [8]. As a result of the polarisation of OH group, water molecules co-ordinated by multiple-charge exchangeable cations in layer silicates possess pronounced acidic properties. This is indicated by pyridinium ion formation when pyridine is adsorbed at partially hydrated montmorillonite and kaolinite [7]. The acidity of water molecules firmly bound to the exchangeable cations of layer silicates may be estimated quantitatively by studying the electronic spectra of sorbed bases, i.e., H a m m e t t indicators, with different acid-base equilibrium constants. For example, it was shown that on the surface of Mg-kaolinite partially dehydrated by evacuation at 100~ there exist active centres possessing the acidity equal to, or higher t h a n that created by 71% H2SO 4 [8]. Water molecules polarised in the field of multi-charge exchangeable cations play a major role in catalytic reactions, which take place at layer silicate surfaces.

666 2.3. Z e o l i t e s Natural zeolites have been known for a long time [16]. More t h a n 40 types have already been discovered, but until 1960s the investigators were dealing in practice with specimens of zeolites of volcanic origin which are found in small quantities and are thus of no interest from the point of view of practical utilisation. However, when large sedimentary deposits of natural zeolites were discovered in the late 1960s and early 1970s in various regions of the world, zeolites become of interest to engineers due to their unique adsorption, ion exchange, filtration and catalytic properties, their ease of quarrying and low cost.

Table 6 Characteristics of natural zeolites Zeolite

Structural formula

Channels in hydrated structure dt, Direction n* CrossVm~, g/cm3 section, cm3/cm~

7"*

E, mgequiv/g

nin

Clinoptilolite

Na6[(A102)6(SiO2)30]. 24H20

(!00) (001) (001) Morde- Nas[(AIO2)8(SiO2)40]" (001) nite 24H20 (010) Phillip- (K,Na)5[(A102)5(SiO0,~]. (100) site 10H20 (010) (001) ChabaCa2[(AIO2)4(SiO2)s]. • zite 13H20 2_(001) Erionite (Ca, Mg, Na2, K 2 ) 4 . 6 9 [(A102)9(SiO2)27]-27H20

8 10 8 12 8 8 8 8 8

0.40• 0.44• 0.41x0.42 0.67• 0.37• 0.42x0.44 0.28x0.48 0.33 0.36x0.37

0.34

2.16

3.5-4 2.6

0.26

2.13

3-4

0.30

2.122.24

4-4.5 4.7

0.48

4.5

5.0

8

0.36•

0.36

2.052.1 2.02

-

3.8

2.6

* n - Number of tetrahedra in rings, * ' 7 - Mohs' scratch hardness.

Industrial deposits of eight zeolite minerals are now known: clinoptilolite, mordenite, phillipsite, chabazite, erionite, ferrierite, laumontite and analcime [17]. The last three zeolites still have found no practical use. The territory of Ukraine possesses large industrial deposits of clinoptilolite, mordenite and analcime, including one of the world's largest deposits of clinoptilolite tufts in T r a n s c a r p a t h i a n Ukraine. Table 6 summarises information [16-18] concerning chemical composition, structure and properties of natural zeolites of interest in practical utilisation both due to their availability in nature and their physicochemical characteristics. Zeolites possess negatively charged three-dimensional aluminosilicate framework. Hydrated exchange cations N a +, K +, C a 2+, Mg2+ are located in

667 primary m i c r o p o r e s - cavities (channels) of this framework and neutralise its negative charge. Theoretical exchange capacity of natural zeolites varies in the range of 2.6-5.0 mg-equiv/g (see Table 6); actual exchange capacity is usually lower. This is associated with the fact that a portion of K § and Ca 2§ cations are located in non-exchange positions, and also with the presence of non-zeolite impurities in the samples of natural ion exchangers. As an example, the clinoptilolite-rich tuff from Sokirnitsa field in Transcarpathian Ukraine, which contains 60-70% zeolite was used in our studies. The chemical composition (wt. %) of the tuff was as follows: SIO2-67.29; TiO2 - 0 . 2 6 ; A120 ~ - 12.32; Fe20 ~ - 1.26; FeO - 0.25; MgO - 0.99; CaO - 3.01; Na20 - 0.66; K20 - 2.76; H 2 0 - 1 0 . 9 0 ; t o t a l - 9 9 . 7 0 ; Si:A1 r a t i o - 4 . 6 3 . The tuff has a cation-exchange capacity of 1.44 mg-equiv/g. The exchangeable cation composition (mg-equiv/g) was: N a * - 0 . 2 1 ; K § - 0 . 2 2 , and Ca 2§ 1.01. Diameters of micropores in natural zeolites are about 0.5 nm (see Table 6). The micropore volume of zeolites Vmi (cm3/cm 3) was determined from the quantity of zeolite water. In addition to primary microporosity determined by the crystal structure, natural zeolites also have secondary packing pores formed by space between microcrystals due to their non-dense packing in the aggregates. Specific surface area of secondary pores for clinoptilolite tuff from Transcarpathian Ukraine is S = 20 m2/g, as determined from benzene adsorption data. Molecules of this substance do not penetrate into zeolite channels of clinoptilolite and are adsorbed on the surface of secondary pores only. Total volume of secondary pores for clinoptilolite from Sokirnitsa deposit is V~=0.110cm3/g, constituted by the volume of macropores Vma = 0.066cm3/g (mercury porosimetry data), mesopores Vine- 0.037 cm3/g (benzene adsorption data), and secondary micropores Vmi-- 0.007 cm3/g (also estimated from benzene adsorption data). From the data obtained by mercury porosimetry, the effective radius of macropores ref= 1000nm and 3 5 0 n m (main), and mesopores ref- 27.5 nm and 11 nm (main) in clinoptilolite were estimated. 2.4. Other n a t u r a l s o r b e n t s Other natural sorbents used in the practice of the purification of drinking, technical and waste water are perlite, asbestos, bauxite, and magnesite. Perlite is the volcanic glass rock consisting mainly of obsidian. The chemical composition (wt. %) of perlite from one of the world's largest Aragats industrial deposit (Armenia) is as follows: S i 0 2 - 76.0; A120 3 - 14.0; Fe203 - 0 . 7 ; C a O - 1.0; MgO-0.2; N a 2 0 - 3.9; K 2 0 - 4.6; calcination l o s s - 3 . 2 . The chemical composition (wt. %) of perlite from Beregove industrial deposit (Ukraine) is: SiO 2 - 74.7; A120 3 - 12.6; Fe20 3 - 1.2; C a O - 1.2; M g O - 0 . 1 ; Na20-2.5; K20 - 4 . 0 ; calcination l o s s - 4.1. Both these perlites are quite similar in chemical composition to American perlite "Sil-Flo" [19]. Bulk density of natural non-Bloated perlites is 1.40-2.22 g/cm 3. The porosity of the best sorts amounts to 30-40%. The specific surface area of non-bloated perlite determined from the adsorption of benzene and nitrogen is 1.5-2.0 m2/g, while

668 with respect to water adsorption this area is of the order of 10 m2/g [4]. Perlite is a heterogeneously porous material, whose structure contains macro-, meso- and micropores. These latter are easily accessible to water molecules and, in a lower d e g r e e - to nitrogen and hydrocarbon molecules. Among natural dispersed materials, asbestos begins to acquire ever increasing value. This species can be subdivided into two g r o u p s - amphibolic and chrysotilic, which differ in composition, structure and properties. World resources of amphibolic asbestos are limited and cannot meet constantly increasing needs of industry. At present, the output of amphibolic asbestos constitutes only 5-10% of the total quarry of asbestos fibres, and the efforts of scientists are concentrated on the development of the technology for the preparation of synthetic amphiboles [20,21]. The shortage and relatively high cost of amphiboles h a m p e r their use in water purification processes. Chrysotilic asbestos is more promising in this respect. It was shown [22] that despite the fibrous texture of chrysotile (idealised chemical formula H4Mg3Si2Og) its structure is similar to that of kaolinite. With respect to adsorption properties, chrysotile together with halloysite should be regarded as an individual subgroup of layer silicates. In fact, the tubular shape of primary halloysite and chrysotilic asbestos particles [22] is seen under the electron microscope. Internal radius of primary fibres of the latter mineral often amounts to 5.5 nm, while the external radius is 13 nm. Primary chrysotile particles in natural state are aggregated into thicker fibres, with usual length of 1.2-1.3 cm [22]. The external surface area of chrysotilic asbestos varies from 4 to 60 m2/g [22-24], and depends on the degree of dispersion. Tubular channels of chrysotile are filled to a significant extent with amorphous material or bandshaped mineral particles. According to calculations of [22], only 12-15% of the volume of chrysotile internal pores are free. Molecules of water, ammonia and other substances are adsorbed in the unoccupied space of the pores. Boehmite y-A1OOH, diaspore a-A1OOH and hydrargillite AI(OH) 3 are the rockforming minerals of bauxites. Iron oxides are virtually always present in bauxites, while kaolinite, calcite, rutile and other minerals are frequently encountered therein [25]. The chemical composition (wt. %) of bauxite from the industrial Tikhvin deposit (near St. Petersburg, Russia) is as follows [26]: A1203- 51.1; F e 2 0 3 - 16.1; SiO 2 - 14.1; C a O - 2.0; TiO 2 - 2.4; calcination l o s s - 14.3. The specific surface area of bauxite is - 5 0 m2/g, total pore volume ~- 0.2 cm3/g, effective pore radius 20 nm [26]. The main raw material used for industrial production of magnesium oxide is magnesite MgCO 3. Magnesium oxide is formed by the heating of magnesite to 550-600~ during this process the specific surface area of the material increases from 3 to 115 cm2/g [27]. The effective radius of intracrystallite pores of MgO obtained as a result of magnesite calcination at 550~ is 4 nm; the volume of these pores is 0.37 cm~/g. In addition to primary intracrystallite pores, secondary packing pores with r = 110 nm are present in the MgO structure [27]. Specific surface area of the sorbent decreases to 33 cm2/g and the radius of intracrystallite

669 pores increases to 18 n m [27] as a result of sintering of m a g n e s i u m oxide which occurs upon increasing its calcination t e m p e r a t u r e to 800~ N a t u r a l porous coals are very promising for the removal of organic impurities from drinking and waste water. The industrial seam of such coal is exploited in Donbas coal-mining region (Ukraine). We have studied the adsorption-desorption of hexane and w a t e r vapours on this coal [28]. The obtained results are shown in Figure 1. It is seen t h a t structure of this adsorbents contains both micro- and mesopores. Calculations show t h a t pore volume is equal to Vmi- 0.048 cm3/g, Vine(r-1.7; 3.3 n m ) - 0.109 cm3/g. According to mercury porosimetry data, this coal is characterised by macropores with volume V m a - 0.02 cm3/g and rma >_ 100 nm. The quantities of acidic (C = 260 ~g-equiv/g) and basic (C = 66 pg-equiv/g) groups on the coal surface were e s t i m a t e d from reverse conductometric titration of aqueous suspensions of the sorbent by alkali and acid.

a, mmol/g

a, mmol/g

I 0.8 k

/

8

i

0.4

.

,

~

0.4

0.8 (a)

P/Po

0.4

0.8

p/po

(b)

Figure 1. Adsorption (open circles)- desorption (filled circles) isotherms for (1) hexane and (2) water vapours at natural porous coal.

670

Q

U S E OF N A T U R A L S O R B E N T S FOR P U R I F I C A T I O N OF WATER FROM DISPERSED IMPURITIES

Dispersed i m p u r i t i e s - clay particles, humic substances, bacteria and viruses, finely emulsified oil and petroleum products are removed from water using filtration, coagulation, flocculation and flotation methods. In present section the application of natural sorbents in these processes is considered. 3.1. U s e o f n a t u r a l s o r b e n t s as f i l t e r i n g m a t e r i a l s in p r e c o a t e d filters One common method of water purification from dispersed particles is the filtration through a porous medium; natural sorbents are efficiently applied for this purpose. In particular, diatomite, tripoly, perlite, and chrysotile-asbestos are widely used as filtering materials in precoated filters [29], often called diatomaceous earth filters, because diatomite is the main filtering material used in precoated filters [30,31]. Various designs of these filters are considered in a number of books [39,31,32]. These filters are used to purify water in large municipal houses and swimming pools [29], for the tertiary treatment of waste water at petroleum refining plants [33], for the removal of radioactive suspended particles from waste water of nuclear power plants [34]. In Beykersfield (USA) water supply station was built by Boyle Engineering Company with daily output 95 000 m 3. The use of precoated filters at this plant makes it possible to reduce capital expenditures by approximately 35% [35]. Application of precoated filters had resulted in a better removal of disperse impurities from water as compared with grained filters. Their use at the concluding stage of premembrane t r e a t m e n t of low-turbidity ocean water reduces water-quality fouling index FI = SDI (silt density index) down to 0.3, while the use of cartridge polymers filters results in the decrease to 0.4-0.5 only [36]. Silt density index or fouling index characterises the presence of highly dispersed impurities in water fed to membrane treatment. Usually this p a r a m e t e r is calculated from the following expression [37,38]: 1 - tl/t 2 SDI = ~ 100 15

(3)

where t I is the time during which 500 cm 3 of the water under investigation is filtered through Millipore cellulose-acetate microfilter with pore diameter of 45 ~m under pressure of 210 kPa, and t 2 is the time during which next 500 cm 3 of water is filtered through this microfilter after elapse of time t I plus subsequent 15-minute filtration in adopted conditions. With proper intake (from a coastal well), seawater contains almost no suspended solids: in fact, the FI water-quality index decreases from 100-200 for ocean water intake to 6 for a coastal well [36]. However, seawater is characterised by the presence of phyto- and zoo-plankton (especially during warm seasons) and a significant amount of dissolved organic substances. According to

671 the d a t a p r e s e n t e d in [39], truly dissolved compounds account for 89% of organic m a t t e r in seawater. The presence of p h ytopla nkton and dissolved organic substances in s e a w a t e r results in the need for the p r e - t r e a t m e n t of w a t e r by filtration in steps, for example according to the scheme r e c o m m e n d e d in [36]: sand filter - carbon filter - cartridge ultrafilter. I m p l e m e n t a t i o n of this scheme reduces the FI w ater-q u ality index from 6 to 0.4-0.5. Additional t r e a t m e n t of s e a w a t e r on a precoated diatomite filter leads to a f u r t h e r reduction in FI value to 0.3 [36]. It is known that the structure of natural sorbents includes micro- (r__ 1.52 nm), meso- (r= 2-100 nm) and macropores (r > I00 nm). Virtually only the largest meso- and macropores of filtering materials participate in processes of purifying water from dispersed impurities. Therefore, to produce the most effective filtration powders natural diatomite and tripoly are calcinated at a temperature of 1000~ Small-size pores disappear due to sintering, and the radius of macropores and their number in the structure of amorphous silica increase. In addition, clay minerals (layer silicates) which are always contained as impurities in tripolies and diatomites during high-temperature calcination are converted into high-temperature non-swelling anhydrous phase. Finally the thermal treatment results in the transformation of hydrophilic silica into hydrophobic one. A characteristic feature of the thermal behaviour of silica which distinguish it from aluminium and iron oxides is the fact that rapid reversible rehydroxylation of the surface of this oxide is observed only for samples whose heat treatment temperature had not exceed 400~ [40,41]. Thus by means of heat treatment one can achieve a certain ratio of hydrophilic silanol to hydrophobic siloxane groups number at silica surface. In some cases filtering powders should have a chemically inert surface. In order to deactivate the surface of natural silica and, at the same time, to decrease the calcination temperature down to 900~ alkaline fluxes (usually soda ash) are introduced into raw silica material in a quantity of 2-5 mass % [42]. As a result, at high temperatures low-melting compounds like Na20-2SiO 2 (m.p. 821~ are formed, which coat the SiO 2 particles with a thin film and thus deactivate the impurity Al- and Fe-centres. To obtain more inert materials, the diatomaceous earth is treated with mineral acids before calcination; this results in the elimination of impurity aluminium and iron oxides. Natural perlite is also subjected to short-time thermal treatment at 1000~ to produce high quality filtration material (filterperlite). The bloating of perlite takes place as a result of rapid dehydroxylation and the emission of steam; this increases the initial volume of perlite by 15-20 times, decreases its bulk density to 80-149 kg/m 3 and increases the porosity to 85-90% [32]. The efficiency of filtration purification of water and other liquids from dispersed impurities depends greatly on the correct choice of powders. These must provide a high rate of purification with a necessary purity of the filtrate. High filtration rates are achieved using more coarsely dispersed powders possessing large-size pores. However, the application of these materials increases

672 the risk of colloidal particles breakthrough through the filter. Therefore it is necessary to use powders characterised by optimum pore dimensions which would provide a sufficiently fast rate of purification and necessary purity of the filtrate. In the USA more than 30 brands of filtering powders are produced on the basis of diatomite alone, which differ from each other in their degree of dispersion, chemical composition, porosity and other characteristics [19]. For example, Manville company which is exploiting a deposit of high-quality diatomites in California, produces nine commercial brands of filtering powders under the name "zelite". The filterzel powder contains ca. 70% particles finer than 10 pm, and only 2.5% of particles are larger than 40 pm. Precoated layer of this powder catches finest particles from water with size down to 0.1 pm. The comparatively coarsely dispersed powder zelite 545 contains 24% of fraction larger than 40 ~m and only 5.5% finer than 10 pm. US companies Dicalite and Eagle-Pitcher also produce filtering powders based on diatomite and perlite. US company Great Lakes Carbon Corp. produces filtering powders based on bloated perlite. Perlite powders outperform the diatomite ones in the permeability of pre-coating layer, but are inferior to them in what regards the fineness of the filtration. High-quality filtering powders based on diatomites and bloated perlites are also produced in Great Britain, France, Italy and Japan. In Russia, Inzensk filtering-powder plant (Ul'yanov province) produces grade A and B diatomites. Kirovograd plant (Ukraine) processes tripoly to produce filtering powders Kisel'gur No.1 and Kisel'gur No.2, fine- and coarse-dispersed respectively. A number of plants in Ukraine manufactures filterperlite. Some characteristics of filtering materials can be found in [19,31,43]. With such a wide choice of filtering materials which differ from each other in structure, chemical composition, degree of dispersion etc., the selection of the most appropriate one for any concrete process is not difficult. The efficiency of the filtration depends not only on the total volume of the macroporous space and the size of macropores but also on the form of pores present in porous material. Large options for regulating the form of pores in the precoated filtrated layer are provided by the blending of filter powders with various shape of particles, in particular diatomite or filterperlite with chrisotileasbestos. Blended filtration materials are already used in the practice of purification of liquids from dispersed impurities [44]. In order to obtain operating filtrated layer, the consumption of powder must be 200-800 g per 1 m 2 surface of supporting filtering elements. The quantity of auxiliary (filter-aid) materials introduced into treated water is usually equal to the mass of dispersed admixture. However the consumption of diatomite 800 g per 1 m 2 and three- or fourfold excess of filter-aid materials is necessary to provide high quality of drinking water purification [35]. Precoated filters enable one to purify water not only with respect to dispersed admixtures, but also to substances dissolved as molecules and ions. To regulate ion and molecular composition of water, finely ground vermiculite, glauconite and

673 active carbon are used in the compositions of precoated powders. Such filters efficiently adsorb organic substances, silicon and iron compounds, etc., from water [43]. In view of extensive use of membrane methods, it is especially important to remove iron compounds at the stage of premembrane water treatment [45]. Precoated filters which employ diatomite, bloated perlite, activated carbon, etc. as auxiliary filtering materials, are already used for rather long time to remove iron from water [46]. However, the quality of such premembrane treatment is insufficient, therefore the method of iron removal from water by passing through a ceramic cartridge at which ferric hydroxide was preliminary deposited has become widespread. The layer of Fe(OH) 3 which grows in the cartridge acts as an active element for adhesion flocs of ferric hydroxide forming as a result of oxidation Fe(II) ~ Fe(III). In connection with the importance of the removing of iron compounds from water, we have paid attention to glauconite as a filter-aid material [45]. This natural mineral was formerly rather widely used in water treatment to remove disperse impurities from water [29]. Two-layered grained filters of crushed anthracite and glauconite are used efficiently at industrial water treatment stations in Canada to remove iron and manganese compounds from water preliminary treated with soda and slaked lime (to bring water pH value up to 9) and then chlorinated [47]. At the territory of Ukraine, a large commercial deposit of glauconite sand exists; its physicochemical characteristics are presented in Section 2.2, Table 4. Grains (pellets) of glauconite in its natural state are covered with a film of iron hydroxides. The latter, being genetically connected with the rock-forming mineral, are tightly held at the grains of glauconite sand. This makes it unnecessary to create a modifying hydroxide layer on the surface of grains of the filtering material for removing iron from water, which is a fairly long and tedious operation. Film of ferric hydroxide plays an important role in the autocatalytic oxidation reaction Fe(II)--> Fe(III) and in the adhesion of flocs Fe(OH) 3 on glauconite particles. The use of glauconite as a filter-aid material reduces iron content in the water to c < 0.1 mg/dm 3 [45]. It is seen from Table 7 that precoated perlite filter which employs diatomite and glauconite mixture as the filter-aid material (with mass ratio of mixture

Table 7 Extraction of some radioisotopes (%) from model solutions using precoated perlite filter Auxiliary material Filterperlite Diatomite/glauconite mixture Diatomite/glauconite mixture

pH

54Mn

7 7 9

64 80 83

6~ 79 75 70

137Cs 27 98 95

89+9~ 0 72 78

674 components 1:1) purifies radioactive waste with respect to 54Mn and 6~ compounds in dispersed state and to dissolved ions la7Cs and sg+9~ [48]. The removal of radioactive ions laTCs and sg+9~ from water takes place due to the application of glauconite, the natural ion exchanger.

3.2. Use of grained clinoptilolite and other grained materials in water purification Natural materials for water purification are used most frequently in grained form. Quartz sand is the traditional material for grained filters. However, an insufficient intergrain (packing) porosity of the sand, its poor adhesion characteristics relative to disperse impurities contained in the river water, high consumption of water for its washing and other factors have led to a situation when the filtration unit became the bottleneck of water purification technological scheme. Therefore other materials such as claydite, aggloporite, haydite, etc. are now being increasingly used instead of quartz sand [49,50]. Claydite and haydite are prepared by thermal shock method at 1000-1100~ from bentonite clays and argillaceous slates, respectively [51]. Aggloporite is produced by sintering of tripoly- and silica clay containing rocks, respectively. These materials possess higher intergrain (packing) porosity, as compared with quartz sand (Table 8).

Table 8 Physicomechanical characteristics of filter materials Material Quartz sand Clinoptilolite Claydite Aggloporite Haydite Granite sand

Particle size, Bulk density, Density, Porosity, Mechanical strength, % abradability crushability mm kg/m3 g/m3 % 1.00 1.15 1.18 0.84 1.10 0.80

1500 900 780 1030 650 1660

2.46 2.20 1.91 2.29 2.18 2.72

37-45 58-53 48 54 65 46

0.20 0.32 0.17 0.20 0.90 0.11

1.25 2.40 0.36 1.50 4.90 1.40

Intergrain pores provide the retention of phytoplankton, disperse particles (clay minerals, humic substances etc.) and their aggregates, that is, the products of coagulation and flocculation. Moreover, to ensure efficient retention of colloidsize dispersions of SiO 2, Fe(OH)a, Fe (III) and A1 (III) complex compounds with fulvic acids etc., the structure of filtering material grains should possess internal macropores with effective radius r _> 100 nm [45]; however, there are no such pores in the structure of quartz and granite sand or aggloporite. The situation becomes more complicated, because along with principal purpose of filter

675 materials, t h a t is, to purify drinking water from substances responsible for its turbidity and colour, they should remove the m a x i m u m possible amount of normalised heavy metal, a m m o n i u m and aluminium cations, traces of radioactive impurities. This creates additional impediments for the use of quartz sand and other above-mentioned materials as the filter media. In our publications [1,52-56] the application of clinoptilolite as a filtering material instead of quartz sand for the purification of river water was shown to be promising. Clinoptilolite possesses sufficiently high cation exchange capacity and can adsorb heavy metal ions and other substances dissolved as cations in water. In structure of clinoptilolite there exist internal pores with r = 350 nm (see Section 2.3), which provide for the removal of colloidal and molecularly dissolved substances, in particular solubilised complexes of fulvic acids with cations A1 and Fe (III). Grains of T r a n s c a r p a t h i a n tuff are sufficiently strong, and do not slack in water. Clinoptilolite possesses good physicomechanical properties (Table 8). Crushed clinoptilolite-rich tuff (particle size 1-3 mm) was tested as the filtering material for coloured surface waters of low, medium and high turbidity. The results show t h a t the application of clinoptilolite had increased both the mud-trapping capacity of the filter by 25-45%, and the filtration rate from 6-8 to 8-10 m/hr without affecting the water quality. In addition, the amount of water required for filter backwashing was reduced by 15-20%, and the aluminium and iron content in cleaned water was reduced from 0.8-0.4 and 0.25-0.15 to 0.5-0.3 and 0.15-0.08 mg/dm 3, respectively, as compared with quartz sand filter [54]. Clinoptilolite filter provides for higher decrease of water turbidity and colour, especially during first 8-10 hr of the filtration cycle; the latter was increased by 4-16 hr. Zeolite filter more effectively removes phytoplankton. The data presented in Table 9 show that the use of the clinoptilolite filters at commercial water supply station (daily output 12 000 m 3) had resulted in a significant e n h a n c e m e n t of water purity. Average purification degree with respect to disperse impurities was 65.6% for clinoptilolite filter, while for reference filter it was only 45%. Results similar to those presented above were

Table 9 Purification of Uzh River water (Uzhorod City, T r a n s c a r p a t h i a n Ukraine) using commercial coal/quartz (A) and clinoptilolite (B) filters Month, 1991 July August September October November December

Suspended solids (mg/dm a) Initial A B 4.4 19.6 33.0 2.9 7.4 10.9

0.9 0.4 0.9 1.2 1.4 1.2

0.2 0.5 0.5 0.8 1.2 0.6

Aluminium (mg/dm a) Initial A B 1.00 1.87 0.65 0.60 0.72

0.45 0.81 0.38 0.29 0.35

0.33 0.47 0.22 0.24 0.27

676 obtained later by another investigators who used clinoptilolite from Tedzami deposit (Georgia) to purify Moskva river water [57]. Requirements to the quality of filtering materials used in premembrane t r e a t m e n t technology are continually increasing, in line with efforts of production engineers to combine coagulation and filtration (contact coagulation method [58]), or flocculation and filtration ("line coagulation" [59,60] or, more properly "line flocculation" method) in one unit. In both cases, the coagulant or flocculant is fed into the water to be treated immediately before granular filters. This leads to more thorough removal from water of disperse impurities, which are deposited directly in the filters intergranular space. Thus filtering materials for premembrane t r e a t m e n t should possess as high porosity as possible, and dimensions of secondary intergranular pores should be sufficient enough to ensure efficient capture of disperse particles and their a g g r e g a t e s - the products of coagulation and flocculation, and the retention of these species in pores. Pre-treatment of water supplied to membrane treatment according to the schemes: a) filtration through multilayered filters and b) coagulation - settling filtration with the use of quartz sand and anthracite as filtering materials, results only in the decrease of water quality index F I = SDI to 25 [59]. Replacement of quartz sand by clinoptilolite should undoubtedly reduce FI value. Nonetheless, to achieve the value FI = SDI = 3-5 required for normal operation of membranes, the final stage of pre-treatment should be the microfiltration through polymer cartridge microfilters with def- 5 ~tm. To improve the quality of pre-treatment, the "line coagulation" method is widely used. It represents a version of reagent filtration, and includes the introduction of the organic cation polyelectrolyte of Superflox573 type immediately before high-rate filters, loaded with a material possessing high porosity and specific granulometric composition. Aluminosilicate filtering material (so-called A filter) with def= 0.4-0.6 mm and layer height h = 1 m is usually employed [59,60]. The amount of flocculant added is usually 1-5 mg/dm 3. Application of "line coagulation" (or, more properly, "line flocculation") method makes it possible, using a single apparatus, to reduce FI value from 50 (for the initial water) to 2 [59]. Artesian waters often contain a significant amount of bivalent iron which, subject to the action of oxygen contained in the air, is converted to Fe (III), hydrolysed and forms colloidal suspensions of iron basic salts, up to ferric hydroxide. In the USA, for example, more than 25% of municipal wells provide water which possesses increased contents of iron and manganese compounds [61]. Usually the concentration of iron in artesian waters amounts to 2-5 mg/dm 3, but groundwater springs exist whose iron contents amounts to 20 mg/dm 3 [62]. In such cases water must pass through grained filter with aeration (to oxidise Fe 2§ -> Fe ~§ immediately before the filtration. In the studies concerned with the removal of iron from water by filtration (see, for example [46,63]), primary attention was paid to questions of essential importance, namely to the oxidation of bivalent iron by air oxygen in the process

677 of water aeration, to the maintaining of optimum pH value, and to the rate of filtration in order to achieve the most complete oxidation. However, practically all known results are related to filters with quartz sand. At the same time, the choice of the type of filtering material essentially determines both the rate of ferric hydroxide film formation on the surface of grains (this film playing an important role in autocatalytic oxidation reaction Fe2§ Fe3+), and the rate of the adhesion of this hydroxide at filtering material grains. Clinoptilolite is a promising filtering material for the removal of iron compounds from artesian waters. This mineral possesses more developed surface, intergranular porosity and, what is very important, the capability for cation exchange. Iron ions which displace Ca 2§ and Na § cations from clinoptilolite exchange centres, act as nuclei for the formation of ferric hydroxide film securely held at the surface of clinoptilolite grains. Therefore, simple replacement of quartz filter by two-layered clinoptilolite one with grain size of 20-40 and 0.8-2.1 mm leads to significant increase in the duration of the filtering cycle and an increase in the quality of removal of iron from underground water [45]. This technology was applied for underground water purification at Beregove region (Transcarpathian Ukraine). This artesian water contained 4.6-7.0mg/dm 3 of Fe(II) and 0.15-0.20 mg/dm 3 of Fe(III). Almost all iron was removed from water by clinoptilolite filter with preliminary aeration: iron content was decreased to 0.3-0.05 mg/dm 3 [64]. Our data [65] also indicate that clinoptilolite is significantly better than quartz sand for the removal of iron compounds from river water. The electric charge of filter grains in water is a very important factor for the removal of the particles of silica compounds: as latter are charged negatively in water, filter grains must have positive charge in water. Such charge is characteristic to the particles of low-silica bauxite and sorbent based on calcined dolomite, which are the materials generally used for water purification in heat power engineering [29]. 3.3. U s e o f n a t u r a l m a t e r i a l s as a d a g u l a n t s in w a t e r p u r i f i c a t i o n Layer silicates montmorillonite, kaolinite etc., although being among dispersed impurities which should be removed from drinking water, rather paradoxically can themselves be used to purify drinking and waste water from suspended particles. For example, it was indicated in [66] that the addition of small (0.5-0.7 g/dm 3) amounts of ground montmorillonite, palygorskite or vermiculite leads to virtually complete removal of coli bacteria from water. Combined use of these minerals with flocculants or small amounts of coagulants is recommended for more intensive and complete precipitation of dispersed silicate particles with bacteria and viruses adhered to them. The use of aluminium sulphate and other coagulants, in particular, during cold season, is inefficient [67] for the purification of highly coloured natural waters with low turbidity, i.e., waters containing many coloured humic materials and small amounts of disperse mineral impurities. In these cases so-called "turbidityforming agent", that is, finely dispersed montmorillonite, preferably in sodium

678 form, is supplementary added to water at water t r e a t m e n t plants to improve the flocculation process. This reduces flocculation time by 30-80%, and decreases the consumption of aluminium sulphate [67]. In Hungary and Austria the method of clarification of drinking and waste water from dispersed impurities was developed, which includes consecutive t r e a t m e n t of water with montmorillonite or other layer silicate in sodium form (0.1-0.3 g/dm 3) and high-molecular flocculants (0.3-0.8 g/dm 3) [68]. It is instructive to examine briefly the colloid chemical aspect of water purification from dispersed impurities upon adding small amounts of montmorillonite [1,4]. It is known that montmorillonite in sodium form is dispersed spontaneously in water down to elementary plates about 1 nm thick [69]. Particles of montmorillonite and other layer silicates in water carry electrical charges of two signs: basal faces of particles are negatively charged, while the side ones and edges of particles at pH < 6 are positively charged [69-71]. Negative charge of basal faces is due to the non-stoichiometric isomorphous substitutions in the structure of montmorillonite, mainly those of Al 3+ by Mg 2+. The excessive negative charge is compensated by exchange cations located at basal faces. In water, the exchange cations pass into diffuse part of electrical double layer, and oxygen atoms at basal faces of clay particles carry negative charge. Side faces and edges of montmorillonite particles are charged predominately positively, as the result of basic dissociation of =AI-OH and -Mg-OH groups at pH _< 8.5. The number of negatively charged surface sites of clay particles exceeds the number of positively charged sites, therefore clay particles in water at pH _> 3 carry a net negative charge [69-72]. The fact that charges of basal planes and side faces of layer silicate particles are opposite to each other leads to some interesting features in the behaviour of their aqueous dispersions, in particular, plane-to-edge type of particles association. This interesting effect is called internal mutual flocculation [73]. The presence of two types of charges at the surface of layer silicate particles results in the fact that in an aqueous medium clay particles can interact with dispersed impurities which carry either positive or negative charges on their surfaces, i.e., they play the role of universal adagulants-precipitants of dispersed impurities from natural and waste waters. In fact, positively (negatively) charged impurities interact with negatively (positively) charged surface sites of layer silicate particles, stick together into large aggregates and precipitate. This effect is described by Deryagin's theory of heterocoagulation (for particles of both types of colloidal degree dispersion)- heteroadagulation (for different size of particles of the precipitated substance and the precipitator) [74]. To accelerate the precipitation of aggregates formed by dispersed particles in water, high-molecular flocculants can be introduced into the system: the adsorption of flocculant macromolecules on particles of several neighbouring aggregates results in their bonding together, and thus intensify the precipitation.

679 Efficient technologies for natural and waste waters treatment, like Retamix [75] and Flygtol [76] have been developed, essentially based on the application of bentonite adagulants and macromolecular flocculants such as polyacrylamide (PAA) and polymethylvinylpyridium chloride (PPC). In Ukraine bentonites were used for wastewater treatment of paper-making factories with respect to finely dispersed cellulose fibres [77]. The interaction of negatively charged cellulose particles with positively charged side faces of bentonite particles results in the increase of particle size; presence of flocculant macromolecules leads to the settling of grown particles which form a dense sediment. Its capacity (~ 7%) is lower than that characteristic to the traditional method of coagulation with aluminium sulphate. The content of dispersed substances in water decreases to 3-5 mg/dm ~. Permanganate oxidability also decreases in comparison with initial water due to sorption of water-soluble organic acids on bentonite. Table 10 shows the results of analysis for waste water and water purified by Cherkasy m o n t m o r i l l o n i t e - natural and soda (Na2CO 3) activated (consumption of adagulant 100mg/dm3), and by flocculants: PAA (2.5mg/dm 3) or PPC (1.25 mg/dm3).

Table 10 Application of Cherkasy montmorillonite for treatment of paper-making plant waste water Adagulant and Flocculant Initial waste water Natural montmorillonite, PAA Natural montmorillonite, PPC Activated montmorillonite, PAA Activated montmorillonite, PPC

Suspended substances, mg/dm a 90 7 4 5 3

Permanganate oxidability, mg O2/dm a

Sediment amount, %

10.0 8.8 8.3 7.5 7.1

12.0 7.0 7.5 6.5 7.0

A version of this technology was developed for the particular application in process water treatment of paper-making plant producing coloured paper [78]. At first stage of the process natural bentonite and flocculant were used to purify water from cellulose fibre, sizing substances and other dispersed impurities, while at second stage valuable component, i.e., anion dye of direct scarlet type was adsorbed from water on bentonite modified by aluminium sulphate. Bentonite saturated with dye was then used as the paper filler increasing the depth of its colour, with the total saving of the dyestuff. In view of the development of membrane methods of water treatment and recuperation of its valuable components, the following technique for removing of the dyestuffs from wastewater of paper-making factories was proposed [2]. At the

680 pre-membrane stage natural bentonite and flocculant are used to remove cellulose fibre, sizing substances and other dispersed impurities from water, while at the second stage water is delivered to electromembrane apparatus for dyestuff removal which takes place in the concentration chamber. The simplest method to decrease the amount of water pollution at papermaking factories and recycling of pollutants is partial or complete replacement of kaolin as a paper filler, for white bentonite. The latter has much higher sorption capacity as compared with kaolin, and when introduced into the paper composition, retains finely dispersed cellulose fibre and adsorbs the dissolved organic substances, i.e., products of cellulose oxidative destruction, thus considerably improving the quality of water [2,79]. An efficient, almost wasteless technology of process water premembrane t r e a t m e n t of textile factories and plants of artificial leather was developed [2,80]. This technology comprises two-stage reagentless and reagent foam flotation in pneumatic units of equipment by air dispersion with a porous partition, adagulation-adsorption t r e a t m e n t of water with bentonite clays, either natural or those modified with aluminium sulphate, and two-stage filtration through granulated filter with quartz and clinoptilolite medium. The technology ensures sharp decrease of surface active substances (SAS), dyestuffs quantity, of the values of chemical (COD) and biological (BOD) oxygen demand etc. (see Table 11) and, in general, provides the purified water quality required for a stable operation of reverse osmosis membranes for water demineralisation. In comparison with traditional water t r e a t m e n t methods (coagulation, ozonation, etc.), the technology is characterised by small amount of reagents spent (at the flocculation stage not more than 10 mg/dm 3 of the flocculant is used, while at the adagulation stage the maximum amount of clay is 50 mg/dm 3) and allows for a utilisation of slimes in the main production process. For example, 100% of the slime formed during adagulation-adsorption treatment is used as a filler for artificial leathers. The flotation condensate formed (0.05-0.1% of the water capacity treated) is burned in a special furnace. This technology makes it possible

Table 11 Amount of pollutants (mg/dm 3) in initial and treated wastewater from artificial leather manufacturing enterprise Pollutants Suspensions Direct dyestuffs Dispersed dyestuffs Anionic SAS Nonionic SAS COD BOD20

Initial water 50-200 50-100 50-150 50-150 50-100 200-300 120-170

Treated water 0-2 0-4 0-2 0.5-1 0.5-1.5 50-70 5-15

681 to increase the quality of t r e a t m e n t by 30-50% in comparison with traditional technologies. It also decreases power demand at least by 30%, sharply decreases (by 5-7 times) the amount of slimes obtained, and requires twice as small production areas. This technology has successfully undergone pilot-production tests at one of the enterprises manufacturing artificial leather. A version of this technology employing natural porous coal was developed. It was mentioned above that to improve the coagulation of dispersed impurities in high coloured and low turbidity natural water, the turbidity forming agent is introduced. To increase the effect of the treatment, to decrease the content of residual aluminium, as well as to reduce the amount of the sediments formed, it was suggested in [81] to use powdered clinoptilolite with the content of rock-forming mineral not less than 70% (group A) and particle sizes < 0.25 mm instead of bentonite. This is, in fact, a waste material (to 40%), formed in the production of crushed clinoptilolite of 2-3 and 3-5 mm fractions, used as a granulated medium of high-rate filters for river water treatment (see above, Section 3.2). Clinoptilolite is most efficiently used when combined with the coagulant solution, with mineral-coagulant ratio 1:1 (calculated on A1203 basis). This results in saving 25% of the coagulant, increasing the t r e a t m e n t quality (Table 12), and decreasing the amount of the sediment formed by 1.3-1.5 times as compared to the traditional technology of coagulation which employs no additives. For initial Dnipro River water with parameters listed in Table 12, the optimal coagulant portion determined by test coagulation is 20mg/dm 3 (calculated on A1203 basis). If clinoptilolite is used in amounts of 7.5-15 mg/dm 3, this portion decreases to 15 mg/dm 3, which is very important in terms of coagulant saving.

Table 12 Quality parameters of Dnipro River water treated with the aid of adagulants Adagulant Initial water Bentonite Clinoptilolite Ditto

Portion, mg/dm 3

Turbidity, mg/dm 3

Colour, degree

Residual aluminium, mg/dm 3

10 7.5 15

5.1 1.2 0.9 0.4

49 17 14 8

0.23 0.19 0.12

Natural porous coal possesses structural mesopores with r = 1.7 and 3.3 nm, and is characterised by high enough adsorption properties with respect to molecules of dyestuffs and SAS with molecular mass M = 370-620 [28]. Fulvic acid also has M = 250-1000 [1]; it should thus be expected that natural porous coal will adsorb fulvic acid from natural waters. In fact, introducing 50 mg/dm ~ of

682 powdered natural porous coal into Dnipro river water simultaneously with usual quantity of aluminium sulphate results in the additional 25% decrease of c o l o u r from 20 down to 15 degrees.

0

U S E OF N A T U R A L S O R B E N T S F O R THE P U R I F I C A T I O N OF W A T E R F R O M ORGANIC M O L E C U L E S A N D IONS

Physicochemical principles for rational selection of natural sorbents used in the adsorption purification of waste waters from organic substances were presented in our earlier publications [1,4,82]. These studies had shown that natural sorbents can be efficiently used to remove surface active substances (SAS), dyestuffs and some high-molecular substances from waste waters. These are the problems considered in following section. 4.1. N a t u r a l s o r b e n t s in p r o c e s s e s o f w a t e r p u r i f i c a t i o n f r o m s u r f a c e active substances Low-molecular SAS which possess detergent properties can be divided into three large groups: non-ionic, cationic and anionic ones. Anionic SAS, which are extensively employed in textile industry, are removed easily from water by flotation methods involving cationic flocculants, and coagulation with traditional aluminium sulphate used as the coagulant. Adsorption methods are also employed for water purification with respect to anionic SAS [83,84]; natural adsorbents most suitable for this purpose are bauxites and magnesium oxide, because their surfaces acquire positive charge in water. One of principal physicochemical properties of layer silicates (clay minerals) is their cation-exchange ability. This property of clay minerals can be applied to remove cation-active SAS widely used in flotation processes of ferrous metal ores etc., from industrial sewage [85]. The main criteria for the selection of natural ion exchangers as the sorbents of cationic SAS are their cation-exchange capacity and the accessibility of exchange centres for large enough organic cations. According to these criteria, the best natural exchangers are montmorillonite and vermiculite. Figure 2 shows adsorption isotherms of dodecyl- and cetylammonium chloride on Ca-form of Pyzhevsk montmorillonite, and isotherms of Ca 2§ cations desorption from these layer silicates [86]. The analysis of these data leads to the conclusion that cation-active substances are sorbed by montmorillonite according mainly to ion exchange mechanism. However, the comparison of the amounts of dodecyl- and cetylammonium adsorbed by montmorillonite with the amount of displaced Ca 2§ cations shows that a portion of cation-active substances is sorbed molecularly on the mineral. Such hyperequivalent sorption is characteristic for cationic SAS, whose structure possess hydrocarbon chains not less than 10 carbon atoms long [87]. X-ray [88] and IR-spectroscopy [89] data indicate that the association of sorbed organic cations and salt or amine molecules takes place.

683 a, mg-equiv/g

1.0

0.5

I 1

I 2

I 3

(a)

1.0

0.5

A ,F

0.1

I

I

0.2

0.3

Ce, mmol/dm3

(b) Figure 2. Isotherms of (1) sorption of (a) dodecyl- and (b) cetylammonium chloride and (2) desorption of Ca 2§ cations for montmorillonite at 20~ Ce - equilibrium concentration of SAS in solution.

684 Cations of SAS are sorbed not only on the external surface, but also on the internal surface of the mineral. The value of first basal reflection on the diffractograms of organomontmorillonites is d0ol=1.67-1.70nm. Therefore organic cations are packed flatly into two layers within montmorillonite interlayer space with the thickness equal to Ad = (d001 - 0.94) = 0.73-0.76 nm. The spacing between the silicate layers depends on the a m o u n t of organic cations penetrated into the interlayer gaps, or more specifically, on the relation between the values of S O, the area of total silicate surface per one exchange centre (0.57 nm 2 per unit charge for Pyzhevsk montmorillonite, 0.365 nm 2 per unit charge for Kovdor vermiculite) and S c, the maximum area which can be occupied by organic exchange cation (about 0.1 nm 2 per one -CH 2- group of an alkyl chain). If S c < S 0, then organic cations fit closely to the silicate layers which results in the formation of the monolayer of organic cations between the silicate layers. When S O< S c < 2S0, flat organic bilayer is formed in the silicate interlayer region. When the value of S c exceeds 2S 0, then organic cations form a layer with individual cations oriented uniformly at some angle with respect to the silicate layer surface [90]. Internal geometric surface area of montmorillonite (~ 750 m2/g [7,69]) m a n y times exceeds its external surface area S = 30-60 m2/g (see Table 4). It is just the accessibility of internal surface of this mineral which leads to its high adsorption properties with respect to cationic SAS. From solutions with the concentration close to critical concentration of micelle formation (CCM), these substances are adsorbed at the external surface of layer silicates in the micellar form. Obviously this can explain the fact that total sorption of long-chain organic cations of the cetylpyridinium type at montmorillonite from solutions with an equilibrium concentration of the order of 3 mmol/dm 3 amounts to 3-4 mg-equiv per 1 g of sorbent [87,91]. This value 3-4 times exceeds the exchange capacity of this mineral (Table 4). The technology of waste water purification for ore beneficiating factory [92] was based on adsorption studies. The pesticides diquat (1,1'-ethylene-2,2'-dipyridilium dibromide) and p a r a q u a t (1,1'-dimethyl-2,2'-dipyridilium dichloride) are also sorbed by layer silicates according to a mechanism of ion exchange [93]. Therefore both montmorillonite and vermiculite can be efficiently used to remove these pesticides from water [1]. Many polar molecules are sorbed by layer silicates due to co-ordination mechanism, which involves the formation of surface complex compounds with exchange cations of these minerals [7]. However, often it is energetically more favourable for organic molecules not to displace water molecules from the first coordination sphere of layer silicate exchange cations, but to form stable hydrogen bonds with them according to the scheme: Me n+... OH2-organic molecule. This is just the mechanism of polyvinyl alcohol [94-96], polyethylene glycol [97,98] and non-ionic SASs sorption at montmorillonite surface [99,100]. These non-ionic surfactants are the polyethylene glycol ethers of fatty alcohols, acids, alkylphenols, etc. with composition CnEm,where n is the n u m b e r of carbon atoms

685 in hydrocarbon chain, m the n u m b e r of oxyethylene groups in polyethylene glycol chain. The sorption of non-ionic SAS at clay minerals occurs mainly because their structure comprises significant n u m b e r of oxyethylene groups (m = 5 to 50, usually m = 6-16). These groups can participate in specific interaction with exchange cations or in a hydrogen bond formation with adsorbed w a t e r molecules via the u n s h a r e d pair of electrons of oxygen atom. IR-spectra show t h a t oxygen atoms of oxyethylene groups interact with exchange cations through bridge water molecules according to the scheme [98]:

Me"+... O - H - - - O ( I H

CH 2 CU2

(4)

The energy of oxyethylene groups interaction with bridge w a t e r molecules is sufficient to overcome electrostatic attraction forces existing between aluminosilicate layers and h y d r a t e d cations which compensate excess negative charge. As a result, the molecules of non-ionic SAS p e n e t r a t e into the interlayer space of montmorillonite. The thickness of interlayer space Ad depends on the concentration of non-ionic SAS in the solution, the n a t u r e of exchange cation (Na § or Ca2+), the length of oxyethylene and hydrocarbon chains in molecule CnE m etc. When equilibrium concentration of SAS is low (c = 105-10 .4 g/dm3), then a single (d001 = 1.39 nm, Ad = 0.45 nm) and subsequently double (do01 = 1.75 nm, Ad = 0.81 nm) layer of SAS molecules in a flat position is formed in the interlayer space of montmorillonite [99]. When Na-montmorillonite contacts with more concentrated C12E14 solution, then in addition to the reflection d001 = 1.8 nm, another reflection which corresponds to do01 = 4.2 nm becomes clearly seen [1], evidencing t h a t within some portion of Na-montmorillonite interlayer spaces C12E14 molecules are oriented at some angle with respect to silicate layers. Figure 3 shows adsorption isotherms of oxyethylated alkylphenol CsH17-C6H40-(CH2-CH2-O)TH (C8PhE 7, agent OP-7, CCM = 0.1 g/dm 3) at Caforms of montmorillonite and hydromica [100]. It is seen t h a t the adsorption values of non-ionic SAS at montmorillonite samples exceed significantly those obtained for hydromica which possesses specific surface area S = 125 m2/g. This can be a t t r i b u t e d to the fact t h a t internal surface of montmorillonite participates in the absorption of OP-7. Limiting adsorption value for CsPhE 7 at Pyzhevsk montmorillonite is a~ = 360 mg/g. For the adsorption of wetting agent "Nonisol-250" (an ester of polyethylene glycol and oleic acid), the non-ionic SAS which possesses higher molecular mass, at Ca-montmorillonite surface, a ~ = 5 5 0 m g / g [101]. It was reported in [102] t h a t limiting value for the adsorption of non-ionic SAS CgE 5 - C9E50 at Na- montmorillonite for equilibrium concentrations c e > CCM can a m o u n t to a| = 1.0 g per I g of the adsorbent. Our

686 data presented in [100] show that 1 g of Pyzhevsk montmorillonite in Na-form can extract 0.8 g of OP-7 agent from an aqueous solution. The adsorption capacity of the best grades of activated carbon for non-ionic SAS does not exceed 0.1 g/g [28,100], with their cost being much higher than that of clay minerals.

a, g/g

1 9

0.3

2 0.2

0.1

I

I

I

0.2

0.4

0.6

I

Ce, %

Figure 3. Adsorption isotherms of oxyethylated isooctylphenyol (OP-7) ta Ca-form of." (1) Pyzhevsk montmorillonite, (2) Cherkasy montmorillonite and (3) Cherkasy hydromica.

687 The capability of montmorillonite (bentonite) clays for efficient adsorption of non-ionic surface-active substances (NSAS) from water was utilised in the development of industrial technology for the treatment of stratal water at gas production fields [100]. When the NSAS solution is prepumped into gas wells to reduce the collector properties of the stratum, some portion of the NSAS is carried away to the surface along with stratal water. This water sometimes contains up to I 000 g/m 3 of NSAS, i.e. of OP-7, prevocell EO, prevocell WOF-100, ditalan OTS-45, syntanol DS-10, etc. To remove these from water, commercial bentonite powder can be used. Table 13 generalises the industrial test results of the stratal water treatment of Pynyanu gas deposit (Ukraine, L'viv province) to remove prevocell EO using Cherkasy bentonite powder [100]. It is seen that minimum permissible concentration (MPC) levels of NSAS (0.5 g/m 3) can be achieved by one-stage stratal water treatment with Cherkasy bentonite. In this case the sorbent demand is 17-20 kg per 1 kg of the NSAS removed.

Table 13 NSAS concentration (C) in initial stratal water and water after treating with the aid of bentonite Sorbent consumption, kg/m 3 1.4 4.8 17.0

C, g / m 3 Initial water 48 270 950

Treated water 0.5 0.4 0.5

The order of technological operations is as follows: stratal water which contains NSAS is delivered from the collection tank to the mixer into which bentonite powder was fed. The mixture is stirred until clay is completely dissolved. Then the pulp is moved into the settling tank for natural phase separation. A rather high concentration of salts (8-21 g/dm 3) in stratal water contributes to an efficient precipitation of the solid phase. The clarified water is delivered into the capacity tank of treated water, while the slime (spent sorbent) accumulated on the slime site platform, is disposed for the utilisation as a drilling fluid component.

4.2. Bentonite clays in processes of water purification from cation dyestuffs Cation (basic) dyes: crystal violet, malachite green, fuchsine, rhodamine, safranine, methylene blue etc. exist in water as organic cations. These dyes are widely used in paper industry for paper colouring, due to high affinity of organic cations with respect to negatively charged sites of cellulose macromolecules [1].

688 Publications [103-105] summarise the studies of adsorption properties of various clay minerals with respect to methylene blue, crystal violet and malachite green. It was shown that montmorillonite, the rock-forming mineral of bentonite clays, possess the highest adsorption capacity with respect to these dyes. The adsorption of cationic dyes takes place both at the external surface of montmorillonite, and in its interlayer gaps. After the adsorption of crystal violet, first basal reflection of Na- and Ca- forms of Pyzhevsk montmorillonite doo I = 2.06 nm and 1.96 nm, respectively. Table 14 comprises the data concerning the adsorption of malachite green dye at Oglanly montmorillonite (Turkmenistan), which possesses cation exchange capacity 0. 84 mg-equiv/g. It is seen that for equilibrium concentrations ( 3 + 3 0 ) . 1 0 .4 mol/dm 3, the adsorption of this dye takes place according predominantly to cation exchange mechanism, with high selectivity of the ion exchanger with respect to organic cations. Spectral analysis data [104,105] suggest that the observed differences in the adsorption of dyes and desorption of inorganic cations (a > a') are related to the fact that the dyes are adsorbed not only in the form of individual organic cations, but also as their associates with molecules. Table 14 The amounts of malachite green dye (a) adsorbed by Na form of Oglanly montmorillonite and displaced Na § cations (a') for various equilibrium concentrations (C~) of the dye in the solution C e. 104, mol/dm 3

a, mol/kg

a', g-equiv/kg

3 28 33 39 65

0.52 0.64 0.75 0.88 1.31

0.40 0.53 0.58 0.62 0.62

The contents of dyes in waste waters of paper-making plants does not exceed 10 mg/dm 3 [1]. It was shown in [106] that the introduction of 100 mg/dm 3 of bentonite and 0.6 mg/dm 3 of the flocculant is sufficient for the complete decolouration of such water. The data obtained in the studies [103-106] were used in the development of technology which applies bentonite clays for the purification of waste water discharged by a plant which produces cationic dyes used as analytic indicators [107].

689

4.3. P h y s i c o c h e m i c a l principles of the application of clay m i n e r a l s for the r e m o v a l of high m o l e c u l a r organic c o m p o u n d s from waste waters In this section the advantages of the application of clay minerals for the purification of industrial waste waters from polyvinyl alcohol and protein compounds are discussed. Polyvinyl alcohol (PVA) is a non-ionic water-soluble polymer, which is widely used in various industrial processes; in textile industry, for example, PVA is applied for textile dressing. Macromolecules of this polymer are almost incapable of biologic destruction; therefore, to remove it from industrial waste water, physicochemical methods are to be used, among which the adsorption methods are most promising. It was shown by X-ray studies [94] that the adsorption of PVA from water solutions leads to the increase in first basal reflection of montmorillonite Na- and Ca-form to the values d001 = 3.00 nm and 1.83 nm, respectively. Thus one can expect high values of PVA adsorption on montmorillonite. This conclusion was supported in our adsorption studies [96], see Figure 4. The conditions of adsorption experiment were as follows: polyvinyl alcohol with molecular mass a, g/g

0.5

0.4

0.3

0.2 ( ~ qP

()

0.1 ~-(~ ()

'

I

I

I

0.2

0.4

0.6

Co, %

Figure 4. Adsorption isotherms of PVA on Pyzhevsk montmorillonite: (1) Na-form, (2) Ca-form.

690 M = 63000 and the n u m b e r of acetate groups less t h a n 2.5% was used as an adsorbate, concentration of adsorbent in the solution was 2.5% mass, adsorbent/solution contact time was 24 hours. Dispersion pre-processing of montmorillonite significantly increases its adsorption capacity with respect to PVA. The results presented above show that bentonite clays can be used to remove PVA and accompanying organic substances from industrial waste waters [1]. Sodium bentonites were long been used for the cleaning of vines and beer [108]. In this process the particles of montmorillonite act both as a d a g u l a n t and adsorbent. In fact, for pH values below isoelectric point pH i = 4.5-5, protein macromolecules and their associates are positively charged in water medium, and efficiently interact with negatively charged sites of montmorillonite particles. The neutralisation of charges leads to the aggregation of particles and their sedimentation. Our data [109] concerning the adsorption of bovine serum albumin with molecular mass M ~ 65000 on montmorillonite Na-form (Prodaneshti, Moldova) which possesses cation exchange capacity E = 0 . 9 5 mg-equiv/g show (see Figure 5) that this adsorbent is characterised by high adsorption capacity with

a, g/g o

o

!

o

0.8

0.4

o

3

4

8

C<, g/dm~

Figure 5. Adsorption isotherms of albumin on Prodaneshti montmorillonite: (1) Na-form, (2) Ca-form and (3) Cherkasy hydromica.

691 respect to dissolved albumin molecules. The conditions of adsorption experiment were: pH 3.6, temperature 10~ albumin concentration in water solution c = 0.1-1.8%, solution to adsorbent ratio 100 c m 3 : 1 g, adsorbent/solution contact time 48 hours. It follows from X-ray data that the adsorption of albumin macromolecules takes place in montmorillonite interlayer regions; this results in the increase of its first basal reflection to d0o1 = 2.3 nm. In addition to serum albumin, the adsorption of casein, chymotrypsin, lysozyme and ovomucoid also takes place in Na-montmorillonite interlayer gaps [110]. The thickness of protein layer in interlayer gaps depends on the type of adsorbed protein, its molecular mass, concentration and pH value of the solution, and varies within the range Ad = 0.6-6 nm [110]. These results show that clay minerals can be used as adagulants and adsorbents for the removal of protein substances from waste waters disposed by dairy and meat industrial plants [1,111], where the contents of proteins can amount to 1-7 g/dm ~. Up to recent time, the main attention at the industrial enterprises of this type was paid to the extraction of fat from waste waters, and its subsequent utilisation. Waste waters polluted by proteins in most cases were treated by biological cleaning process; this have led to unrecoverable loss of proteins which otherwise could be used as a valuable product for feeding stuff. We have developed a comprehensive flotation/adsorption technology for the cleaning of waste water at meat industrial plant. At the first stage the water was acidified to pH 3.5, and the fats were extracted using the flotation. Then bentonite or hydromica clay suspension was introduced into water, with total consumption of clay being not more than 0.5-5.0 g per 1 dm 3 of cleaned water. To intensify the sedimentation of disperse impurities, up to 5 mg/dm 3 of flocculant was also introduced into the system. The organomineral sediment separated from clarified water was dehydrated and sterilised by pressurised heating to 120-130~ This solid sediment, containing 40-50% mass of protein, can be utilised as an additive to the feeding stuff. This technology was tested in industrial conditions.

Q

U S E OF N A T U R A L S O R B E N T S FOR WATER P U R I F I C A T I O N FROM I N O R G A N I C CATIONS

An important feature of layer silicates and natural zeolites, as compared with ion-exchange resins, is their increased selectivity with respect to large cations [18,112]. In this section the reasons for such high selectivity of these inorganic exchangers with respect to Cs § Rb § K § and NH4 § ions will be discussed, and the application of clay minerals and zeolites for specific processes of waste water purification from large cations will be considered.

692

5.1. S e l e c t i v i t y of layer s i l i c a t e s and zeolites with r e s p e c t to large a l k a l i - m e t a l ions Our and literature d a t a on t h e r m o d y n a m i c constants K a of Na § exchange for other alkali metal and a m m o n i u m cations in organic cationite, layer silicate and zeolites [1,18,112-116] are s u m m a r i s e d in Table 15.

Table 15 T h e r m o d y n a m i c constants K a for exchange of Na § ions of various exchangers for other alkali-metal ions and a m m o n i u m ions Ka Exchanger Sulfonated Polystyrene (Dowex-50) Kaolinite Glauconite Montmorillonite Palygorskite Zeolite X Chabazite Erionite Clinoptilolite Mordenite

Li + 0.5 0.5 0.6 0.7 0.20 0.06 0.10 0.20 0.07

K+ 1.5 7.8 2.3 3.2 3.2 15.0 8.9 28.2 7.2

NH~ 1.3 2.2 10.0 4.0 8.5 4.6

Rb + 1.6 12.3 9.1 9.7 22.4 42.7 30.2

Cs + 1.7 18.2 17.6 31.6 24.4 31.6 68.4 63.1 69.1 51.2

The analysis of the exchange constants for alkali metal ions in a synthetic cationite h a d shown t h a t the selectivity of this type of ion exchange m a t e r i a l increases along the series Li < Na < NH 4 < K < Rb < Cs, i.e., corresponding to the increase of the ion crystallographic radius. The observed sequence of relative alkali metal ion affinities for the ionite has been a t t r i b u t e d [117] to the differences in the energy of electrostatic interaction between the counterions and the ionite matrix, which is inversely proportional to the square of the h y d r a t e d ion radius. The radii of h y d r a t e d alkali metal ions follow a series t h a t is the reverse of t h a t for the radii of the u n h y d r a t e d ions. Comparison of the data in Table 15 shows, however, t h a t the selectivity of certain mineral ion exchange m a t e r i a l s with respect to K § NH~ and Rb § ions is 10-25 times higher t h a n t h a t of organic cationites, and their selectivity with respect to Cs § ion is 50 times higher. Such high affinity of large ions for the surface of zeolites and layer silicates can no longer be ascribed solely to physicochemical factors, in particular, to differences in the radii and polarizabilities of the h y d r a t e d ions. The explanation of these p h e n o m e n a m u s t be sought for in the individual characteristics of mineral ion-exchange materials, primarily their s t r u c t u r a l characteristics.

693 The data presented in Table 15 show that the constants for the exchange of NH~, K § Rb § and Cs § ions for Na § ions of montmorillonite, kaolinite, glauconite and palygorskite are similar to those characteristic for the exchange of these ion pairs in synthetic zeolite X. This is quite natural, since exchanged alkali metal ions which possess large size, are localised in the pseudohexagonal oxygen voids in this zeolite and layer silicates [115]. These ions continue to remain partially embedded in the pseudohexagonal structural voids even in completely waters a t u r a t e d specimens. The interaction of Rb § and especially Cs § ions with ion exchanger matrix leads to the significant changes in its structure, which are ultimately reflected in the physicochemical characteristics of montmorillonite, in particular, in the increase + + of its external surface to 150-190 m2/g [7,16]. Cations NH 4 , K , Rb § and Cs § in general are irreversibly immobilised within the interlayer region of vermiculite (0.6-1.0 charge units per unit cell), similarly to the m a n n e r in which K § ions are immobilised between the silicate layers of mica (layer charge is 1.0-2.0 units). High selectivity of chabazite, erionite, clinoptilolite and mordenite with respect to cations K § Rb § and Cs § is caused by the localisation of these ions in eightmember oxygen rings of zeolites structure [115,116]. Free section dimensions of eight-member oxygen rings is comparable with the sizes of large alkali metal cations. The efficient interaction between exchange cations and structural oxygen atoms of eight-member rings increases the selectivity of these zeolites with respect to K § Rb § and Cs § cations (Table 15). Thermodynamic exchange constant K a provides the information about the mean selectivity of all types of exchange centres which exist in the structure of an ion exchanger, with respect to the sorbed cation. The most selective ion exchanger centres may be determined from the analysis of the correction of the selectivity coefficients K S' at low coverage levels, particularly at N~=0.2. Corresponding values for Cs+-Na§ of zeolites with high silicon contents are Ks'= 100-150 [116]. These data demonstrate the high selectivity of chabazite, erionite, clinoptilolite and mordenite with respect to Cs § cations. It is interesting + to note that the position of NH 4 cation in the selectivity order of n a t u r a l zeolites with respect to alkali metal cations (Cs > Rb > K > NH 4 > Na > Li) does not agree with its ionic radius (r = 0.143 nm). However, this case is a clear example that the exception proves the rule. In contrast to spherically symmetrical alkali metal ions, NH4 § ion exhibits pronounced tetrahedral structure. Being localised in 8member ring, it is involved in an efficient ion-matrix interaction not with six oxygen atoms, as in the case of Cs § Rb § or K § cations, but with two (for almost flat rings) or three (for puckered rings) oxygen atoms only. This also leads to the decrease in the energy of a m m o n i u m ions interaction with the framework, which results in the decrease in the values of K s' and K a [116]. Calorimetric studies of ion-exchange equilibria on clinoptilolite involving unicharged cations [118] confirm this conclusion. The energies of interaction of the compensating ions with ion exchanger matrix, and therefore the exchange constants, depend both on the

694 commensurability between the exchange cations and silicon-oxygen structural rings, and on the charge of the ion exchanger matrix, i.e. on the level of the AI~+-~ Si 4+ isomorphism. An increase in the charge of the matrix (a decrease in the SiO2 :A1203 ratio) leads to an increase in the contribution of electrostatic repulsion between neighbouring cations WAA into the total energy of their adsorption by ion exchanger. The contribution of WAA is especially significant, when large size cations Cs § Rb § K § are located in exchange positions, this results in the decrease in the Cs (Rb, K)-NaZ exchange constants. This effect is illustrated by Figure 6 reproduced from our study [119].

Ka

\ \

1

b

\ \ Q

80

\ \

\3 \ \ \ \ \ \

40

\ \ \

4

b

\ \ \ \

0.2

I 0.3

I 0.4

I 0.5

(A1)~

Figure 6. Ka values for Cs - Na-zeolite exchange vs equivalent proportion of aluminium in zeolite structure (A1)c: (1)erionite, (2)offretite, (3) low charge chabazite, (4)high charge chabazite.

695 The similarity between the sizes of entering large cations and free section dimensions of oxygen-silicon rings in which they are localised, also produces significant effect on both K a and the charge of the matrix. Thus the dependency of K a o n the ratio SiO 2 : A1203 and (A1)c = NA1/(Nsi + NA1) where NA1 and Nsi are the numbers of aluminium and silicon atoms in an elementary cell, respectively, should be analysed only for those zeolites for which the localisation sites are characterised by similar free section sizes. Chabazite, offretite and erionite are interesting in this regard. Their structure contains only 8-member rings (chabazite, erionite) or 8-member rings in combination with 12-member rings (offretite), which are too large for cations to be localised within. The minimum sizes of 8-member rings free section for all these three zeolites are the same. This fact enables one to postulate that for all these zeolites the structural factor plays a similar role in the determining of K a value. Figure 6, which includes also the data on high-charge synthetic chabazite [120], shows a linear dependency of Cs-NaZ exchange constant on the charge of zeolite matrix. Thus for natural zeolites a clear relation exists between the level of isomorphism in the structure and the exchange constant (free energy of exchange). It is necessary to take into account the composition of natural exchange complex of natural zeolites. The crystallisation of zeolites took place under different geological conditions: some of them were crystallised in the presence of calcium cations, while o t h e r s - in the presence of sodium and potassium cations. It was shown [121,122] that sodium sample of low-silicon calcium clinoptilolite in its natural state reveals a higher selectivity to double-charge cations whose dimensions are similar to those of Ca 2§ ion. On the contrary, sodium sample of high-silicon, sodium-potassium clinoptilolite in its natural state reveals a higher selectivity with respect to unicharged cations possessing large size. The thermodynamic exchange constant for the system Cs-Na-clinoptilolite decreases from K s = 69-70 (Table 15), the value characteristic for Hector (USA) and Dzegvi (Georgia) high silicon sodium-potassium clinoptilolite, to K a = 23-25 [118] for Sokirnitsa (Ukraine) low silicon calcium clinoptilolite. These data show that natural zeolites possess the property of memory with respect to cations which took part in their crystallisation, and to cations which possess charge and size similar to those of ions in initial samples. It is important that high selectivity of clinoptilolite with respect to large size alkali metal cations in both cases takes place. Thus, the sorption of large cations on natural high silicon zeolites, in addition to thermodynamic selectivity factor, is characterised by a more significant crystallochemical (geometric) factor, which results from the localisation of these cations in 8-member structural rings with the dimension of free section being comparable to the sorbed cation size. This conclusion may be considered as one of the formulations for the crystallochemical principle of selectivity of natural high silicon zeolites with respect to large cations.

696 In our earlier publication [115] it was stressed that selective sorption of large cations in 8-member zeolite structural rings, and the formation of stable complexes of alkali and alkali-earth metal cations with macrocyclic ligands are the phenomena of r a t h e r general character. Later [116] we have shown that the relation exists between the crystallochemical principle of selectivity of mineral ion exchangers, and the principle of the closest correspondence of the macrocycle opening to the size of complex-forming metal ion [123].

5.2. Use of z e o l i t e s for the p u r i f i c a t i o n of water from i n o r g a n i c c a t i o n s High selectivity of natural zeolites with respect to large size ions is used for the removal of radioactive 137Cs, 9~ and other ions from nuclear wastewaters, and for the removal of ammonium ions from municipal, industrial and agricultural effluents. A critical analysis of corresponding methods of treating these wastewaters was presented in [1,124]. Here the analysis of most interesting publications concerning the application of natural zeolites for the removal of inorganic ions from waste waters will be presented. Ames [125] was the first to demonstrate the ion-exchange specificity of clinoptilolite for the removal of radioactive caesium and strontium from low-level waste streams of nuclear installations. This process was extensively studied at Hanford and several other nuclear-test stations in the United States in the 1960s. Millions of gallons (1 american g a l l o n - 3.78 dm 3) of low-level 137Cs wastes have been processed through zeolite ion exchangers since that time [124]. The "saturated" zeolite columns were removed from the system, buried as solid waste, and replaced with new drums containing fresh clinoptilolite. Similar process was developed to remove 137Cs from high-level effluents using chabazite-rich ore [126]. Natural zeolites are capable of extracting species such as 9~ 13VCs, 6~ 59'63Ni, 45Ca, ~lCr selectively in the presence of high concentrations of competing ions. Natural zeolites are not only considerably less expensive than organic ionexchange resins; they are much more resistant to nuclear degradation. Due to their silicate nature, these zeolites also react rapidly in cement or glass systems, entraining the radioactive species into the final concrete or vitreous products. Therefore the studies concerning the use of natural zeolites for radioactive waste waters purification are in progress. Here the works of Robinson et al. [127,128] are to be especially mentioned, where it was shown that chabazite zeolites (both natural and synthetic) are the most effective materials for the removing of trace quantities of Cs and Sr from radioactive waste water. The amount of purified water is 3 000 000 dm3/year. The use of natural zeolites for the radioactive waste water t r e a t m e n t have been studied in a number of countries, including Canada, Great Britain, France, Bulgaria, Mexico, Japan, Germany, Russia, Ukraine [124]. The removal of ammonium ions from municipal, agri-industrial and industrial waste waters becomes increasingly important at the present time. Not only is + NH 4 toxic to fish and other forms of aquatic life; it also contributes greatly to the rapid growth of algae and leads to the eutrophication of lakes and rivers.

697 Therefore the concentration of ammonium ions in reclaimed water must not exceed 0.2-0.5 mg/dm 3 [18,129]. House waste water normally contains up to 30 mg/dm 3 of ammonia. A very promising method of the removal of ammonia from waste water is the ion exchange using natural zeolite clinoptilolite as the ion exchanger. Experimental-industrial and industrial testing carried out in the USA [1,124,129] have shown that the use of this mineral makes it possible to decrease the ammonium ion content in waste water by 93-97% and to attain the residual ammonia concentration in water of 0.2-0.5 mg/dm 3. However, in the development of a technology which employs natural zeolites for the t r e a t m e n t of municipal waste water and waste water of industrial plants to remove ammonia nitrogen, in addition to the ion exchange unit it is mandatory to have an efficient design of the unit used to regenerate exchange filters. Presently existing methods for the regeneration of spent natural zeolites may be divided into biological, reagentless and reagent methods. The critical analysis of these methods was presented in our review [54]. Considering the high acid resistance of clinoptilolite, we proposed its regeneration with an acid [130]. This enables the restoration of the sorption capacity of mineral ion exchanger with respect to ammonium ions (by its conversion to the H-form) and production of a solution of ammonium fertiliser, for example ammonium nitrate or sulphate, in a single technological operation. At Cherkasy (Ukraine) Industrial Association "Azot", the efficiency of the application of Sokirnitsa clinoptilolite for the t r e a t m e n t of waste water containing 100-260 mg/dm 3 NH4 § with its subsequent regeneration by 2-2.5% H2SO 4 [54], was industrially tested. It was shown that the most efficient process is the "hungry" (60-70%) regeneration of the spent filter. Here 3 kg-equiv of H2SO 4 are consumed per l k g - e q u i v of clinoptilolite exchange capacity. In particular, 60% regeneration of filter containing 10.6 tons of clinoptilolite had required 66 m 3 of 2.2% H2SO 4 (1462 kg). As a result of the regeneration, 4.91 kgequiv of NH4 § was removed from the entire clinoptilolite media. In 11 cycles of sorption and subsequent regeneration, the dynamic exchange capacity of the clinoptilolite (DEC of approximately 0.5 kg-equiv/kg) was practically unchanged. Animal waste waters possess high concentration of ammonium and potassium ions (up to 350 and 650 mg/dm 3, respectively). A comprehensive technology of such water t r e a t m e n t at swine complexes was developed [130]. This technology incorporates both mechanical and biological pre-treatment, the phosphate precipitation with the aid of calcium hydroxide, and the sorption of potassium and ammonium cations on clinoptilolite filters. Spent clinoptilolite is mixed with the phosphates after their neutralisation, and used as a fertiliser for lowproductivity sod-podzol soils of light granulometric composition [54]. At present, n a t u r a l zeolites clinoptilolite and phillipsite are extensively used in the USA, J a p a n and Italy to remove ammonium ions produced by the vital activity of fish during its transport or cultivation in special ponds [17,131]. We will examine this problem with regard to the t r e a t m e n t of water in a closed

698 system for the incubation of trout spawn [132]. According to the specifications, the concentration in water of NH4 § ions, a product of the metabolism of the spawn, should not exceed 0.4 mg/dm 3. The water should also contain K § Mg 2§ and Ca 2§ with the concentrations not less t h a n 50 mg/dm 3. Since the process of spawn incubation takes place at 6-10~ and this t e m p e r a t u r e is unfavourable for biological water processing, the t r e a t m e n t with n a t u r a l clinoptilolite was chosen as the efficient one. According to the specifications for the concentration of Ca 2§ Mg 2§ and K § in water [28], Sokirnitsa (Ukraine) clinoptilolite whose exchange complex comprises these ions was used for the treatment. From the rate of NH4 § production in a system with volume of 80 dm 3 of water containing 6 000 trout spawns (3 mg of NH4 § per hour), MPC(NH4 § = 0.4 mg/dm 3, the size of the zeolite filter was estimated and optimal conditions of its operation were determined with the consideration of the dynamic exchange capacity of clinoptilolite. 5.3. U s e of n a t u r a l s o r b e n t s as d e c o n t a m i n a t i o n a g e n t s for t h e e l i m i n a t i o n o f t h e c o n s e q u e n c e s o f t h e a c c i d e n t at C h o r n o b y l nuclear power plant High selectivity of clinoptilolite with respect to radioactive isotopes 137Cs, 9~ etc., was used by us to develop in 1986 the technology of the drinking water purification from the radioactive contamination [133]. The technology includes the t r e a t m e n t of water by powdered clinoptilolite at pH 9.5-9.8, followed by the processing with aluminium sulphate. After the precipitation of dispersed impurities in the clarifiers, subsequent tertiary t r e a t m e n t of water was performed by means of filtration through the 2.0 m thick layer of clinoptilolite with graininess 1-3 mm. It was shown by special experiments t h a t this purification method ensured the reduction in water radioactivity from 10 .7 to 3.10 .9 Ci/dm 3. The advantage of this scheme was in the fact t h a t it could be implemented at the existing equipment of the Dnipro water supply station. Since the impermeable "wall in soil" was constructed around the Chornobyl nuclear power plant, the problem had arisen of the purification of drainage waters prior to their discharge into the Prypiat' river. To perform this purification, the process was recommended which employed the filters with clinoptilolite tuff medium. The graininess of clinoptilolite was 2-3 mm, filtrating layer height 2 m, filtration linear rate 2 m/h; 15-20 mm crushed stone served as a bed-layer. Initial water radioactivity being 106-10 .7 Ci/dm 3, the purification degree amounted to 70-80%, and the service life of the filters was one year. The degrees of water purification with respect to various radionuclides, ensured by the filtration, were as follows: 95% for 137Cs, 80% for 9~ 50-60% for radionuclides of heavy elements, 15-20% for l~ The filters were designed in such a way t h a t the worked-out sorbent could be discharged; the resulting waste material which possessed the radioactivity level of 10 .5 Ci/kg was buried. Water t r e a t m e n t by powdered clinoptilolite and brown or porous coal, as well as the clearing with the help of coagulants and flocculants were recommended, in order to achieve higher purification degree.

699 Clay minerals were also extensively used for the decontamination of clothes, machinery and building materials. When formulating the problem, we have proceeded from the fact that the methods developed before 1986 were essentially relied on the use of surfactants, and could not ensure the decontamination required by safety norms. When standard method was applied to process the machinery and building materials, a serious problem of the decontamination of waste waters in large quantities had arisen. Therefore an essentially new sorption-adhesion method employing clay materials was proposed for the decontamination of clothes, machinery and building materials. Clay particles in an aqueous phase carry electrical charges of both signs. Due to this property, clays act as universal adagulants and can efficiently bound radioactive dispersed particles. Clay minerals possess particularly pronounced cation exchange properties, and display increased selectivity with respect to caesium, strontium and barium ions. This selectivity can be significantly enhanced by introducing special complexing additives into the system [134,135]. Thus, the application of clays as decontamination agents makes it possible to remove radionuclides in both dispersed and ion-soluble state from the contaminated surface. Clay minerals are high-disperse substances with developed external surface [1,7]. This property, their high wettability and the trend of particles to form a coagulation structure in aqueous media ensure high adhesive-enveloping properties of aqueous dispersions. At the same time, high wettability of clay particles facilitates the washing of the aggregates which they form with radioactive particles and radionuclide ions by a water jet from contaminated surface. In our work [133], montmorillonite and palygorskite from Dashukovka (Cherkasy province) deposit were used in the form of aqueous pastes (12-15%) and suspensions (2-7%). Clothes were treated with the 2% montmorillonite suspension at 10-12~ during 1 min with permanent stirring in a special washing machine, and then rinsed twice with water at 15-20~ Building materials and tractor machinery were treated with the pressure-feeded jet of 5-7% clay suspension or by clay paste (palygorskite and its mixture with montmorillonite). A paste was applied to the contaminated surface, held for 0.5-1 h and washed away by a water jet. Initial and final contamination levels of clothes, machinery and building materials were measured using the detectors DP-5V and RAM-63 at specially furnished sites or in premises where the radioactive background was below 0.03-0.1 mR/h. It is seen from Table 16, which summarises residual contamination levels and decontamination coefficients K d for the proposed t r e a t m e n t and those obtained using standard SAS solutions, that the use of suspensions and pastes is preferable to traditional decontamination methods. The application of this technology had provided a quite simple solution to the problem of waste water decontamination. Radioactive pollution was concentrated within clay slime, which could be easily precipitated from water; to accelerate the precipitation process, polyacrylamide was added (2.5-3.0 mg/dm3). The slime was

700 buried, while clarified water was subjected to tertiary purification clinoptilolite filters, and then discharged for subsequent natural filtration.

using

Table 16 Decontamination efficiency of standard SAS-containing solutions as compared with clay dispersions and pastes Initial

Processed items Uniform and working clothes Lorries and bulldozers

Concrete slabs, bricks, roofing slate, etc.

contamination level, mR/h 1.2 1.7 2.9 2.0-2.5 8.0-10 50-70 100-200 300-400 6.0-8.0 80-100 200-250

Standard SAS solutions Residual level, mR/h Kd 0.23 0.42 0.44 1.5-2.0 4.0-6.0 25-40 15-20 30-40 2.0-4.0 2.5-3.5 3.5-8.0

4.3 4.0 6.6 1.3-1.2 2.0-1.7 2.0-2.5 6.0-10 10 3.0-2.0 32-30 57-31

Clay suspensions and pastes Residual level, mR/h Kd 0.10 0.07 0.09 0.2-0.3 0.5-0.6 5.0-8.0 3.0-5.0 6.0-8.0 0.7-1.5 0.7-1.5 1.5-2.5

12.0 24.3 25.6 10-3.3 16-10 10-8.0 33-40 50 8.6-5.3 114-66 130-100

To summarise, the comprehensive application of clay minerals in the works concerned with the elimination of the consequences of Chornobyl nuclear accident, involving their properties as sorbents, ion exchange agents, adagulants and filtering materials, was proved to be extremely efficient. 6. A P P L I C A T I O N OF A D S O R P T I O N - A C T I V E M A T E R I A L S B A S E D ON N A T U R A L S O R B E N T S IN T H E P R O C E S S E S OF W A T E R PURIFICATION Extensive studies are performed recently concerned with the development of materials based on natural adsorbents. Sometimes the heat t r e a t m e n t or the substitution of natural exchange complex by other simple inorganic cation is quite sufficient to achieve the required improvement of the physicochemical and technical characteristics of materials based on natural disperse minerals. In the first case the structure of the mineral is changed predominantly (geometric modification); in the second case, the exchange complex is changed (ion-exchange modification). However, the improvement of natural minerals often involves several steps, leading to significant changes both in their structure and surface chemistry; in this way semisynthetic sorbents, a new class of sorption-active materials, are obtained. As defined in [136], semisynthetic sorbents are the

701 composite materials prepared from natural minerals via their chemisorption modification by organic and inorganic compounds, deposition of simple or complex oxides or other treatment, resulting in the formation of sorbents whose surface and porous structure differ from those characteristic for the original mineral, and combining the useful properties of natural and synthetic sorbents. In our recent publication [137], physicochemical principles underlying the production of new materials (sorbents, supports, catalysts, filtering powders, fillers for polymeric media) based on natural minerals are examined. In this section we consider the problems related to the preparation of new materials applicable for water purification, by simple or more complicated chemical t r e a t m e n t of natural minerals. 6.1. M o d i f i c a t i o n by a l u m i n i u m and iron salts In order to impart to natural sorbents the ability to adsorb organic and inorganic anions, these sorbents should be modified. This modification should preferably be performed using the substances and apparatuses which are themselves applied in the water t r e a t m e n t technology. From this point of view, it is desirable to modify the sorbents by aluminium sulphate, because A13§ ions and the products of their hydrolysis are strongly adsorbed by natural dispersed minerals, leading to the change of particle surface charge in water from negative to positive [71]. Perlite and bentonite modified by aluminium salts adsorb anionic SAS and dyestuffs from water rather efficiently [78,138]. In the technology of drinking water conditioning it is very important to remove fluorine ions from the water, because the increased (> 1 mg/dm 3) content of these ions causes various diseases [62]. In Ukraine some water sources (especially underground) contain 2-9 mg/dm 3 of fluorine ions; a number of water sources in the USA, Italy, Spain and England are also polluted with excessive amount of fluorine. All existing methods of removing fluorine-ions from water [62] have some disadvantages. The most promising method is the use of crushed clinoptilolite premodified by aluminium sulphate, acting as a specific absorbent of fluorine ions [56,139]. The testing of this method was performed with artesian water in Lubny town (Ukraine). The filter with clinoptilolite medium was employed, composed of granules 1-3 mm in size; filtering layer was 2 m thick. The modification of clinoptilolite was performed using 0.5% aluminium sulphate solution passed through the layer at a rate of 10 m/h for 1.0-1.5 h. The excessive amount of modifier was then washed out of the granulated medium with initial fluorine-containing water; this washing was terminated when the concentration of aluminium in the filtrate decreased to 0.5 mg/dm 3. The rinse water was collected in a special tank. At the final stage of filtering cycle, when the content of fluorine-ions in the filtrate had increased to 1.0-1.2 mg/dm 3, the rinse water was mixed with the initial water in a proportion (~ 1:14) which corresponded to aluminium concentration in the water supplied to the filters of ~ 1 mg/dm 3. This resulted in a considerable increase of the filter cycle.

702 At the filtration rate 5 m/h, approximately after 32 h of the filter operation, when the concentration of fluorine ions in the treated water increased to 1.01.1 mg/dm 3 (Table 17), the filter was switched over to the regeneration with aluminium sulphate, while the defluorination of water was performed by another filter, connected in parallel. The mean specific capacity of modified clinoptilolite is 500 mg/dm 3.

Table 17 Concentration (C) of fluorine in initial water and filtrate when providing water defluoration using clinoptilolite modified by aluminium sulphate Filter operating time, h 1 9 10 14 18 24 32

C, mg/dm 3 Initial water 2.0 2.2 2.3 2.1 2.2 2.4 2.1

Filtrate 0 0 0.15 0.5 0.8 0.7 1.1

There exist several mechanisms of fluorine ion absorption: anion exchange, formation of alumofluorine complexes and molecular adsorption of fluorine salts. These problems are discussed in more detail in [139]. The inversion of mineral particle charge from negative to positive means that strong sorption of hydroxo cations of aluminium takes place. This leads to the fact that it is not A13§ ions which act as counterions, but r a t h e r the hydroxyl groups of the hydroxo cations: i.e. the minerals acquire the ability for the anion exchange. The use of clinoptilolite modified by aluminium sulphate in the process of defluoration of natural waters is based on this ability. Studies of the interaction of aluminium and iron hydroxides or oxyhydroxides with clay minerals had shown that when a few percent of oxide is added to the clay, surface properties of the clay are controlled by the oxides. The amount of polycation required to produce the charge inversion is small. The results presented in [140] show clearly that A1 polycations are much more efficient in the blocking of negative charge than Fe polycations. It was supposed that this blockage is due not only to the higher charge of A1 polycations (z = 0.49 per gram-atom) as compared with Fe(III) polycations (z - 0.2 per gram-atom), but also to their shape. The spherical shape of Fe polycations limits the blocking of negative sites on the clay surface. On the contrary, planar A1 polycations cover a larger area, thus neutralising and blocking greater number of negative sites. These polycations can also occupy interlayer space of montmorillonite (see below).

703 However, the sorption of A1 polycations does not lead to the changes in the surface area of clay minerals with the rigid s t r u c t u r e - kaolinite, hydromica. On the other hand, the surface area of the kaolinite increases significantly (from 7.7 to 36.0 m2/g) when Fe polycations are added. This can possibly be ascribed to the 'porosity' of clusters formed by Fe polycations. The developed surface of kaolinite and hydromica clays modified by iron (III) salts, and more pronounced complexing properties of Fe 3§ ions as compared with A13§ ions, makes these sorbents promising for water t r e a t m e n t applications. For example, clays with Fe (III) polycations in the exchange complex are good adsorbents for the removal of anionic dyestuffs and non-ionic surfactants from water [2,137]. A remarkable success in the adsorption science was the development of methods for the preparation of semisynthetic microporous sorbents based on layer silicates with an expanded structural cell and basic salts of aluminium, titanium, chromium, etc., the synthesis of so-called pillared clays or pillared interlayer clay sorbents (PILC sorbents). Data were first presented in [141] concerning the sharp increase in the accessible surface and the formation of a thermostable material with open slit-like pores of width ~ ~ 0.8 nm as the result of the implantation of basic aluminium cations into the interlayer spaces of beidellite. In the former Soviet Union, a semisynthetic microporous sorbent based on montmorillonite and basic aluminium salt was obtained for the first time at our Institute [142]. Dozens of papers and patents have been published, dealing with the methods of preparation of PILC sorbents, studies of their properties and their practical application; some of these results are summarised in publications [143-145]. The data extracted from these papers, and also from [146], concerning the specific surface and effective pore radii in PILC sorbents, as compared with aluminosilicate catalysts possessing 13% A1203 produced by Davison Chemical Company (USA) are presented in Tables 18 and 19. It is seen that the sorbents obtained are generally microporous materials.

Table 18 Thickness of interlayer spacing Ad and specific surface area S determined from nitrogen adsorption in PILC sorbents based on montmorillonite Interlayer oligomer

Ad, nm

S, m2/g

[A1,304( ~ H)24] 7+

0.91

400

SisO12(OH)s

0.96

300

[Zr4(O H) ~Clx](~x)+

0.121

290

[TiO(OH)z] x

0.156

300

[Cr2,4(OAc)0,4(O H) 5,s]+

0.71

300

704 Table 19 Effective pore radius r and specific surface area S determined from nitrogen adsorption by All,~ - PILC sorbents and aluminosilica gel S, m2/g r, nm

A l i a - PILC

30 5 - 30 1- 5 <1 Total

1 4 109 362 476

Aluminosilica gel 0 7 568 0 575

Microporous sorbents based on montmorillonite and basic aluminium salts are most widely used. They have been prepared by the substitution of interlayer Na § or Ca 2§ cations of the initial mineral by oligomeric cations with composition varying from [A11304(OH)24(H20)12] 7+ to [AI1304(OH)30(H20)6] + [147]. Among these, the [All~O4(OH)2s(H20)s] ~+ cation is considered to be the most stable one in aqueous solution within pH range 3.7-4.3 and OH :A1 mole ratio range n = 1.2-2.3 [148]. Oligomeric A1132cations occupy about half of the internal surface of montmorillonite (S~ 7 5 0 m / g ) , making the remaining half of surface accessible for adsorption and catalysis. The advantages of the PILC-sorbents as compared with synthetic zeolites are their more opened porosity, which improves kinetics of the sorption and catalytic processes, and their relative cheapness, which is no less important for their application in water t r e a t m e n t processes. In what regards the water purification process, it is important that the mesopores also exist in the structure of (All3)-montmorillonite, with radius r = 1.65 nm [142]. The data obtained earlier [2,142] and presented in Table 20 clearly indicate that, when passing from the initial sodium to Al~3-montmorillonite, the hydrophility coefficient ~ sharply decreases, though certainly does not achieve the value characteristic for hydrophobic zeolites of ZSM-5 type, possessing SiO 2 :A1203 ratio 83 [149].

Table 20 Volume V of water (1) and hexane (2) adsorbed on the minerals at different relative pressure P/P~; hydrophility coefficient ~ of the minerals V, cm3/g, at P/Po Mineral 0.2 0.4 0.8 1 2 1 2 1 2 ~ - V1/V~ Na-Montmorillonite 0.034 0.015 0.092 0.019 0.253 0.032 7.91 A113-Montmorillonite 0.114 0.115 0.148 0.122 0.233 0.141 1.65 ZSM-5 0.037 0.186 0.048 0.186 0.080 0.212 0.38

705 In general, pure PILC-sorbents are hydrophilic materials. However, the presence of hydrophobic sites and co-ordination-unsaturated A13§ atoms belonging to All3 clusters creates favourable conditions for the sorption of a number of organic substances on these materials. For example, in [150] a selective sorption of dioxins by All~-Sorbents was reported. Molecules of smaller size, like phenol, are sorbed on PILC-sorbents from water less selectively as compared to their sorption by activated carbon [151,152], there exist, however, the methods capable for increasing the sorption capacity of these materials. One such method is an additional hydrophobization of the PILC-sorbents with long-chain cetylpyridinium cations [153]. Semisynthetic sorbents which possess a rather developed positive charged mesopore surface, adsorb large molecules of anionic dyestuffs much better than activated carbon does [154]. There exist efficient methods to increase the external, readily accessible surface of the PILC-sorbents [155]. Therefore, these materials may find their application in complex water treatment, for the removal of organic substances possessing different molecular masses.

6.2. P r e p a r a t i o n of c l i n o p t i l o l i t e m o d i f i e d by m a n g a n e s e d i o x i d e and its a p p l i c a t i o n for the r e m o v a l of iron and m a n g a n e s e ions f r o m artesian water Underground water specific to some regions of Ukraine, USA and other countries contains an increased amount of Fe(II) ions (up to 20 mg/dm 3) and Mn(II) ions (up to 3 mg/dm ~) [45,47]. Health standards approved in Ukraine require much lower content of these ions, 0.3 mg/dm 3 and 0.1 mg/dm 3, respectively. While Fe(II) ions can be removed from artesian water quite successfully using the aeration at the filters with grained filtering material [155], the existing methods for the removal of Mn(II) ions [45,155] are rather inefficient. Therefore it is necessary to develop special filtering materials which act simultaneously as the catalyst for Mn(II) ions oxidation, and the oxidant itself. The method for the formation of manganese dioxide film on the natural clinoptilolite tuff grains and the application of this material for the removal of residual iron ions and practically all manganese (II) ions from artesian water were presented in [156,157]. The 1-3 mm fraction of zeolite tuff (Sokirnitsa deposit, Ukraine) containing ca. 60-70% of rock-forming mineral (cation exchange capacity 1.29 mg-equiv/g) was used in the studies. To form the catalyst-oxidant MnO 2 film on the filtering material surface we had processed the clinoptilolite tuff grains with the solution of manganese nitrate salt; next the Mn(II) introduced into the zeolite by ion exchange was oxidised by potassium permanganate. To determine maximum amount of the MnO 2 film bonded tightly to the surface of the grains, samples with different quantity of Mn(II) introduced into the exchange complex (0.02 to 0.16 mg-equiv/g) were treated by 0.5% KMnO 4 solution in the presence of 1% NaC1 solution, and then

706 the excess of potassium p e r m a n g a n a t e and MnO 2 unbounded to the grain surface was washed out and evaluated. To evaluate the a m o u n t of m a n g a n e s e dioxide bonded tightly to the zeolite, the MnO 2 film was removed from the clinoptilolite surface in course of the t r e a t m e n t by the mixture of sulphuric acid and sodium sulphite Na2SO 3, and subsequent processing by KNO 3 solution; then the q u a n t i t y sought for was calculated from the content of Mn(II) ions in the resulting solution. The methods of preparation and application of modified clinoptilolite were tested both in laboratory and industrial regimes. The laboratory studies were performed in dynamic conditions using the model water which possessed no iron, but contained 2.44 mg/dm 3 of Mn(II), 2 mg-equiv/dm 3 of Ca(II) and 2.5 mg-equiv/dm 3 of Na § at pH 8.1. Glass column 10 m m in d i a m e t e r and 200 m m high was employed; the feed flow linear rate of cleaned w a t e r was 2 m/h. For the industrial testing actual a r t e s i a n water of Beregove region ( T r a n s c a r p a t h i a n Ukraine) was used, with the characteristics listed in Table 21. Almost all iron was preliminary removed from water by usual clinoptilolite filter with the aeration, which decreased the iron content in water from 4.8-7.0 mg/dm 3 to 0.5-0.05 mg/dm 3. To remove Mn(II) ions from water, the pilot plant equipped with a filter, 65 m m in diameter and 2 m load height was employed; feed flow rate was from 6 to 7 m/h. The weight of the filtering material was 6.6 kg, m e a n d i a m e t e r of loaded grains was from 1 to 3 mm.

Table 21 Quality p a r a m e t e r s of a r t e s i a n w a t e r prior to cleaning Fe (total) Fe(III) Mn Hardness Alkalinity Chlorides Sulphates

4.8 - 7.2 mg/dm ~ 0.15 - 0.20 mg/dm ~ 1.1 - 1.4 mg/dm 3 2.4 mg-equiv/g 2.4 mg-equiv/g 14.0 mg/dm 3 7 mg/dm 3

Ammonia Nitrates Nitrites SiO3 -2 Oxidizability (permanganate) pH

0.2 mg/dm 3 not detected traces 15.5 mg/dm 3 1.44 mg 0 2 / d m 3 6.9- 7.1

It was shown by special preliminary experiments t h a t to ensure the acceptable quality of modified clinoptilolite the a m o u n t of Mn 2§ ions introduced into its exchange complex should not exceed 0.15 mg-equiv/g; these ions are adsorbed mainly at external surface of zeolite grains. To accomplish this, clinoptilolite was processed by 1 % solution of Mn(N03)2; the a m o u n t of the solution used was 0.7 dm 3 per 1 kg of zeolite. To prepare Mn02.Cli, 0.3 % solution of Z M n O 4 was t r a n s u d e d through the layer of Mn-Cli grains at the rate of 2 m/h; the consumption of the solution was 1.1-1.2 dm ~ per 1 kg of zeolite. In the process of oxidation of adsorbed bicharged m a n g a n e s e by ZMn04, the formation of both tightly bonded and weakly bonded (physisorbed, i.e., removed

707 easily from the zeolite grains surface) manganese dioxide takes place at the surface of zeolite grains. The quantitative analysis of Mn (II) had shown that the amount of tightly bonded MnO2 was 0.4% by mass, see Figure 7. Mass portion of physisorbed MnO2 increases sharply when the amount (e) of Mn (II) introduced into the natural zeolite exchange complex exceeds 0.05 mg-equiv/g. For e = 0.15 mg-equiv/g, the amount of weakly bonded MnO2 was 0.1% by mass. External surface area of clinoptilolite grains was ca. 20 m2/g; the quantity of exchange cations at this surface was 0.08 mg-equiv/g. This leads to the conclusion that manganese dioxide is chemisorbed at external surface of zeolite grains only.

ach, %

a,h, %

05

-

0.10

0.4

-

0.08

0.3

-

0.06

-

0.04

-

0.02

0.2

"~

,...

0.1 I

I

0.05 0.10 0.15 e, mequiv/g Figure 7. Dependence of the amount of (ach) chemisorbed and (aph) physisorbed MnO2 on Mn(II) amount introduced into clinoptilolite exchange complex.

It can be supposed that oxygen atoms which belong to broken - S i - O- bonds existing at external clinoptilolite surface, are comprised into the structure of MnO 2 surface microclusters. Table 22 illustrates rather high performance of the removal of Mn(II) ions from artesian water using modified clinoptilolite. The MnO 2 film bonded tightly to the grains acts both as the catalyst for Mn(II) oxidation by oxygen, and as the efficient oxidant. Both factors lead to the transfer of dissolved Mn(II) ions into insoluble compounds of Mn(III) and Mn(IV). Active MnO2 deposed at the surface of clinoptilolite forms an intermediate complex MnO 2. 02 with oxygen dissolved in water. Thus the oxidation process can be schematically presented by following chemical reaction:

708

(5)

Mn(IV)O2 + Mn(II) --> Mn(III) + Mn(IV)

Oxidation properties of oxygen are due to the destruction of O - O bond in i n t e r m e d i a t e complex. The deposed m a n g a n e s e dioxide acts also as oxidant, transferring Mn(II) ions to insoluble m a n g a n e s e oxide: Mn 2§ + 3MnO2. Cli -~ 2Mn20 a 9 Cli

(6)

Table 22 Results of pilot studies for deironing and d e m a n g a n a t i o n of w a t e r using modified zeolite Mn 2§ mg/dm 3

Fe (total), mg/dm 3 Filtration time, h 1 24 48 97 121 142 162 168

Filtrate after

Initial deironing on water 4.9 4.9 5.1 5.2 5.0 5.2 5.1 5.0

natural zeolite 0.22 0.17 0.3 0.19 0.12 0.45 0.28 0.25

Filtrate after deironing on Initial modified water

zeolite 0.02 traces traces traces traces 0.03 0.13 0.25

1.25 1.16 1.34 1.37 1.18 1.42 1.44 1.42

Filtrate after Filtrate after 1st deironing demanganation on modified stage zeolite 1.05 1.06 1.21 1.22 1.12 1.37 1.38 1.35

n/d n/d n/d n/d n/d 0.02 0.05 0.10

n/d- not detected

It was shown by pilot experiments t h a t 1 kg of modified clinoptilolite is sufficient to clean 600-700 dm 3 of a r t e s i a n w a t e r without regeneration. The activity of the worked-off filtering m a t e r i a l with respect to Mn (II) can be completely restored by its regeneration with weak potassium p e r m a n g a n a t e solution. The consumption of 0.1% Z M n O 4 solution for the regeneration of catalytic and oxidising ability of 1 kg of modified clinoptilolite was 0.7-0.9 dm a, linear regeneration rate was 2 m/h, wash water consumption was 1.0-1.5 dm 3 per 1 kg of r e g e n e r a t e d material. Thus, modified clinoptilolite with high disperse m a n g a n e s e dioxide bonded tightly to the grains surface was shown to be efficiently used for the removal of residual iron ions and complete d e m a n g a n a t i o n of a r t e s i a n w a t e r during a r a t h e r long time.

6.3. M o d i f i c a t i o n of m i n e r a l s o r b e n t s by p o l y p h o s p h a t e s It is known t h a t polyphosphates possess high complexing properties with respect to heavy metal ions; for example, the stability constants of complexes

709 formed by Co 2+, Ni z+, Cu 2+ and Zn z+ ions with polyphosphates are 107-101~ [158]. One can expect t h a t mineral sorbents modified by polyphosphates will exhibit the selectivity with respect to heavy metal ions. We have developed the sorbents used for the removal of heavy metal ions from water on the basis of alumina, kaolinite and metakaolinite, this latter material is the product of the t h e r m a l t r e a t m e n t of kaolinite at 600~ [159]. The polyphosphate adsorbs tightly on positively charged sites at the surface of these minerals [160]. Kaolinite from Glukhov deposit (specific surface S ~ 70 m2/g) was used as the substrate of complexing sorbents with grafted polyphosphates. Its adsorption capacity at tripolyphosphate is 28mg/g; however, the adsorption from concentrated solutions of phosphates can amount to 200 mg/g or more, due to the polyeondensation reaction in which sorbed phosphate ions interact with nonsorbed ones [159]. Figure 8 shows Ni 2+ and Co z+ sorption isotherms and distribution coefficients for Glukhov kaolinite modified by sodium tripolyphosphate. A steep rise of the isotherms indicates t h a t the sorbent possesses a high selectivity with respect to

6

5

4

3

I

I

I

I

pH lgK,,

a, mg/g 500

-

5

40O

-

4

300

-

3

2O0

-

2

-

1

1

100 I I

I

I

I

50

1O0

150

200

C~, mg/dm~

Figure 8. Adsorption isotherms a (Ce) and distribution coefficient Kd(pH) dependencies for (1) Co 2§ and (2) Ni 2§ ions on kaolinite modified by sodium tripolyphosphate.

710 these ions. The values of distribution coefficients K d calculated for the Henry region at pH 6 were approximately 105 cm3/g, which correspond to the values characteristic to chelate synthetic sorbents. The selectivity decreases with the increase of acidity, see Figure 8. It can be thus concluded that the regeneration of phosphate complexing sorbent is possible. In fact, the washing of the sorbent with acid at pH 2 results in the desorption of virtually all cations which were previously adsorbed. The composition of (Me) : (P207) 4 complexes was shown to be 1 : 2 from the analysis of diffusion reflection spectra of the complexing sorbent with sorbed Ni 2§ and Cr 3+ ions [159]. Non-swelling kaolinite prepared by the granulation with oxynitrates and modified by tripolyphosphate was used for the purification of wastewater of galvanic plating section at Kievpribor electrical devices factory. The characteristics of wastewater were: pH 5.5-5.8, Ni 2§ concentration 200 mg/dm 3, the admixtures present: Na2SO 4, NaC1, H3BO 3 and formaldehyde. Static adsorption regime was used. The contact of 1 dm 3 of the wastewater with 5 g of the adsorbent resulted in the decrease of Ni 2§ concentration in water to 0.5 mg/dm 3. It can be concluded that complexing sorbents based on layer silicates and alumina modified by condensed phosphates can be used in fine purification of waters with respect to heavy metal ions. 6.4. D e v e l o p m e n t of oil-adsorbing materials and their application for

the removal of oil and petroleum-containing pollutants from water Oil and petroleum-containing products are among the most common water pollutants. To clean the water, a number of physicochemical methods is used: coagulation, flotation, filtration and adsorption. Application of filtration/adsorption methods for the removal of oil and petroleum products requires the development of new efficient materials. In this contribution the results are presented of the studies concerned with the preparation of hydrophobized bloated perlite, surface-porous adsorbent based on this bloated perlite, and coal-mineral adsorbents, and their application for the cleaning of water. 6.4.1. Hydrophobized bloated perlite Bloated perlite (0.10-0.35 mm fraction) which is used in the construction industry was taken as the initial material for the preparation of hydrophobized bloated perlite. This material was modified by water solutions (2-5 %) of sodium ethylsiliconate, sodium methylsiliconate, or water emulsions of polyethylhydrosilicone and polymethylhydrosilicone, with the concentration of 2-5 % with respect to the organic substance. The main reaction which takes place during the modification by polyalkylhydrosilicone is the interaction of hydride groups of the hydrophobizer with hydroxyl groups of the perlite surface:

711 R

R

I

I

-O-Si-H+HO-Si-

~

-O-Si-O-Si=

I

I

+ H2

(7)

Oligomeric organosilicones inoculated to the surface of perlite and other silica materials via the reaction-capable hydride and hydroxyl groups undergo the polycondensation process transforming into continuous polymeric film. This process of surface polycondensation can be enhanced by the temperature increase to 200-250~ The mechanism of chemisorptional inoculation of organosilicone modifiers to the silica materials is described in more details in [1,161]. We have studied the modification of a model a d s o r b e n t - geometrically homogeneous macroporous Silochrom with mean pore radius r = 57 nm [162], by polymethylhydride silicone, R3Si(OSiRH)nOSiR 3 (where R = C H 3, n = 10-15). Modification of the initial Silochrom was performed by the application of an optimum amount (2.5 mg/m 2) of the modifying substance from a 1% solution in toluene and subsequent thermal fixation of the modifying layer in a vacuum (1.3 N/m 2) over 2 h at 230-250~ Values of the specific surface area SAr, determined from low-temperature argon desorption were found to be 44 m2/g and 33 m2/g for the initial and modified silica, respectively. The physicochemical characteristics of the initial and modified Silochrom are presented in Table 23. The decrease in the specific surface area from 44 m2/g to 33 m2/g is due to a 'sealing-up' of the smaller pores with the modifying agent. It should be noted that the modified sample was hydrophobic: hence, the BET capacity of the arbitrary monolayer of adsorbed water, am, had decreased more than 10-times with respect to the initial sample, while the molecular area of the adsorbed water molecule, coil2o, had correspondingly increased and become equal to 3.04 nm 2. The specific heat of wetting of the modified Silochrom by water, q, had decreased to a value of 36 mJ/m 2, which is characteristic of graphitized carbon black [163]. An indication of the strong surface hydrophobization arising from polymethylhydride silicone processing is provided by the increase in the contact angle (0 = 95 ~ for the modified quartz as compared to 0 = 5 ~ for the nonmodified form). This value of the contact angle agrees well with the results obtained by Trau et al. [164] for quartz modified by the adsorption of various organosilicone substances from their solutions in hexane (0 = 82-96~

Table 23 Physicochemical characteristics of initial and modified Silochrom Silochrom

342r, m/g

am, mmol/g n _ C 6 H 1 4 H20

C0H2~, nm

q, mJ/m 2

Initial Modified

44 33

0.133 0.060

0.208 3.04

250 36

0.185 0.118

0, degree 5 95

712 For the preparation of hydrophobized perlite the pilot plant was constructed and tested jointly by Institute of Colloid Chemistry and Chemistry of Water and Kyiv Polytechnic Institute [1]. The performance of this plant was 30 m3/day. The operation of this plant is based on the impregnation of bloated perlite with the hydrophobizer solution in the hydrophobizing column followed by the drying in the air-fountain dryer using the products of natural gas combustion at temperature of 200-250~ At present two industrial plants for the preparation of hydrophobized perlite operate in Ukraine. The data presented in Table 24 show that the hydrophobization of perlite results in the formation of water repulsive substance (contact angle of the hydrophobic material with respect to the wetting by water is 95 ~ possessing the increased capacity with respect to oil.

Table 24 Comparative physicochemical and technological properties of initial and polymethylhydride silicone modified bloated perlite Property

Initial

Bulk density, kg/m 3 40- 250 170- 180 Water adsorbance per 24 hours, mass % Specific surface: 0.7 with respect to nitrogen, m2/g 10 with respect to water, m2/g 0.01 Limiting adsorption volume with respect to benzene, cm~/g 5 Contact wetting angle, degrees 0.5- 2 Oil absorption capacity, g(oil)/g(adsorbent) Sinks Buoyancy in water

Modified 40- 250 10- 20 1.0 7 0.07 95 6.0- 7.5 Floats on the surface

An important technological property of hydrophobized perlite is its buoyancy even when it is saturated with oil and petroleum products. This property is extensively used for the elimination of accidental petroleum discharges in water reservoirs. The systems are designed for the application of the adsorbent on oil slick and its collection together with bounded oil using special collecting ship. These developments are summarised in [165]. The specific feature of this adsorbent, its buoyancy in water, is used to enhance the flotational cleaning of waste waters containing petroleum products [166]. This method is based on the introduction of hydrophobic perlite (particle size 0.15-1.2 mm) in the stream of dispersed air into the flotation apparatus. The adsorbent particles together with air bubbles come up to the surface trapping the oil droplets. This leads to the increase of separation efficiency for oil-in-water emulsions as compared to traditional flotation method.

713 Hydrophobized perlite is employed also as the filtering material for precoat filters used for the removal of oil from technological water at nuclear power stations [167]. For the filtration rate of 2.8 m/h the dynamic oil capacity of modified perlite was found to be 0.2 g per 1 g of the adsorbent. This is 4 to 5 times higher t h a n the oil capacity of filter-perlite used in nuclear energetic. Therefore the application of hydrophobic perlite results in threefold increase of the filtering cycle and corresponding decrease of the amount of worked out radioactive slurry to be buried; in addition, the residual concentration of oil in water decreases from 1.0-0.7 to 0.3-0.5 mg/dm 3.

6.4.2. S u r f a c e - p o r o u s a d s o r b e n t b a s e d on h y d r o p h o b i c b l o a t e d perlite To increase the performance of the removal of oil products from water the method for the preparation of surface-porous adsorbent based on perlite modified by polymethylhydride silicone [168]. The essence of the method is the thermooxidation of the modifying layer at 320-350~ This thermoprocessing results in the removal of methyl groups [161] and the formation of surface micropores. The data obtained using the gas chromatographic version of molecular probes method [169] show that the dimensions of these micropores are approximately 0.6 x 1.2nm. Table 25 illustrates the sharp increase of the adsorbent specific surface and the improved performance of the material with respect to the removal of petroleum products from water. Table 25 Specific surface S of initial and polymethylhydrosilicone-modified perlite and the efficiency of its application for the removal of petroleum products from industrial waste water Perlite Initial Modified Modified Modified Modified

Treatment temperature, ~ 180 320 320 320

S, m2/g Ar 1.0 1.0 205 205 205

n-C6H14 2.0 8.0 105 105 105

Petroleum products concentration, mg/dm 3 Initial Final 1.0 1.0 0.5 1.0 2.9

1.0 0.8 0.2 0.2 0.3

The dynamic experiment in which the removal of petroleum products from water using the perlite was studied is discussed in [168]. It was found that the increase of petroleum products initial contents in water from 1 to 3-6 mg/dm 3 did not affect the cleaning depth, but results in the decrease of the time during which the protecting properties are retained. It appears expedient therefore to apply the

714 thermooxidated modified perlite at final stages of cleaning, when oil products contents does not exceed ca. 1 mg/dm 3. 6.4.3. C o a l - m i n e r a l a d s o r b e n t s Coal-mineral adsorbents (CMA) recently have been studied extensively because they are believed to be a promising class of adsorption-active materials [170]. Such materials prepared on the basis of adsorbents and catalysts coked in the adsorption catalytic processes can be successfully used for water cleaning [171]. It was shown in [172] that CMA represents the perfect grained material for the removal of emulgated oils from water. Its oil adsorption capacity (0.55 g/g) exceeds significantly that of quartz sand (0.06 g/g) and anthracite (0.2 g/g), and is only somewhat lower than that of active carbon (0.80 g/g). In the three stage filtration scheme the use of CMA enables one to decrease the contents of emulgated petroleum products in water from 24 to 0.56 mg/dm 3 [172]. The semi-industrial tests of CMA applicability for the cleaning of petroleumcontaining waste water were performed at the oil-refining plant at Ufa (Russia). The results of the tests are summarised in Table 26; the column (30 cm 3 volume) was packed by CMA fraction of 0.25-0.5 mm, linear filtration rate was 2 m/h. The regeneration of the adsorbent is r a t h e r simple and requires three volumes of hot water (80-90~ with pH value adjusted at 10 by soda, and subsequent processing with two volumes of hot and normal temperature water. After the cleaning of 200 volumes of water containing 60 mg/dm 3 of petroleum products and 340 mg/dm 3 of mechanical impurities the contents of petroleum products in the water cleaned with initial adsorbent had decreased to 11.5 mg/dm 3, and for the regenerated one this value was found to be 10.7 mg/dm 3.

Table 26 Removal of petroleum products and mechanical impurities from water using CMA Amount of water passed, dm 3 0 0.2 1.0 2.0 6.0 9.0

Petroleum products, mg/dm 3 86.0 3.6 4.7 8.4 20.0 18.0

Mechanical impurities, mg/dm 3 72 14 2 7 22 22

The results of these tests were used for the design and construction of the experimental furnace for the preparation of CMA from coked adsorbent and for

715 the development of the industrial method for cleaning of waste waters containing petroleum products. 7.

CONCLUSIONS

The data are presented concerning the structure and adsorption properties for principal types of natural adsorbents: clay minerals, zeolites, silica. Physicochemical backgrounds of the application of natural adsorbents for water purification with respect to disperse impurities and substances dissolved as molecules and ions are discussed. The methods and technologies are proposed for the use of natural clays and zeolites to remove disperse impurities, cationic and non-ionic surfactants, cationic dyes and protein substances, large-size cations Cs § K § NH4 § from water. Natural adsorbents are shown to be efficient in the deactivation of water, workwear, machinery and construction materials during the elimination of consequences of Chornobyl nuclear disaster in 1986-1987. The methods for the modification of natural adsorbents are developed which enhance the adsorption ability and selectivity of these materials with respect to fluorine anions, anionic dyes, oil and petroleum-containing products, heavy metal cations. The coked alumosilica adsorbents and catalysts are shown to be efficient materials for the preparation of coal-mineral adsorbents used in water purification. The data are presented concerning adsorption properties of pillarclays, with reference to their possible application for cleaning of water from organic impurities.

REFERENCES 1. 2. 3. 4. 5. 6. 7.

Yu.I. Tarasevich, Natural Sorbents in Water-Treatment Processes, Naukova Dumka, Kyiv, 1981 (in Russian). Yu.I. Tarasevich, J. Water Chem. and Technol.*, 16, 11 (1994) 21. V.G. Berezkin, V.P. Pakhomov and K.I. Sakodynski, Solid Supports in Gas Chromatography, Khimiya, Moscow, 1975 (in Russian). Yu.I. Tarasevich, Soviet Progress in Chemistry,** 44, 2 (1978) 16. V.T. Bykov, in: M.M. Dubinin and V.T. Bykov (eds.), Natural Sorbents, Nauka, Moscow, 1967, 77 (in Russian). N.I. Zalevsky, in: M.M. Dubinin and V.T. Bykov (eds.), Natural Sorbents, Nauka, Moscow, 1967, 112 (in Russian). Yu.I. Tarasevich and F.D. Ovcharenko, Adsorption sur des mineraux argileux, Inst. Fran~. Petrole, Malmaison, 1980.

* Before 1992 - Soviet J. Water Chem. and Technol. - English version of Khimiya and Technologiya Vody (Translation published by Allerton Press, Inc., 150 Fifth Avenue, New York, NY 10011; (212) 924-3950). ** After 1992 - Ukrainian Chem. J.- English version of Journal Ukrainskij Khimicheskij Zhurnal (Translation published by Allerton Press, Inc.).

716 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

27. 28. 29. 30. 31. 32. 33.

Yu.I. Tarasevich, Structure and Chemistry of the Surface of Layer Silicates, Naukova Dumka, Kyiv, 1986. R. Glaeser, S. Beguinot and J. M~ring, C. R. Acad. Sci., 274D (1972) 1. G.W. Brindley, in: W.D. Kingery (ed.), Ceramic Fabrication Processes, Wiley, New York, 1957, 7. J. Uytterhoeven, Bull. Groupe Franc. Argiles, 10 (1963) 69. H.G. Vasiljev and F.D. Ovcharenko, in: E. Wolfram (ed.), Proc. Intern. Conf. Colloid Surface Sci., Akad. Kiado, Budapest, 1975, 19. J.L. Ahlrichs, C. Serna and J.M. Serratosa, Clays and Clay Miner., 23 (1975) 119. E. Ruitz-Hitzky and J.J. Fripiat, Clays and Clay Miner., 24 (1976) 25. E. Mendelovici and D. Carros Portillo, Clays and Clay Miner., 24 (1976) 177. R.M. Barrer, Zeolites and Clay Minerals as Sorbents and Molecular Sieves, Academic Press, London- New York, 1978. G.V. Tsitsishvili, T.G. Andronikashvili, G.N. Kirov and L.D. Filizova, Natural Zeolites, Khimiya, Moscow, 1985 (in Russian). D.W. Breck, Zeolite Molecular Sieves, Inter. Sci. Publ., New York, 1974. A.A. Lur'ie, Sorbents and Chromatographic Supports, Khimiya, Moscow, 1972 (in Russian). L.F. Grigor'eva, T.A. Makarova, E.N. Korotkova and O.G. Chigareva, Synthetic Amphibolic Asbestos, Nauka, Leningrad, 1975 (in Russian). D.V. Kalinin, N.D. Deniskina and G.G. Lokhova, Amphibolic Asbestos, Its Synthesis and Origin in Nature, Nauka, Novosibirsk, 1975 (in Russian). W.A. Deer, R.A. Howie and J. Zussman, Rock-Forming Minerals, vol. 3: Sheet Silicates, Longmans, London, 1962. R.K. Ailer, The Colloid Chemistry of Silica and Silicates, Cornell University Press, Ithaca, New York, 1955. R.J. Atkinson, J. Colloid and Interface Sci., 42 (1973) 624. I. Kostov, Mineralogy, Oliver and Boyd, Edinburgh and London, 1968. T.G Plachenov, G.M. Belotserkovsky and V.F. Karel'skaya, in: F.D. Ovcharenko (ed.), Natural Mineral Sorbents, Akad. Sci. UkrSSR Press, Kyiv, 1960, 43. W.F. Vleesschauwer in: B.G. Linsen (ed.), Physical and Chemical Aspects of Adsorbents and Catalysts, Akademic Press, London-New York, 1970, 285. Yu.I. Tarasevich and V.S. Rak, Colloid Journal, 60 (1998) 129. V.A. Klyachko, I.E. Apertsin, Treatment of Water for Industrial and Municipal Water Supply, Stroyizdat, Moscow, 1962 (in Russian). L.A. Alekseev, Vodosnabzhenie i Sanitarnaya Tekhnika, II (1976) 33. M. Kahler and J. Voborsky, Filtration of Beer, SNTL-Technicke Literature, Praha, 1981 (in Czech). V.A. Zhuzhikov, Filtration, Khimiya, Moscow, 1968 (in Russian). V.G. Perevalov and V.A. Alekseeva, Purification of Waste Water of Oil Fields, Nedra, Moscow, 1969 (in Russian).

717 34.

35. 36. 37. 38. 39.

40. 41. 42. 43. 44. 45. 46. 47. 48.

49. 50. 51. 52. 53. 54. 55. 56.

57.

Yu.V. Kuznetsov, V.N. Tchebetovsky and A.G. Trusov, Principles of Water Purification from Radioactive Contaminations, Atomizdat, Moscow, 1974 (in Russian). R.W. Stephens, Public Works, 101 (1970) 68. H.W. Pohland, Desalination, 32 (1980) 157. Film-Tec Reverse Osmosis Systems, Pt.7: Pretreatmet for Film-Tec Reverse Osmosis Systems, Corporate Headquarters, Minneapolis, 1987. N.I. Dubnitskaya, S.A. Perlov, Bumazhnaya Promyshlennost', 6 (1987) 5. L.A. Kul'sky, I.T. Goronovsky, A.M. Koganovsky and M.A. Shevchenko, Reference Book of Properties, Methods of Analysis, and Treatment of Water, Naukova Dumka, Kyiv, 1980 (in Russian). A.C. Zettlemoyer and E. McCafferty, Croat. Chem. Acta, 45 (1973) 173. L.T. Zhuravlev, Langmuir, 3 (1987) 316. Ch.L. Thomas, Catalytic Processes and Proven Catalysts, Academic Press, New York and London, 1970. I.K. Skobeev, Filtering Materials, Nedra, Moscow, 1978 (in Russian). I.M. Leichkis, Filtration with Use of Filter-aid Materials, Tekhnika, Kyiv, 1975 (in Russian). Yu.I.Tarasevich, Soviet J. Water Chem. and Technol., 13, 7 (1991) 72. G.I. Nikoladze, Removal of Iron from Natural and Recycled Waters, Stroyizdat, Moscow, 1978 (in Russian). T. Viraraghavan, E.L. Winchester and G.L. Brown, J. Amer. Water Works Ass., 79, 8 (1987) 43. V.P. Shvedov, V.E. Ignatov and A.A. Kolirin, in: B.P. Nikolskyi (ed.), Inorganic Ion-exchanger Materials, Leningrad University Press, Leningrad, 1974, 294 (in Russian). A.I. Zhukov, I.L. Mongait and I.D. Rodziller, Methods of Treatment of Industrial Waste Water, Strojizdat, Moscow, 1977 (in Russian). R.I. Ayukaev and V.Z. Mel'tzer, Production and Use of Filtration Materials for Water Treatment, Strojizdat, Leningrad, 1985 (in Russian). S.P. Onatsky, Production of Claydite, Strojizdat, Moscow, 1987 (in Russian). Yu.I. Tarasevich, G.G. Rudenko, V.A. Kravchenko and V.E. Polyakov, Khimiya i Tekhnologiya Vody, 1 (1979) 66. Yu.I. Tarasevich, G.G. Rudenko and V.A. Kravchenko, Khimiya i Tekhnologiya Vody, 4 (1982) 169. Yu.I. Tarasevich, Soviet J. Water Chem. and Technol., 10, 3 (1988) 21. A.P. Kartashov, V.A. Kravchenko and Yu.I. Tarasevich, Soviet J. Water Chem. and Technol., 11, 7 (1989) 67. Yu.I. Tarasevich, in: Program and Abstracts, Zeolite'93, 4th Intern. Conf. on Occurrence, Properties and Utilization of Natural Zeolites, Boise, Idaho, 1993, 199. M.M. Senyavin, V.A. Nikashina and V.A. Tyurina, Soviet J. Water Chem. and Technol., 8, 6 (1986) 64.

718 58. 59. 60. 61. 62. 63. 64.

65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79.

80.

L.A. Kul'sky, P.P. Strokach, Technology for Treating Natural Waters, Vyshcha Shkola, Kyiv, 1986 (in Russian). J.C. Schippers, J. Verdouw and J.M. Hofman, Desalination, 32 (1980) 103. J.C. Schippers, H.C. Folmer, J. Verdouw and H.J. Scheerman, Desalination, 56 (1985) 109. Information of Ground Water Associates Inc., J. Amer. Water Works Assoc., 76, II (1984)67. E.F. Zolotova, G.Yu. Ass, Removal of Iron, Fluorine, Manganese and Hydrogen Sulfide from Water, Stroyizdat, Moscow, 1975 (in Russian). V.I. Stankyavichus, Removal of Iron from Water by Filtration, Moklas, Vilnnius, 1978 (in Russian). Yu.I. Tarasevich, V.E. Polyakov, I.G. Polyakova, Program and Abstracts: Zeolite'97, 5th Intern. Conf. on Occurence, Properties and Utilisation of Natural Zeolites, Naples, Italy (1997) 282. G.G. Rudenko, Yu.I. Tarasevich, V.A. Kravchenko and A. G. Sidorovich, Soviet J. Water Chem. and Technol., 5, 1 (1983) 73. L.A. Kul'sky, Theoretical Principles and Technology of Water Conditioning, Naukova Dumka, Kyiv, 1983 (in Russian). G.G. Rudenko, I.T. Goronovsky, The Removal of Impurities from Natural Waters at Water Supply Stations, Budivernik, Kyiv, 1976 (in Russian). Szenbanyszati tr6szt in Tatabansa, verfahren zum Kl~iren von feststoffwasser-suspensionen, Austria Patent No. 288 981 (1971). H. van Olphen, An Introduction to Clay Colloid Chemistry, Intersci. Publ., New York - London, 1963. D.J.A. Williams and K.P. Williams, J. Colloid and Interface Sci., 65 (1978) 79. A. Delgado, F. Gonzalez-Caballero and J.M. Bruque, J. Colloid and Interface Sci., 113 (1986) 203. S.L.Swartzen-Allen and E. Matijevid, J. Colloid and Interface Sci., 50 (1975) 143. H. van Olphen, J. Colloid Sci., 19 (1964) 313. B.V. Deryagin, Kolloid. Zhurn., 16 (1954) 425. A. Lux, Entropie, 11-12 (1972) 41. W. Brecht, H.L. Dalpke, Wochenblatt ftir Papierfabrication, II-12 (1972) 400. V.P. Svitel'sky, B.F. Ometsinsky, Yu.I. Tarasevich and V.M. Kovrizhkin, Khimiya i Tekhnologiya Vody, 3 (1981) 374. Yu.I. Tarasevich, V.E. Doroshenko and V.S. Rak, Soviet J. Water Chem. and Technol., 7, 5 (1985) 129. V.P. Svitersky, V.M. Kovrizhkin, Yu.I. Tarasevich and A.O. Orazmuradov, in: Cellulose, Paper and Cardboard, Lesprom, Moscow, I0 (1987) 25 (in Russian). G.A. Kozub, Yu.I. Tarasevich and V.E. Doroshenko, J. Water Chem. and Technol., 18, 5 (1996) 35.

719 81. V.T. Ostapenko, Yu.I. Tarasevich,, A.E. Kulishenko and N.A. Sinelnik, Soviet J. Water Chem. and Technol., 12, 9 (1990) 66. 82. Yu.I. Tarasevich, Soviet Progress in Chemistry, 43, 9 (1977) 930. 83. A.M. Koganovsky and N.A. Klimenko, Physicochemical Methods of Purification of Waste Water from Surface-Active Substances, Naukova Dumka, Kyiv, 1974 (in Russian). 84. V.V. Pushkarev, D.I. Trofimov, Physicochemical Features of Purification of Waste Water from SAS, Khimiya, Moscow, 1975 (in Russian). 85. Yu.Yu. Lur'ie and A.I. Rybinkova, Chemical Analysis of Industrial Sewage, Khimiya, Moscow, 1974 (in Russian). 86. Yu.I. Tarasevich, S.V. Bondarenko and A.I. Zhukova, Kolloid. Zhurn. 42, (1980) 1128. 87. D.J. Greenland and J.P. Quirk, in: Clays and Clay Minerals. Proc. 9th Natl. Conf., Pergamon Press, New York, 1962, 484. 88. W.D. Johns and P.K. Sen Gupta, Amer. Mineral., 52 (1967) 1706. 89. Yu.I. Tarasevich, G.V. Lantuch, A.I. Zhukova and S.V. Bondarenko, Teoret. i Eksperim. Khimiya, 18 (1982) 470. 90. Yu.I. Tarasevich, E.V. Aksenenko and S.V. Bondarenko, in: A. D@rowski and V.A. Tertykh (eds.), Adsorption on New and Modified Inorganic Sorbents, Elsevier, Amsterdam, 1996, 539. 91. W.U. Malik, S.K. Srivastava and D. Gupta, Clays and Clay Minerals, 9 (1972) 369. 92. A.N. Tsarev, M.A. Ponomarev and N.I. Dronova, Obogaschenie rud, 2(1973) 8. 93. Yu.I. Tarasevich, A.I. Zhukova and S.V. Bondarenko, Colloid Journal, 58 (1996) 522. 94. D.J. Greenland, J. Colloid. Sci., 18 (1963) 647. 95. A.A. Panasevich, G.M. Klimova and Yu.I. Tarasevich, Kolloid. Zhurn. 38, (1976) 1188. 96. G.M. Klimova, A.A. Panasevich and Yu.I. Tarasevich, Ukrainskij Khim. Zhurn., 44 (1978) 386. 97. R.L. Parfitt and D.J. Greenland, Clay Minerals, 8 (1970) 305. 98. G.M. Klimova, A.A. Panasevich, Yu.I. Tarasevich and E.G. Sivalov, Kolloid. Zhurn., 42 (1980) 238. 99. H. Schott, Kolloid. Z.-Z. Polym., 199 (1964) 158. I00. A.A. Panasevich, G.M. Klimova, V.P. Maksimova and Yu.I. Tarasevich, J. Water Chem. and Technol., 12, 12 (1990) 55. I01. G.M. Brindley and M. Rustom, Amer. Mineral., 43 (1958) 627. 102. M. H~ibner, Chem. Erde, 36, 8 (1977) 20. 103. Yu.I. Tarasevich, V.E. Polyakov, G.M. Klimova and A.A. Panasevich, Soviet Progress in Chemistry, 45, 5 (1979) 37. 104. Yu.I. Tarasevich, E.G. Sivalov and V.S. Rak, Khimiya i Tekhnologiya Vody, 2 (1980) 117.

720 105. V.E. Doroshenko, Yu.I. Tarasevich and E.G. Sivalov, Soviet J. Water Chem. and Technol., 4, 2 (1982) 19. 106. C.H. MSbius and T.H. Gfinther, Wochenblatt ffir Papierfabrication, 15 (1974) 559. 107. A.B. Kudin and O.N. Berman, Vodosnabzhenie i Sanitarnaya Tekhnika, 1 (1987) 17. 108. R.E. Grim, Applied Clay Mineralogy, McGraw-Hill, New York, 1962. 109. Yu.I. Tarasevich, V.A. Smirnova, L.I. Monahova and V.M. Ropot, Kolloid. Zhurn., 37 (1975) 912. II0. R.D. Harter and G. Stotzky, Soil Sci. Soc. Amer. Proc., 37 (1973) 116. 11 I. Yu.I. Tarasevich, V.A. Yurasova, L.I. Monahova and V.M. Ropot, Izv. Akad. Nauk Mold. SSR, Ser. Biol. i Khim., 6 (1986) 52. 112. C.B. Amphlett, Inorganic Ion Exchangers, Elsevier, Amsterdam, 1964. 113. F. Helfferich, Ion Exchange, McGraw-Hill, New York, 1962. 114. N.F. Chelishchev, V.F. Volodin and V.L. Krykov, Ion-Exchange Properties of Natural High-Silicon Zeolites, Nauka, Moscow, 1988 (in Russian). 115. Yu.I. Tarasevich, Soviet Progress in Chemistry, 44, II (1978) I. 116. Yu.I. Tarasevich, Soviet J. Water Chem. and Technol., 1 I, 4 (1989) 22. 117. R. Griesbach, Austauschadsorption in theorie und praxis, Akademie-Verlag, Berlin, 1957. 118. Yu.I. Tarasevich, V.E. Polyakov and L.I. Badekha, in: Proc. 1995 Intern. Conf. on Ion Exchange, Japan Ass. on Ion Exchange, Takamatsu, 1995, 349. 119. Yu.I. Tarasevich, Ukrainskij Khim. Zhurn., 63, 12 (1997) 98. 120. R.M. Barter and J. Klinowski, J. Chem. Soc. Faraday Trans., 68 (1972) 1956. 121. Yu.I. Tarasevich, M.V. Kardasheva and V.E. Polyakov, J. Water Chem. and Technol., 18, 4 (1996) I. 122. Yu.I. Tarasevich, M.V. Kardasheva and V.E. Polyakov, Colloid Journ., 59 (1997) 754. 123. K.B. Yatsimirskii and Ya.D. Lampeka, Physicochemistry of Complexes of Metals with Macrocyclic Ligands, Nauk. Dumka, Kyiv, 1985 (in Russian). 124. F.A. Mumpton (ed.), Mineralogy and Geology of Natural Zeolites, Intern. Committee Natur. Zeolites, New York, 1993. 125. L.L. Ames, Amer. Mineral., 45 (1960) 689. 126. B.W. Mercer, L.L. Ames and P.W. Smith, Nucl. Appl. and Techn., 8 (1970) 62. 127. S.M. Robinson, J.M. Begovich and C.B. Scott, J. Water Poll. Contr. Fed., 60 (1988) 2120. 128. S.M. Robinson, T.E. Kent and W.D. Arnold, in: Program and Abstracts, Zeolite'93, 4th Intern. Conf. on Occurrence, Properties and Utilization of Natural Zeolites, Boise, Idaho, 1993, 172. 129. B.W. Mercer, L.L. Ames, C.J. Touhill, W.J. van Slyke and R.B. Dean, J. Water Poll. Contr. Fed., 42 (1970) R95.

721 130. V.E. Polyakov, Yu.I. Tarasevich and M.M. Medvedev, Khimiya i Tekhnologiya Vody, 1, 2 (1979) 19. 131. P. Ciambelli, P. Corbo, V. De Simone and C. Porcelli, in: Zbornic Prednasok Konf. Slovzeo'84, Vysoke Tatry, Czechoslovakia, 1984, vol. 2, 69. 132. L.N. Zhir-Lebed, Yu.I. Tarasevich and V.E. Polyakov, in: Trudy Kaliningr. Tekhnol. Instit. Rybnoj Promishlennosti, Kaliningrad, 1985, 25 (in Russian). 133. Yu.I. Tarasevich, J. Water Chem. and Technol., 18, 3 (1996) 6. 134. V.S. Rak and Yu.I. Tarasevich, Ukrainskij Khim. Zhurn., 55 (1989) 799. 135. Yu.I. Tarasevich and V.S. Rak, Kolloid. Zhurn., 55 (1993) 160. 136. Yu.I. Tarasevich, in: M.M. Dubinin and V.V. Serpinsky (eds.), Adsorption and Adsorbents, Nauka, Moscow, 1987, 209 (in Russian). 137. Yu.I. Tarasevich, Theoret. and Experim. Chem.*, 32 (1996) 231. 138. Yu.I. Tarasevich, R.M. Verlinskaya, M.P. Nesterova and A.B. Gornitsky, Soviet J. Water Chem. and Technol., 8, 6 (1986) 34. 139. V.A. Kravchenko, N.D. Kravchenko, G.G. Rudenko and Yu.I. Tarasevich, Soviet J. Water Chem. and Technol., 12, 7 (1990) 99. 140. G.M. Oades, Clays and Clay Miner., 32 (1984) 49. 141. G.W. Brindley and R.E. Sempels, Clay Miner., 12 (1977) 229. 142. Yu.I. Tarasevich, V.E. Doroshenko, V.M. Rudenko and Z.G. Ivanova, Kolloid. Zhurn., 48 (1986) 505. 143. D.E.W. Vaughan, in: R. Burch (ed.), Pillared Clay: Catalysis Today, Elsevier, Amsterdam, 1988, vol. 2, 187. 144. D.E.W. Vaughan, in: W.H. Frank and T.E. White (eds.), Perspectives in Molecular Sieve Science, Amer. Chem. Soc., Washington, 1988, 308. 145. R.M. Barrer, Pure and Appl. Chem., 61 (1989) 1903. 146. A.J. Lopez, J.M. Rodriguez, P.O. Pastor, P.M. Torres and E.R. Castellon, Clays and Clay Miner., 41 (1993) 328. 147. R.A. Schoonheydt, J. van den Eynde, H. Tubbax, H. Leeman, M. Stuyckens, I. Lenotte and W.E.E. Stone, Clays and Clay Miner., 41 (1993) 598. 148. J.Y. Bottero, J.M. Cases, F. Fiessinger and J.E. Poirier, J. Phys. Chem., 84 (I 980) 2933. 149. Yu.I. Tarasevich, V.E. Polyakov, Kh. I. Minchev and V. Zh. Penchev, Kolloid. Zhurn., 55 (1993) 128. 150. T. Nolan, K.R. Srinivasan and H.S. Fogler, Clays and Clay Miner., 37 (1989) 487. 151. R.C. Zielke and T.J. Pinnavaia, Clays and Clay Miner., 36 (1988) 403. 152. V.E. Doroshenko, Yu.I. Tarasevich and G.A. Kozub, J. Water Chem. and Technol., 17, 6 (I 995) 7. 153. K.R. Srinivasan and H.S. Fogler, Clays and Clay Miner., 38 (1990) 287. 154. V.E. Doroshenko, Yu.I. Tarasevich and V.S. Rak, Soviet J. Water Chem. and Technol., 1 I, 6 (1989) 27.

* English version of Teoreticheskaya i Eksperimental'naya Khimiya (Translation published by Plenum Publishing Corporation, USA).

722 155. M.J. Hammer, Water and Waste-Water Technology, Interscience, NewYork, 1975. 156. Yu.I. Tarasevich, V.E. Doroshenko, V.M. Rudenko and Z.G. Ivanova, Soviet J. Water Chem. and Technol., 12, 5 (1990) 72. 157. V.E. Polyakov, I.G. Polyakova and Yu.I. Tarasevich, Khimiya i Tekhnologiya Vody, 19, 5 (1997) 483. 158. E.A.Prodan, L.I. Prodan and N.F. Ermolenko, Tripolyphosphates, Nauka i Tekhnika, Minsk, 1969 (in Russian). 159. G.M. Klimova and Yu.I. Tarasevich, J. Water Chem. and Technol., 14, 12 (1992) 37. 160. J.W. Lyons, J. Colloid. Sci., 19 (1964) 399. 161. A.V. Nazarenko, Yu.I. Tarasevich, S.V. Bondarenko and G.V. Lantuch, Theoret. and Experim. Chemistry, 25 (1989) 753. 162. Yu.I. Tarasevich, A.I. Zhukova, E.V. Aksenenko and S.V. Bondarenko, Adsorption Sci. and Technol., 15 (1997) 497. 163. A.C. Zettlemoyer, J. Colloid Interface Sci., 28 (1968) 343. 164. M. Trau, B.S. Murray, K. Grant and F. Grieser, J. Colloid Interface Sci., 148 (1992) 182. 165. O.V. Bezorud'ko, K.A. Zabela, A.A. Krupa, A.A. Pashchenko and Yu.I. Tarasevich, Hydrophobic Perlite and its Use for Purification of Water from Oil Impurities, Institute of Oil-Gas Industry, Moscow, 1985. 166. Yu.I. Tarasevich, L.K. Patiuk and A.A. Panasevich, Khimiya i Teknologiya Vody, 7, 6 (1985) 67. 167. L.K. Patiuk, Yu.I. Tarasevich and V.L. Zabolotnykh, Khimiya i Teknologiya Vody, 4 (1982) 546. 168. S.V. Bondarenko, A.V. Nazarenko and Yu.I. Tarasevich, Khimiya i Teknologiya Vody, 17 (1995) 500. 169. Yu.I. Tarasevich, A.I. Zhukova, E.V. Aksenenko, S.V. Bondarenko and A.V. Nazarenko, Adsorption Sci. and Technol., 10 (1993) 147. 170. R. Leboda and A. D~browski, in: A. D~browski and V.A. Tertykh (eds.), Adsorption on New and Modified Inorganic Sorbents, Amsterdam, Elsevier (1996), 115. 171. Yu.I. Tarasevich, Khimiya i Teknologiya Vody, 11 (1989) 789. 172. L.A. Kul'sky, Yu.I. Tarasevich, E.A. Shevchuk and Z.G. Ivanova, Khimiya i Teknologiya Vody, 12 (1990) 15.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

723

Activated carbon filtration in drinking water production: New developments and concepts S.G.J. Heijman and R. H o p m a n Kiwa Research and Consultancy, P.O. Box 1072, 3430 BB Nieuwegein, The Netherlands

1. INTRODUCTION Despite the efforts of the past ten years to prevent pollution, the concentration of pesticides in surface water in the Netherlands is increased. Polar pesticides are nowadays widely used because the accumulation in the food chain is a minor problem compared to their outmoded apolar counterparts. A disadvantage of this development is the fact t h a t these polar pesticides are easily washed to surface waters t h a t are used for drinking water production. For instance in the river Meuse in the year 1990 atrazine was detected in concentrations up to 1 pg/1 (Figure 1). Due to the negative publicity the use of this herbicide and consequently the concentrations in the surface water decreased in the following years. During this period the use of the herbicide diuron became more popular. 5,4

2.5

"~ diuron atrazine glyphosate AMPA

2.0

& ~

4,5

I

15

li

~i~ ~ i~ ~rll '

1.0

~

0

~

I

0

;r

II

0.5 /

'

I

!:

i,,

,

J , standard

0.0

-

1990

i(o,1 .g/~)

1991

1992

1993

1994

1995

Figure 1. Pesticides in the river Meuse, period 1990-1995.

1996

724 In 1992 this herbicide was first detected in considerable concentrations. After the 'top' year 1993 the concentrations of this pesticide decreased again, but now the concentrations of glyphosate and the metabolite amino-methyl phosphonic acid (AMPA) were increasing to peaks with more then 2 ~g/1. So the drinking water companies have to face pesticides with very different properties and with large concentration fluctuations. At the m o m e n t g r a n u l a r activated carbon is widely used to remove pesticides in the drinking water t r e a t m e n t process. In order to respond to new pesticides it is necessary to predict the g r a n u l a r activated carbon (GAC) lifetime. The performance of GAC depends not only on the properties of the pesticide molecules but also on the properties of the n a t u r a l organic m a t t e r (NOM) like humic acids and fulvic acids in the water. The concentration of the NOM is typically 1000 times higher compared to the concentration of the pesticide. For all n a t u r a l waters the NOM-concentration and NOM-composition differ, so full-scale experiments are necessary in order to predict the GAC-performance on a new w a t e r t r e a t m e n t location. Full-scale experiments should last at least until b r e a k t h r o u g h of the pesticide occurs (1-4 years). To avoid these expensive experiments some relatively 'fast' laboratory experiments, like bottle experiments and small-column experiments, together with adsorption models are used to predict b r e a k t h r o u g h curves on full-scale.

;:'

pesticide ~

NOM

..........

o

Figure 2. Schematic presentation of one of the mechanisms: NOM molecules block pores and in this way decrease the surface area of the granular activated carbon.

725 The main problem is to translate the outcome of the laboratory experiments to full-scale. The system with porous activated carbon grains and natural water is rather complex. The main difficulties are: o:- The size distribution of the NOM-molecules Typically there is a very large size distribution (MW-distribution from 100 to 10000) ~ The size distribution of the pores in the granular activated carbon Typically the pore size distribution varies from a few AngstrSm to about 1 pm. ~ The polarity distribution of the NOM-molecules. ~176The adsorption/desorption behaviour of the different pesticides ~176Because the NOM-front travels faster through the columns the pesticide molecules are always facing a covered surface (also called pre-loading). ~176The effect of pore blocking in the carbon by NOM-molecules (Figure 2). ~176Kinetic problems of the adsorption and desorption processes of both NOMmolecules and pesticides in the pores of GAC. The accuracy of the break-through prediction for full-scale granular activated carbon columns is rather poor at the moment. The problems with the used models will be discussed.

2. EQUILIBRIUM EXPERIMENTS Adsorption-isotherms are measured in order to get an indication-of the performance of activated carbon for the removal of pesticides. Recently Najm [11] introduced the Equivalent Background Concentration Model (EBC-model) in order to describe the competition between the pesticide molecules and the other organic molecules in natural water (NOM). The EBC-method is based on the postulation that the natural organic matter in natural waters can be replaced by one fictive organic component, which is competing with the pesticide. In fact average adsorption behaviour is defined for the NOM in competition with the pesticide. The adsorption behaviour of the fictive component is defined by EBC-characteristics: the Freundlich equation (q=Kc n) of the fictive adsorption isotherm. The EBC-characteristics are calculated using the Ideal Adsorbed Solution Theory (IAST) [13] for two competing components. The IAST-equations are derived from the Gibbs-equation, assuming that the spreading pressure in a single component solution is equal to that in a multicomponent mixture. The contribution of each adsorbed component to the spreading pressure in a multi-component system is proportional to the amount adsorbed of each component. The sum of the individual contributions equals the total spreading pressure in the mixture. For two competing components the IAST-equations are:

726 1

cl,e- I qlql+ q2I" [n~" I q--~ nl + q~ll nl

(I)

1

C2,e- I qlq2 + q21"[n~ "Iq--~ nl + q~)] n2 Cl,e ql q2 nl

K1 n2 C2,e

K2

: equilibrium concentration pesticide [~g/1] : equilibrium loading of activated carbon with pesticide [~g/mg] : fictive loading of activated carbon with NOM [pg/mg] : Freundlich constant of the pesticide in ultrapure water [ ] : Freundlich constant of the pesticide in ultrapure water [(~tg/mg)/(~tgfl) hi] : Freundlich constant of the NOM in the absence of pesticide [ ] : fictive equilibrium concentration of the NOM [pg/1] : Freundlich constant of the pesticide in ultrapure water [(pg/mg)/(~tgfl) n2]

_ C1,0- C l , e ql _

q2 -

(2)

(3)

ACD

C2,o- C2,e ACD

(4)

Two mass balances are also valid: C1,0 9 initial concentration of the pesticide [~g/1] C2,o 9initial concentration of the NOM [pg/1] ACD 9 active carbon dosage [mg/1] In order to calculate the EBC-characteristics (Ke and n2) of n a t u r a l water in competition with a pesticide two adsorption isotherms should be measured: one adsorption isotherm of the pesticide in ultrapure water (water without NOM) and one adsorption isotherm in the n a t u r a l water (two component system). The adsorption isotherm in the absence of NOM gives the single-solute Freundlich p a r a m e t e r s (nl and K1) of the pesticide (=component 1). In a bottle experiment different amounts of powdered activated carbon are added to the water with a known concentration of the pesticide. Because for each point in the isotherm the solid/liquid ratio is different the NOM/solid ratio is also changing. This is a normal procedure in water research because the pesticide concentration is dictated by the pollution in the n a t u r a l waters. The only parameter, which can be influenced, is the activated carbon dose. Although the solid/liquid ratio is not constant, the Freundlich isotherms obtained with this experimental set-up are linear within the experimental error. If the equilibrium concentration is in the region of the initial concentration (Cl,e~C1,0) the

727 experimental error is increasing. In this region a deviation of the linearity is observed, but this deviation can also be described with the EBC-model [11]. The equilibrium loadings (ql,e) of the pesticide are calculated from the bottle experiment by comparing the initial concentration (C1,0) with the equilibrium concentration (Cl,o) of the pesticide after seven days of equilibration time. The atrazine concentrations are m e a s u r e d with an ELISA-method with a detection limit of 0.05 gg/1. This adsorption isotherm is fitted in an optimizing procedure with the EBCcharacteristics (K2 and n2) as variables. The result can be used to calculate equilibrium concentration of the pesticides in different concentration ranges. In Figure 3 the adsorption isotherms in ultrapure water and in n a t u r a l water with a relatively high initial concentration (C1,0) of the pesticide is used to predict the adsorption isotherm in n a t u r a l water with a lower initial concentration of the pesticide. The major drawback of the EBC-method is t h a t only equilibrium conditions can be predicted. In a GAC-column there are no equilibrium adsorption conditions. Another restriction of the method is t h a t adsorption isotherms are m e a s u r e d with powdered activated carbon. The effect of pore blocking is probably more i m p o r t a n t in grains compared to powder. So the outcome of the EBCmethod will be to optimistic compared to full-scale r e s u l t s . O in ultrapure water 9 isoth. C1,0 - 93 pg/1 0 isoth. C 1 , 0 - 9gg/1 EBC-fit EBC-prediction

100

E 10 D

O ~

.13

0.1

'

I

I

-4D"

I

I

I

I

I

I

I I

I

I

I

I

I

I

I

I

I

I0

Cl,e (pg/O Figure 3. Verification of the Equivalent Background Concentration (EBC) model. The isotherms in ultrapure water and with an initial concentration of 93 gg/1 are used to obtain the EBC-constants (K2 and n2). The EBC-prediction calculated using these constants (dotted line) is a reasonable prediction of the experimental data with the same initial concentration.

728

3. THE HOMOGENOUS SURFACE D I F F U S I O N MODEL (HSDM) Crittenden et al. [1-4] first introduced the Rapid Small Scale Column Test (RSSCT). The idea is t h a t a full-scale column process can be scaled down to a small column experiment, if some dimensionless groups (P~clet, Stanton, and pore-diffusion modulus) are kept constant. After grinding the g r a n u l a r activated carbon is sieved in order to collect a fraction radius about ten times smaller compared to the radius of granular activated carbon used in full-scale columns. With the smaller grains the experimental duration is also decreased with about a factor ten. The dimensions and the process variables of the small-column are calculated using the Homogenous Surface Diffusion Model (HSDM) [15]. In this model the following three mechanisms are considered: 1) Axial dispersion in the bulk fluid of the column caused by diffusion or random fluid movement around the adsorbent particles. 2) External mass-transfer resistance or film transfer caused by the diffusion of the adsorbate from the bulk solution to the adsorbent surface. 3) I n t e r n a l mass-transfer resistance of pore and surface diffusion. To simulate a full-scale column, the amount of spreading in the b r e a k t h r o u g h curve must be identical for the small-column. In this way adsorption isotherms are not necessary. The passed water volume in the small column per unit activated carbon is directly related to the passed volume in a full-scale column. So extensive isotherm or kinetic studies are not required to obtain full-scale b r e a k t h r o u g h prediction from RSSCT. The major shortcomings in this model are: 1) In a large scale columns granular activated carbon is used with a broad size distribution of particles (for instance 0.2-1.4 mm). In the model only one particle-size is used for the calculations. So the particle-size of the full-scale column used in the calculations is more or less a matching parameter. 2) In the first models the intra-particle diffusion coefficient was assumed not to depend on the particle size (Constant Diffusivity or CD-model). In more recent studies [14] the intra-particle diffusivity was found to be proportional to the particle size (Particle dependent Diffusivity or PD-model). 3) Pore blocking by larger NOM-molecules is not incorporated in the model. Pore blocking is very difficult to handle in models as well as in laboratory experiments because the process depends on time and particle size. From our own observation (Figure 4) and from literature [5,10] we can conclude t h a t the breakthrough-curve predicted with the small-scale column test is in most cases an overestimation of the breakthrough-curve in the large columns. So the predicted granular activated carbon lifetime is almost always overestimated. The overestimation is depending on the n a t u r a l water type and the concentration and the properties of the NOM in the n a t u r a l water.

729

0.4 9 - ---

full scale column small scale column iI

~' 0.3

I I I I s f i

/I r/

o

0.2

i t

N

//~*\\ // /

/

iI \1

i

/

9~

/

0.1

/ / iI

=.;

/

9

I

/ iI

. . . . .

0

.11 _--__. .

! .

L...

40

t

I, !

80

I

I

J

L

120

1

160

I

200

passed volume (I/g) Figure 4. Verification of the validity of the homogenous surface diffusion model (constant diffusivity) in predicting the breakthrough curves of full-scale columns. In both full-scale as well as small-scale the initial concentration of bentazon was 0.4 pg/1. Norit ROW 0,8S, Natural water source: Drinking Water Company South-Holland-East, production location 'Het Kromme Gat'. qp=0.2-1.6 m m , rsp=0.l mm, qc=500 mm ,Lic=1400 mm, Qic=1828 l/h, rsc=9 mm, Lsc= 205 ram, Q~c=1 1/h. rip :radius large particle [mm] rsp :radius small particle [ram] Lic : length large column [ram] qc :radius large column [ram] QJc : flow rate in large column [l/h] L~c :length small column [mm] rsc : radius small column [mm] Qsc : flow rate in small column [l/h]

4.

PREDICTION OF BREAKTHROUGH-CURVES BASED ON E X P E R I M E N T S W I T H T H E F U L L G R A I N SIZE

Since the correlation between breakthrough curves in RSSC-tests and fullscale experiments is rather poor, the experiments are not used for predicting fullscale GAC-columns in drinking water production at the moment. In order to predict the lifetime of activated carbon filtration columns additional laboratory experiments are being developed. The most promising is an experiment with the

730 full grain-size as used in the full-scale adsorbers. The uptake of pesticides in the grains is measured as a function of time in a bottle-experiment. The advantage of this approach is that there are no scaling problems with respect to the grain-size. Disadvantage is that the equilibrium time can be as long as five weeks or more and it is necessary to translate the results of a batch experiment to a columnprocess. We have investigated the possibilities of a Linear Driving Force model (LDF-model) in order to predict breakthrough-curves from bottle experiments

[12]. dql - k(ql,e-ql (t)) dt with: k: Linear Driving Force constant

(5)

[s'l].

In the Linear Driving Force-equation (equation 5) the speed of accumulation of the pesticide on the activated carbon is a function of a constant and the difference between the equilibrium loading and the actual loading. To simulate different stages of preloading the natural water in a bottle is refreshed after five weeks of equilibration time. The amount of activated carbon is adjusted to simulate normal preloading in a full-scale column. For instance if 33 mg activated carbon is treated with four times a litre of natural water (three refreshments) the experiment simulates a preloading in the next experiment of 121 1/g. A typical result is shown in Figure 5. It is clear that the preloading influences the adsorption behaviour of the succeeding refreshments. The preloading as expected effects the equilibrium concentration. If we plot the measured speed of adsorption (dq/dt) against the loading (q) we obtain a linear relation (Figure 6). The slope of the lines (k, the LDF-constant) is clearly depending on the preloading of the activated carbon grains. With this experiment we get information about equilibrium loadings and adsorption rates in one experiment with full-size grains. We expect that the mechanism of pore blocking is better incorporated in our experiments and the prediction of full-scale columns is more accurate. With these experiments we hope to predict breakthrough-curves in full scale columns. The next step is to use the obtained LDF-constant in a mass balance incorporated in a partial differential equation [12,16] (Figure 7).

731

2.5

~

[]

2.0 ,

&

1.5

~ 1.0 ~ 0.5 0.0

i

0

i

i

i

i

i

i

1

100 200 300 400 500 600 700 800 900

time (h) Figure 5. Measured curves from succeeding refreshments in bottle experiments. C l,o=2 gg/1, ACD=33 mg/1, sieve fraction-0.6-0.7 mm, Chemviron F400 activated carbon. O 9 first litre; []" first refreshment; A: second refreshment; x" third refreshment.

0.4 0

E" 0.3 0.2 Q

<>

[]

0.1 z&

0.0 0.00 -

,

\

,

0.04

,

,

0.08

u"

0.12

0.16

0.20

loading (mg atrazine/gGAC) Figure 6. Same experiments as in Figure 5 but plotted according to the LDF-equation. The speed of adsorption is decreasing as a function of the preloading. O" first litre; rn: first refreshment; A: second refreshment; x" third refreshment.

732

Dynamic column model Differential fluid phase mass b a l a n c e 8Ci 6 DL + 8vq 6z 6z 8Z

_ ~

+ 8Ci+ 8t

8qi 6t

1-Sb) Sb

_ 0

dz

8qi 8t

,>

represents a set of mass balances and transport equations inside the adsorbent

Figure 7. Model used to translate the outcome of batch experiments to breakthrough curves in columns. The adsorption part of the equation (dq/dt) can be replaced by the LDF-equation (equation 5) with the measured dependence of k from the pre-loading and the measured equilibrium isotherm, z=axial coordinate; v-interstitial fluid velocity; ci=concentration of component i; qi=adsorbed amount of component i; ~b=porosity of packed bed.

1.4

...... 7 minutes - - - 10 minutes -_ 15 minutes 20 minutes

1.2 (].)

.~ 1.0

. . . . . . . .ii. . . . . . . . . .

9 9-"

/

i

~ 0.8 o

o,,

/

/

/

/

/

/

o 0.6 ,,.,,,,.,..,,,,,...,,,,'"

0.4

////

/ / / / / / / / /

/

0.2 0

0.0 0

50

100 150 200 250 300 350 400 4~50

time [days] Figure 8. Calculated breakthrough curves for four different empty bed contact times (7, 10, 15 and 20 minutes), using the model in Figure 7 and the outcome of the batch experiments in Figure 5.

733 The next step is to compare these calculated curves with the b r e a k t h r o u g h curves from pilot-scale experiments. At the moment we evaluate the validity of these model calculation for the prediction of b r e a k t h r o u g h behaviour in full-scale columns.

5.

THE RELATIVE B R E A K T H R O U G H MOMENT OF D I F F E R E N T P E S T I C I D E S AS M E A S U R E D W I T H T H E S M A L L S C A L E C O L U M N TESTS.

In the W a t e r T r e a t m e n t Companies there is a strong need for an easy and fast indication of the performance of granular activated carbon in order to remove different pesticides. Almost always the polarity of the pesticide in casu the octanol/water distribution constant (Kow) is used as a first indication. From Figure 9 we can see t h a t at least for the plotted pesticides a correlation between the polarity and the b r e a k t h r o u g h moment (Breakthrough-moment is defined as

the passed volume of a column (1/g) with a pesticide concentration in the effluent of the column of 0.1 lzg/l. At this point the drinking water companies have to regenerate their GAC-columns) in an RSSC-test is absent.

9 propazin [] bentazon diuron 9 atrazin

2.5 mcyanazin

9

DNOC

simazin

.o

9 pirimicarb 9 metribuzin

1.5 0

200

400

600

800

1000 1200 1400 1600

breakthrough moment (I/g) Figure 9. There is no correlation observed between the polarity of a pesticide and the breakthrough moments in the RSSC-test. Probably also other mechanisms play an important role in the competition between pesticides and NOM-molecules in granular activated carbon adsorption.

734 Bentazon and diuron have almost the same polarity but the GAC-lifetime for diuron is about 30 times larger compared to bentazon. Probably other p a r a m e t e r s as for instance the diffusivity and the size of the pesticides also play an i m p o r t a n t role in the competition between pesticides and NOM-molecules. The RSSC-tests can also be used for comparing b r e a k t h r o u g h behaviour of different pesticides. In the experiments in Figure 10 a mixture of five pesticides is dosed to a small column. In the experiment in Figure 11 a mixture of eight pesticides was dosed to a small column. In this way a large n u m b e r of pesticides have been screened for their relative b r e a k t h r o u g h behaviour (Table 1) in a short time. Because such a mixture of pesticides can be m e a s u r e d online with the same analytical method, this test offers a fast indication of the performance of a certain adsorbent in combination with a n a t u r a l w a t e r type.

35

s

30

--.- simazine

25

-- cyanazine 9 metribuzine - o - propazine

20 t_zi__. 5 ~

~

~

~ 1

0

50

~

u

~

~

I

~

~ I

100

150

~

~

~

i

200

I

250

I

300

350

passed volume (I/g) Figure 10. A mixture of triazines is dosed to a small column and the breakthrough of the

different species is measured online. Cl,o= 2~tg/1, rp=98 gm, Norit ROW 0.8S., natural water source: Nieuwegeins drinking water.

The concentration (in mg/1) of the NOM is about h u n d r e d times higher compared to the total a m o u n t of pesticides in this experiment. Therefore the m u t u a l influence of the pesticides on their b r e a k t h r o u g h behaviour is not significant. The m u t u a l influence of the pesticides in a mixture on the b r e a k t h r o u g h curve of atrazine is shown in Figure 12.

735

100 .3-

90 80

O "~

70

-o-

metribuzine

i

pirimicarb atrazine --m- DNOC diuron

60 II1

o

bentazon

50 40

3o 20 10 0

0

200

400

600

800

1000

1200

passed volume (I/g) Figure 11. A mixture of pesticides is dosed to a small column and the breakthrough of the different species is measured online. C~,o= 2gg/1, rp=98 gm, Norit ROW 0.8S., natural water source: Nieuwegeins drinking water.

Table 1 Relative b r e a k t h r o u g h of pesticides m e a s u r e d with an R S S C - t e s t Pesticide metribuzine propazine cyanazine atrazine simazine bentazon MCPP metribuzine pirimicarb metamitron diuron MCPA DNOC

Relative b r e a k t h r o u g h obtained from Figure 10 a n d 11 0.6 0.7 0.8 1.0 (by definition) 1.5 0.2 0.2* 0.7 0.9 >6 5.9 0.9* 2.2

* 9 same experimental conditions but with three times shorter empty bed contact time.

736

35

[

5 triazines mixture 9 pesticides only atrazine

I.,./// fjf~'J

25

..~

/

.~

/

//

20/1

.Q

[

15-

//

/

/

i/I/I/ //

10 -

,/

///

.

,/

5-

0

I

0

100

I

t

I

200

300

400

500

passed volume (I/g) Figure 12. The influences of different mixtures of pesticides on the breakthrough curve of atrazine in an RSSC-test. The concentration of NOM is about 1600 t.tg/1. In the mixture of nine pesticides the total concentration of pesticides is 18 lag/1. In the mixture of five triazines the total concentration of pesticides is 10 pg/1.

Although the absolute value of the b r e a k t h r o u g h varies with the experimental conditions, the order of b r e a k t h r o u g h of the different pesticides in the RSSC-test was very constant in all our experiments. The order of b r e a k t h r o u g h did not depend on the activated carbon type, the n a t u r a l water type and the empty bed contact time. Even in full-scale experiments (Figure 13) the p a t t e r n of b r e a k t h r o u g h of pesticides is the same. This encouraged us to define a relative breakthrough moment:

relative b r e a k t h r o u g h -

b r e a k t h r o u g h moment of the pesticide b r e a k t h r o u g h moment of atrazin

(6)

In this way we have a fast indication of the behaviour of different pesticides in GAC-columns. The n u m b e r of tabulated pesticides is increasing at the moment and a new pesticide can be put on this Scale after an experiment of about one month.

737

1.5 4 -

pirimicarb

zx

atrazine

AA

-~-_ bentazon ZX&

-~~A

A

o

0.5

/ , ,

~J~ /

0

-

A

0

&

&

. . . .

A

~

_

=

. ~.~__-.

~

9

,.,,

'

J

~

9

.

;~o~> ~

ip_jLi~= "o...>-~

~gO

A

0

~

A A A ~ n

A

&

~

i

50

~_--

100

,

i

150

200

250

time (days) Figure 13. Breakthrough behaviour of pesticides in a full-scale experiment, empty bed contact time = 7 minutes, Chemviron F400, Natural water source: Drinking Water Company Overijssel, location 'Vechterweerd', Ci,0 of the different pesticides was 1,5-2 gg/1.

It is an easy-to-use p a r a m e t e r , but there are also some shortcomings: In case of a bacteriological breakdown of pesticides the b r e a k t h r o u g h of the pesticide is postponed. This is for instance observed in a full-scale experiment for metribuzin. If the influence of the NOM in the n a t u r a l water is diminished (for instance after the removal of NOM with nanofiltration or with activated carbon fibres) the order of b r e a k t h r o u g h of pesticides is disturbed.

6.

N E W D E V E L O P M E N T S IN G A C - F I L T R A T I O N

G r a n u l a r Activated Carbon (GAC) filtration is a relative expensive t r e a t m e n t step in the production of drinking water. Therefore a lot of research projects are dealing with possibilities of how to extend GAC-life time. New developments in adsorptive removal of pesticides are all based on decreasing the influence of the N 0 M on the pesticide adsorption: * the NOM can be removed before GAC-filtration by m e m b r a n e filtration * the NOM adsorption is decreased by size-exclusion in activated carbon fibres * the NOM adsorption decreases after an ozone-treatment.

738 The rapid small-scale column (RSSC) test is a very powerful i n s t r u m e n t in the investigation of these developments. Of course the outcome of these comparing experiments only give an indication of the possible benefits in full-scale columns. 6.1. T h e c o m b i n a t i o n m e m b r a n e f i l t r a t i o n / G A C - f i l t r a t i o n

Water Supply Company of Overijssel (WMO) considers implementing a combination of nanofiltration and GAC-filtration for both the removal of pesticides and colour and softening of the water at the location Vechterweerd. A pre-treatment by nanofiltration has two main effects on the efficiency of the GAC-filtration for pesticide removal: the NOM level decreases drastically and only small NOM molecules enter the GAC step. This results in less competition between NOM and pesticides and less pore blocking and preloading by NOM. Figure 14 shows the results of the rapid small-scale column tests. From this figure it can be concluded that GAC lifetime time is extended enormously: about a factor of 100 !. By using nanofiltration as a pretreatment bentazon showed breakthrough after about 500.000 bedvolumes, while atrazine did not exceed the 0.1 ~g/1 even after more than 500.000 bedvolumes. Under practical conditions the GAC lifetime will even be longer because of the additional pesticide removal by the nanofiltration step.

0.8

0.6

--':~'-- atrazin RSF bentazon RSF --~--atrazin nanofiltr. - - i - . bentazon nanofiltr.

-

~ 0.4

- ~eriment A 1 / (NOM high) S o(j 0.2 -7' / ,/" 0 0

100.000

300.000

ExperimentB (NOM low) 500.000

bedvolumes Figure 14. Increasing GAC lifetime by NOM removal prior to GAC filtration: pesticide breakthrough for rapid sand filtrate (high NOM content) and nanofiltrate (low NOM content). C~,0=0.5 gg/1 Natural Water from location 'Vechterweerd'.

739

An alternative possibility is to use a very short EBCT (1-4 minutes) after nanofiltration to decrease i n v e s t m e n t costs for GAC-filtration. This option is currently u n d e r study at the pilot-plant Vechterweerd. To study the mechanisms, which lead to this significant increase of GAC lifetime in more detail, additional laboratory-experiments were performed. In these experiments diluted rapid sand filtrate (RSF) was used. The diluted RSF had the same NOM content as the nanofiltrate (1 mg/1) but the molecular size distribution in the diluted RSF and the nanofiltrate were obviously different: nanofiltrate contained only molecules with a d i a m e t e r smaller t h a n about 200400 D, while diluted RSF h a d the normal molecular size distribution, ranging from very small to very large molecules. Figure 15 shows the results of a smallscale column experiment with RSF, diluted RSF and nanofiltrate.

0.5

o...o oo .......................

0.6

o~ o: ~ -

/

i

O N

i

0.3 43

-i

.oOO o.o

..""'"

0

~ RS--F(-D~=7 mg/I)

,,,'"

-i

-.... dil. RSF (DOC = 1 mg/I) nanofiltrate (DOC= 1 mg/I)

i e

k.'o:'.., 0.2

,' ,' ,' ,' ,'

9

0.1

,' ,' ,'

0 0

,'

J

0

I

0

I

100

I

I

I

200

I

300

1

I

I

400

I

500

f I

I

I

600

bedvolumes (x 10a) Figure 15. Breakthrough curves for bentazon measured with RSSCT in: drinking water (RSF, NOM - 7 mg/1), diluted drinking water (diluted RSF, NOM = 1 rag/l) and nanofiltrate of RSF (NOM = 1 mg/1) C1,0= 0.5 gg/1, rp=98 gm, Chemviron F400, Natural water source: Drinking Water Company Overijssel, location 'Vechterweerd'.

Since the diluted drinking w a t e r and the nanofiltrate have the same organic carbon (=NOM) concentration one would expect the same breakthrough-curves for both w a t e r types. But with nanofiltrate the b r e a k t h r o u g h - m o m e n t is about ten times later. This is probably caused by the polydispersity of NOM (MW= 100 to 20.000) and the fact t h a t nanofiltration removes the larger size fraction NOMmolecules. This experiment is a strong indication for the importance of pore

740 blocking as a m e c h a n i s m in GAC-filtration. The larger NOM-molecules probably block the micropores (<2 nm) and cover in this way a relatively large adsorption area for the pesticide (MWbentazone-240). 6.2. A c t i v a t e d c a r b o n f i b r e s Another new development, which is evaluated with the RSSC-test, is the performance of new adsorbents as activated carbon fibres (ACF) (Figure 16). These fibres have a large surface area with almost only micropores with a pore radius smaller then 1 nm. One of the advantages of fibres is t h a t intra- particle diffusion problems normally occurring in granules are almost absent. Film diffusion is the rate determining mechanism. The optimal contact time of the m a t e r i a l is therefore very short compared to the g r a n u l a r activated carbon (GAC). A short contact time is very i m p o r t a n t because the carbon fibre columns can be r a t h e r small compared to the GAC-columns and therefore reduces i n v e s t m e n t costs. Another advantage is the larger b r e a k t h r o u g h - t i m e expressed in litres per gram activated carbon of the ACF compared to GAC [8] (Figure 17).

Figure 16. Electronmicrograph picture of activated carbon fibres. The diameter of the fibres is typically 10 gm. The BET-surface area is about 1200 to 1800 mZ/g and the fibres have a very narrow pore distribution of around about 1 nm.

741

o N

1.5

-o- ACF 2,60 s. ACF 3,20 s. ~ --D- ACF4,60s. ACF 6 , 3 0 s.

~

"

I - i - GAC_ !20min:

(i.1 .Q O

0.5 o o o

0

25.000

50.000

75.000

bedvolumes Figure 17. Bentazon removal by granular activated carbon (GAC-) filtration and activated carbon fibre (ACF-) filtration. Ci,0=2 gg/l, GAC = Norit ROW 0,8S, Natural water source: Nieuwegeins drinking water.

This is an effect of both the very fast adsorption kinetics, the higher surface area (40%-80% higher BET-surface) and the effect of size-exclusion of the larger NOM-molecules. A question at this m o m e n t is w h e t h e r or not the fibres can be regenerated after b r e a k t h r o u g h occurs.

6.3. C o m b i n a t i o n o z o n e - G A C - f i l t r a t i o n Another i m p o r t a n t development t h a t is evaluated with the RSSC-test is the biological activated carbon filtration [6]. Ozonation prior to GAC-filtration influences the biodegrability, the polarity and the size of the NOM-molecules. As a consequence activated carbon lifetime is increased significantly. In the RSSCtest [7] (Figure 18) the b r e a k t h r o u g h m o m e n t of atrazine is increased with about 60%, which is a very i m p o r t a n t benefit from economical and from environmental point of view.

742 0.4 - o - with ozone -n- without ozone

[3

/

/

0.3 iI I

/

0.2

]

/

0

I

1/

! I

q~ 0.1 o

0 0

50

100

150

200

250

passed volume (I/g) Figure 18. Atrazine removal from natural water pretreated with ozon and untreated water. Ozon dose = 2 mg/1, C1,0=5 ~tg/l. Norit ROW 0,8S, natural water source: Amsterdam Water Supply location "Leiduin".

7.

CONCLUSIONS

At the moment it is not possible to predict granular activated carbon filtration life time from short laboratory experiments. But experiments with full grain sizes showed some promising results. Pore-blocking by n a t u r a l organic m a t t e r (NOM) is an i m p o r t a n t mechanism in the adsorption behaviour of pesticides in a column. It is therefore i m p o r t a n t to incorporate this mechanism in the models or to take account of this problem in the laboratory experiments. Relative b r e a k t h r o u g h moments of pesticides can be obtained from rapid small scale column tests and can be used as an indication of b r e a k t h r o u g h in full scale columns. The rapid small scale column test is a powerful i n s t r u m e n t to evaluate new materials and new concepts. Promising water t r e a t m e n t concepts for the removal of pesticides are: * nanofiltration/GAC-filtration * ozone/GAC-filtration * activated carbon fibres.

743 ABBREVIATIONS ACF BAC GAC NF NOM RSF

:Activated Carbon Fibre :BiologicalActivated Carbon :Granular Activated Carbon : nanofiltration or nanofiltrate :Natu r al Organic Matter : Rapid Sand Filtration

REFERENCES 1. J.C. Crittenden, J.K. Berrigan and D.W. Hand, J. Water Pollution Control Federation, 58 (1986) 312. 2. J.C. Crittenden, N.J. Hutzler and J.S. Grierke, Water Resources Research,

22 (i 986) 285. 3. J.C. Crittenden J.K. Berrigan, D.W. Hand and B. Lykins, J. of Environmental Engineering, 113 (1987) 243. 4. J.C. Crittenden, D.W. Hand, H. Arora and B.W. Lykins, J. American Water Works Association, 79 (1987) 74. 5. J.C. Crittenden, P.S. Reddy, D.W. Hand and H. Arora, Prediction of GAC Performance Using Rapid Small-Scale Column Tests, AWWA Research Foundation, September, 1989. 6. A. Graveland, Application of biological activated carbon filtration at Amsterdam Water Supply, Water Supply, 14 (1993) 233. 7. R. Hopman, W.G. Siegers, M.A. Meerkerk and J.C. Kruithof, Proceedings IWSA-Workshop on Natural Organic Matter, Poitiers, France, September, 1996. 8. R. Hopman, W.G. Siegers and J.C. Kruithof, Proceedings AWWA Water Quality and Technical Conference, November (1995) 511. 9. R. Hopman, J.P. van der Hoek, J.M. van Paassen and J.C. Kruithof, IWSA 21th World Congress, Madrid, September 1997. 10. D.R.U. Knappe, V.L. Snoeyink, P. Roche, M.J. Prados and M.M. Bourbigot, Water Research, 31 (1997) 2899. 11. I.N. Najm, V.L. Snoeyink, Y. Richard, Journal AWWA (1991) 57. 12. R.G. Peel and A. Benedek, Canadian Journal of Chemical Enginering, 59 (i98i) 688. 13. C.J. Radke and J.M. Prausnitz, Amer.Inst.Chem.Eng.J., 18 (1971) 761. 14. R. Schneider, MS Thesis, University of Karlsruhe, Federal Republic of Germany (I 989). 15. H. Sontheimer, J.C. Crittenden and S. Summers, Activated Carbon for Water Treatment, second edition in English of "Adsorptionsverfahren zur Wasserreinigung", Karlsruhe (DVGW-Forsungsstelle) (1988). 16. B.M. Vliet and W.J. Weber, J. Water Pollution Control Federation, 53 (1981) II, P1585.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

745

Adsorption kinetics in natural waters: a generalised ion-exchange model G. Pan Department of Environmental Sciences, University of Plymouth, Plymouth PL4 8AA, United Kingdom

1.

INTRODUCTION

Adsorption of dissolved species onto suspended particles is widely accepted as a major sink of contaminants in rivers, lakes, estuaries, and oceans [1,2]. In order to understand the fate and distribution of the pollutants, which is of concern to human health, it is necessary to model quantitatively different adsorption behaviour. Ion exchange is known to be one of the major mechanisms in ionic solute adsorption and many investigations have been devoted to equilibrium models based on this perception. In addition to the application of the equilibrium method of adsorption isotherm (such as the Langmuir or Freundlich equations), surface complexation models have been developed over the last three decades which can predict the effect of various solution variables such as pH and ionic strength [3-7]. In natural water systems, however, equilibrium models are often inadequate because adsorption processes can be kinetically controlled when adsorption kinetics are not rapid relative to the gravitational settling of particles [8-10]. In many turbid water systems, adsorption kinetics can therefore be predominant in regulating the concentration levels of pollutants. Although ion exchange kinetics has been well developed in chemical engineering [11,12], there has been little application of these theories to natural water systems. The purpose of the paper is to develop a generalised ion exchange kinetic model, which may supply a theoretical framework for sorption kinetics modelling in natural water systems. The kinetics of ionic solute adsorption can be separated into four stages: (i) adsorption onto the outer-sphere of the hydrated surface; (2) diffusion through the liquid film; (3) surface reaction with activated groups on the solid surface; (4) intraparticle diffusion of the ionic solute accompanied by a progressive shell process towards the unreacted core. In batch rate experiments, the adsorption of metal ions often displays two-phase kinetics: very rapid initial adsorption over a few minutes, followed by a long period of much slower uptake. The second phase has often been observed to be much slower in natural water systems (a time scale of weeks or months [13-18]) than in pure metal hydroxide or mineral suspension systems, where relaxation times are within hours [19,20]. The first, fast, uptake

746 phase however is a characteristic of metal sorption in suspension systems, which is not always common in typical ion exchange systems. Among the four stages mentioned above, process (1) is a non-ion exchange step, which occurs quickly in the first instance, and is the cause of the initial rapid stage of the two-phase kinetics. Because of the complexity in natural systems, the overall kinetic time scale is generally determined by a combination of the other three ion exchange stages. In ordinary ion exchange processes where there are no accompanying chemical reactions and where the counterions in the solid bead diffuse simultaneously so that no moving boundary develops, the rates of the processes are controlled either by film diffusion, intraparticle diffusion, or a combination of both. These behaviours, which can be achieved under controlled experiments, can be well described by the BAM theory (Boyd, Adamson and Myers, [11]) established four decades ago. Film diffusion tends to be the rate-controlling mechanism when high capacity, smaller particles are in contact with slowly moving, dilute solutions. Conversely, intraparticle diffusion tends to dominate when low capacity, bigger particles are in contact with fast moving, concentrated solutions [12]. The initial stage of an ion exchange process is always controlled by film diffusion regardless of whether the later stage is con-trolled by film diffusion or particle diffusion [21]. Over the last two decades there have been two major advances in ion exchange kinetics. One is the discovery of the moving boundary phenomenon in which the shell of products moves towards the unreacted resin matrix when ion exchange processes are accompanied by chemical reactions [22]. The other is the development of various two-step-controlled models based on nonlinear or non-steady state assumptions [23-27]. Most of the latter models are mathematically too complex to be used in environmental chemistry. In this paper, a generalised, yet simple, ion-exchange model is obtained by using a specific mathematical approach, which yields an analytical solution for a nonsteady state, multi-step-controlled partial differential equation. The last fifteen years have seen a number of studies in the kinetic sorption of metals and anions in seawater and other natural waters. The major concern of these studies was to determine the uptake rates based on the overall adsorption kinetic curves [13-18,28,29]. A quantitative model was first proposed by Nyffeler et al. fifteen years ago, in which the sorption process was treated as a simple, overall first order, reversible reaction [15]. Based on this model, two-step or three-step first order reversible kinetic models were developed [9,17,30-32]. These models are mainly used for data fitting. They do not necessarily reflect the kinetic mechanisms of the process [33]. In fact, chemical reaction cannot be treated as the only control step in many cases, since various diffusion steps can be much slower than surface reactions. Further, sorption processes are seldom wholly reversible and few of them can be treated as elementary steps [34,35]. Because of the empirical nature of these models, they do not explain how the basic properties, such as flow status and particle size, can influence adsorption kinetics. To establish a modelling system based on the combination of various

747 transport steps and the surface reaction is therefore important in the field of environmental adsorption. 2.

EXPERIMENTAL

2.1. A d s o r p t i o n k i n e t i c s m o n i t o r A conventional method for monitoring adsorption kinetics in batch experiments is to withdraw aliquots of suspension at different times and then separate the particles from solution by either filtration or centrifugation to stop the reaction. The concentration of solute in the s u p e r n a t a n t can then be measured by different analytical methods. Such a sampling-separatingmeasuring method is tedious and time consuming and is often inadequate to study rapid initial sorption processes. Since the method normally consumes samples it can cause a considerable change in total volume and hence limit the number of experimental points that can be assembled. The lack of sufficient data points in the initial and overall stages of the process can seriously hinder the understanding of the mechanisms of the reaction, because the detailed shape of the kinetic curve is often a reflection of the mechanism. Different control steps can limit different stages of the same kinetic curve. Another experimental artefact that often accompanies this conventional method is non-settling colloids, which are the fine colloids remaining in solution after centrifugation or filtration. It is widely accepted that non-settling colloids can cause considerable errors in the determination of solid-liquid partition coefficients [35,36-38]. In the adsorption of metals in seawater or other electrolyte solutions, all the above mentioned problems can be easily solved by using a polarographic device [39]. This device enables polarography, which is usually used to measure samples of static solution, to be directly used as a monitor in constantly stirred systems. In this case, the artefact of non-settling colloids is eliminated because polarography is only sensitive to truly dissolved ionic species of metals [40,41]. It takes about 10 seconds for each measurement, no samples are consumed and no reagents are added. The assembling of kinetic data is therefore unlimited and can be automatically recorded every half minute. 2.2. M e t h o d s a n d m a t e r i a l s Detailed experimental methods have been published previously [39]. Briefly, adsorption kinetics of cadmium onto amorphous ferric hydroxide and T-MnOOH was studied in a controlled seawater system. The influences of pH, initial concentration of Cd, particle concentration, particle size, and pre-hydration of solids on the kinetic adsorption behaviour were examined. Total volume for each kinetic experiment was 2 litres throughout the experiments and stirring was kept constant.

748

3.

THEORETICAL

3.1. P h y s i c a l h y p o t h e s i s of the m o d e l (i) Individual particles are spherical and homogeneous and shrinking and swelling processes are negligible. (ii) Surface complexation or other fixation reactions can occur between the sorbed ion (A) and the activated surface groups. The kinetic impediment coefficient 0~) of the surface reaction is inversely proportional to the rate constant (k).

(iii) The ionic solute (A) first diffuses into the liquid film of the hydrated particles. The film is treated as planar and Fick's law is assumed to be applicable to both liquid film and solid phases. The diffusion coefficient in the liquid film (D) is generally treated as different from that in the solid phase (D). (iv) After the ion (A) diffuses through the liquid film and reacts with the surface group, a layer of product is formed. This layer is porous so that the reaction can proceed by the intraparticle diffusion of more solute ions through the layer. Thus, a shell progressive mechanism exists because of the continuation of the surface reaction. An intraparticle diffusion mechanism simultaneously exists as the reacted layer progresses towards the-centre of the core. In a simplified case, where the surface reaction is fast, the concentration of the ion within the particle can be treated as being in steady state, i.e., 0 % t = 0. Generally, a nonsteady state of 0C//~t = kC is assumed here. (v) The overall rate of the process is generally controlled by a combination of film diffusion, surface reaction, shell progressive process and intraparticle diffusion. (vi) The system approximately satisfies the infinite bath condition (i.e., total volume is large, with respect to the change of concentration). The above hypothesised adsorption mechanisms are schematically illustrated in Figure 1.

3.2. M a t h e m a t i c a l m o d e l According to the above hypothesis, the material balance for the adsorbate ion within the solid particle is represented by the following partial differential equation: I0 C 2 0C] 0C D --7-+ = r 5r 0t Or ~

(R > r > rc)

(1)

D represents the diffusion coefficient in the solid particle, and t time. The meaning of other symbols is illustrated in Figure 1. The overbars used here designate the solid phase.

749

Liquid film Solid surface Reacted shell Unreacted core cO 0 r (b 0 cO

-!--

u

CR co

CR

0

( 5 + R ) R roo r~ R ( R + f i )

Radial position Figure 1. Schematic illustration of the model in a spherical particle. Lines in the coodinate are theoretical concentration profiles in different phases.

According to the n o n - s t e a d y s t a t e h y p o t h e s i s (iv),

+r--

(2)

In order to solve this partial differential equation under the non-steady state condition, a specific approach is used here. A mathematical assumption is made as:

y = C. r

(3)

Then, y,

OC =~.r+C Or

--

(4)

750

a2C c3C y"=--.r+2~ (~r2 o~r Equation (2) can be r e a r r a n g e d as c~2C ~r 2

(5)

c~C 9r + 2 . . . .

/)r

D

C-r

(6)

Compare Equations (6), (5), and (3), a linear ordinary differential equation is obtained: Z y" = ~ . y

(7)

The general solution for equation (7) is

y = al e - ~ ' r

+ a2 e ~ ' r

(8)

The general solution for Equation (2) can be easily obtained by substituting (8) into (3): = al --e r

.r + a2 --e r

.r

(9)

The boundary condition used at the surface of the particle is (10)

D Equation (10) is the result of Fick's law in the liquid film, where k f = -~-, D is the diffusion coefficient in the liquid film and 8 is the thickness of the film. Kd is the equilibrium solid-liquid partition coefficient, which links the concentrations of A in the liquid film and in the solid particle. K d = CR CR

(11)

The boundary conditions used at the progressive shell between the reacted and unreacted portions of the particle are

751 (C)rc

=0

~{aC]

(12)

=

<--~Jr c

_Kd(drc/ < at )

(13)

The boundary condition used at the outer sphere surface between the liquid film and the bulk solution is C(R+5) = Cbulk

(14)

The initial condition is Cbulk It:0 : C ~

(15)

From the above boundary and initial conditions which are imposed by equations 10 to 15, a 1 and a 2 in the general solution of equation 9 can be obtained. a 1 = -a2e

-2~-~-rc

(16)

kfC~R a2

(17)

=

/ 1-z~ R

k : ~ / e - 2~'rc 9ehf-~R + ~ / kf_ KdD

1 + R~f-~/e- .~~.R R

Substitute (16) into (9), and differentiate it with respect to r,

r=rc

rc2e

1-r c

e

r2 l + r c

(18)

e

Substitute (18) into (13), and rearrange,

rc.e

-f~.( c -R)

-

I

1 e~'(rc -R) M +k2D~Df~C o R "rc "

drc

(19)

752

/

1-R~_

where M =

R

kf /

K d .D

kfC~R

Integrate (19) and rearrange,

t

IKdMR

Kd~ K d R ~ I Kd ~D(R-rc) K d ~ M ~-~(R-rc) kf~COR + kfCOR j - 2)~ Mrce - - -2)~ e

Kd Mrce~f~(rc-R)+ Kd ~ M ~-~(rc-R)e - ~Kd~-~ rce~D(rc-R) 2k 2k kfCOR

(20)

Kd D ~(rc ~ -R) e kfC~R~ Define: ~=rc R

(21)

where ~ represents the extent of the progress of the ion exchange in a particle. Whent=0, ~ = l ; w h e n t = ~ , ~=0. Substitute (21) into (20), assume RD >>SD, and expand the exponential function to the fourth term, KdR5 I(1 2)+ RD (1-3~2+ 2~3)+ R2L (~-3~3 + 3~4 -~5)1 t = ~-~~A -~ 3KdD5 -3~

(22)

Two important kinetic mechanism parameters are defined here: RD PfP = 3K dD5

(23)

R2~. - ~ Qc- 3D

(24)

Equation 22 becomes:

753 KdR8

t = 2D-~A [(1- ~

2)+ pfp(l_ 3~2+ 2~ 3)+ Qc(~- 3~ 3+ 3~ 4 - ~5)]

(25)

The time needed for reaching equilibration, when ~eq = 0, is" _ KdR8 teq - 2DC~ (1+ Pfp)

(26)

From (21), there is

3 = 1- Vreacted = 1- 0 Vtotal

(27)

where 0 is the adsorption fraction, which is the ratio between the reacted volume and the total volume of a solid particle. Substitute (26) and (27) into (25),

t T

-

-

1 -

M

teq

~

-

(28)

-

1 + Pfp

Equation 28 is the mathematical form of the generalised ion exchange model. In practise, the adsorption fraction is often experimentally determined by adsorption density: F 0=~ Feq

(29)

Thus, equation 28 can also be expressed as"

+-

+

teq t _

(30)

_

1+%

+QcI(l-1-q)3Fe /

1-F/4 3Feq)-/1- lFe Fq)

754 3.3. S i m p l i f i e d f o r m s o f t h e m o d e l Two kinetic m e c h a n i s m p a r a m e t e r s , Pfp and Qc, are obtained as a result of the theoretical deduction of the model. From their definitions it can be seen t h a t Pfp r e p r e s e n t s the ratio between the rates of film diffusion and particle diffusion, and Qc reflects the resistance of the surface reaction and the shell progressive process.

The higher Pfp, the more the adsorption tends to be controlled by

particle diffusion compared with film diffusion and vice versa. The larger Qc, the more the surface reaction contributes to the control steps. Equation (23) indicates t h a t film diffusion tends to control the adsorption process u n d e r conditions of high Kd, D , and 5 (i.e., high adsorption capacity of solids, fast diffusion in the particle and slow motion of solution) and low R and D (i.e., small particles and slow diffusion rate in the liquid phase). According to (24), surface reaction and shell progressive steps can limit the whole adsorption process to some extend if the reaction rate constant is low (i.e., ~ is high) and w h e n particle size is big and diffusion in the solids is slow. U n d e r different conditions Pfp and Qc can have different values and hence several simplified models can be deduced from the general model. These simplified models are useful in m a n y environmental and experimental systems and are discussed as follows. (a) W h e n Pfp = 0 and Qc = 0, equation 28 is reduced to: : 1 - ( 1 - 0)2~3

(31)

Equation 31 is a film diffusion controlled model with fast accompanying chemical reactions. It is comparable with BAlM theory [11]. (b) W h e n Pfp = ~ and Qc = 0, (28) is reduced to: = 1 - 3(1 - 0)2//33 + 2(1 - 0)

(32)

Equation 32 is a particle diffusion controlled model, which is the same as Nativ's theory [22] where surface reaction is fast. (c) W h e n Qc = 0, (28) is reduced to:

:

l+Pfp

{ [ 1 - (1-

]+ efp[ 1 - 3(1 -

§ 2(1 - 0) ] }

(3a

This is a combined film diffusion and particle diffusion controlled model, which is similar with Dana's steady state model [21] in nature. (d) W h e n Pfp = 0, (28) is reduced to: x : [ 1 - (1 - B)2~3 ]+ Q c [ (1 - o)l~3 - 3(1 - O)+ 3 0 - B)4~3 - (1 - B)5/~3 ]

(34)

755 This is a combined film diffusion and surface reaction controlled model. (e) W h e n Pfp = ~ and Qc v 0, (28) is reduced to: z" = 1 - 3(1 - 0)~ + 2(1 - O)

(35)

This model is the same as the particle diffusion controlled model (32). This is because w h e n intraparticle diffusion is predominantly slow, no more ions can reach the u n r e a c t e d core to continue the surface reaction, and the process is therefore limited by the particle diffusion. The theoretical kinetic curves generated by the general model and the simplified models are shown in Figure 2. The dotted line with open circles is controlled by film diffusion (31) and the dotted line with open squares is controlled by particle diffusion (32). The solid line with open triangles is controlled by a combination of film and particle diffusion (33) and it can be positioned a n y w h e r e between the two dotted lines. The solid lines with closed symbols are controlled by a combination of film diffusion, surface reaction, shell progressive progress and intraparticle diffusion. They can be positioned anywhere from the particle diffusion line to beyond the film diffusion line.

~

..o-

t

o

o ~

~

~

..

~176

.-'O'" 0.8

~

~

. (3"

Er

.-~"

~

o

~ ~

13~

C O

13"'

. u

o

o

1:r .~

o

.~.

0.6

,'/

Y

g (1) c

O~

o

oo-/,

d

~

) ---o--- Film diffusion model (Eq. 31) --.o--- Particle diffusion model (Eq. 32)

_c: 0.4 o x

ILl

Film-particle combined model (Eq. 33), P=1.5 0.2

r

General model (Eq. 28), P=5, Q=3

9

General model (Eq. 28), P=0.05, Q=0.5 I

0

0.2

0.4

Relative time

0.6

0.8

1

(t/toq)

Figure 2. Functional images of the general model and its simplified forms.

756 4.

RESULTS

Adsorption rates of c a d m i u m onto a m o r p h o u s ferric hydroxide and 7-MnOOH were m e a s u r e d in controlled s e a w a t e r systems. U n d e r some conditions the simplified film-particle diffusion model (33), which is a s i n g l e - p a r a m e t e r model, was enough to describe the adsorption kinetic curves. Results are p r e s e n t e d in Figures 3 and 4, where points r e p r e s e n t experimental data and lines r e p r e s e n t calculated results.

3

A

E 2.5-

I

o .u.~ -~ .o .u. []

.o-O -r~ -o ~3

2-o b

1.5-

P:,

Cd-FeOOH

j:::r ~

1

d -p'

2

0.5 <

0 0

20

40

60

80

100

t (mins) Figure 3. Kinetic adsorption of cadmium onto amorphous ferric hydroxide in a seawater system. (1) Open squares: Co = 20 ppm, Cp = 1.00 g/l, pH = 8.10, particle size = 60-80 mesh. Particles were pre-hydrated in seawater for 3 hours. The dotted line is calculated from equation 28. Prp = 3, Qc = 0, teq = 95 min, F e q = 2.6 mg/g. (2) Closed circles: C0=2ppm, Cp = 0.05 g/l, pH =7.00, particle size < 100 mesh, particles were not pre-hydrated. Solid line (eq. 28): Prp = 0.4, Qc = 0, teq = 101 min, F e q = 0.83 mg/g.

From equation 33 it can be seen t h a t if the model can describe e x p e r i m e n t a l data, t h e n a plot of t against

F /~3]]]}+PfP [ ( -Feq F/~3 l+Pfp{[1(1-F~q ) / leq teq

-

1-3 1

+2 l_=g_r

should

be a s t r a i g h t line with a slope of 1 and intercept of zero. The same principle holds for the general model, equation 28. Results in Figures 3 a n d 4 were introduced to equation 33, and the results were presented in Figures 5 and 6. The fact t h a t data of Figures 3 and 4 were t r a n s f o r m e d into s t r a i g h t lines indicated t h a t adsorption kinetics u n d e r these conditions was controlled by a combination of film diffusion and particle diffusion.

757

6

A

1 E

5

4 c

:

MnooH

3 o Q. o

2

0 _ %,

2

L_

1

0

1

I

I

30

60

90

120

t (rains)

Figure 4. Kinetic adsorption of cadmium onto ,/-MnOOH in a seawater system. (1) Open squares: Co = 20 ppm, Cp -- 2.00 g/l, pH = 6.00, particle size < 100 mesh, particles were not pre-hydrated. The dotted line is calculated from equation 28. Pfp = 6, Qc = 0, teq = 78 min, Feq = 4.98 mg/g. (2) Closed circles: Co = 2 ppm, Cp = 0.50 g/l, pH =8.00, particle size < 100 mesh, particles were pre-hydrated in seawater for 12 hours. Solid line (28): Pfp= 2.5, Qc = 0, teq =180 min, Feq = 2.64 mg/g.

120 Cd-FeOOH >-

9

8 0 - (1): y = 1.0584x - 2.0264 r2 = 0.9908 40-

r~~~v''g~~/ _~_ _ _

_~~~

0

-

-

20

,

,

40

60

(2):y = 0.9546x + 1.0881 , r2:09948

80

100

120

t (mins)

Figure 5. Test of mathematical model (eq. 28) for the kinetic data of Fig. 3.

y

.~_

l+Pfp

758 120 Cd-MnOOH >.

8 0 - (2): y = o.8649x + 5.3546 r2 = o . 9 9 1 ~ ~ ~ ~ ~

~I~

40_~~~,

-

,

,

(1): y = 0.8428x + 4.8283 , r2=098, 58

0-~' 0

20

40

60

80

100

120

t (rains) Figure 6. Test of mathematical model (eq. 28) for the kinetic data of Fig. 4.

5.

DISCUSSION

From Figure 2 it can be seen t h a t normal ion exchange is not an i n s t a n t a n e o u s process. However, m a n y adsorption processes can be fast so t h a t over half of the adsorption can be achieved within a very short initial period, forming a twophase kinetic curve. This is because, in addition to the ion-exchange processes which last throughout the whole range of x, a fast non-ion exchange process can occur initially, depending on the experimental or environmental conditions. This non-ion exchange process is closely related to static electronic interactions. For example, the fast initial kinetics can be enhanced when particles are smaller, more pre-hydrated and dispersed and more charged, and v i c e v e r s a . The non-ion exchange m e c h a n i s m is not considered in the ion exchange model presented here, because the purpose of the model is to reveal the major factors t h a t govern the overall adsorption kinetics by considering both transport and surface reaction processes. In the previous sorption kinetic models in n a t u r a l waters, sorption is t r e a t e d as one or more first order reactions, with no attention paid to various t r a n s p o r t processes. Models developed along the latter lines have to have no less t h a n four p a r a m e t e r s in order to describe empirically two-phase kinetic curves [9]. It is therefore i m p o r t a n t to establish a theoretical system so t h a t further empirical modelling can be developed on a rational physico-chemical basis. In fact, by incorporating a non-ion exchange mechanism into the model presented here, a semi-empirical, t w o - p a r a m e t e r analytical model was able to describe twophase adsorption kinetics. This work will be described in a separate paper. In principle, the two m e c h a n i s m p a r a m e t e r s Pfp and Qc can be theoretically calculated from basic properties of the system such as the diffusion coefficients, particle size, thickness of the liquid film, rate constants and partition coefficients. In practice, where these data are not available, a fitting of data is also of physical

759 meaning. For example, from the modelling results in Figures 3 to 6 it can be seen that the role of particle diffusion increased as the initial concentration and particle size increased, and the influence of the non-ion exchange was negligible under these conditions. Lower pH and shorter pre-hydration time of the solids may be partly responsible for the inhibition of a non-ion exchange process, because surface charge may be lower under these conditions [42, 43]. The fact that Qc equals zero indicated that surface reactions were fast which did limit sorption processes. A steady state t r e a t m e n t was therefore applicable to the sorption processes under the conditions specified in Figures 3 and 4. It must be noted that solid-liquid adsorption kinetics is greatly influenced by the experimental conditions. For the same system, the kinetics can be controlled by different steps under various conditions. One of the merits of the model developed here is that it can predict and explain these changes. Since it is an analytical model, it is easier to use than numerical models, and the modelling can be done using normal spreadsheet software (e.g. Lotus, Excel etc.). Over the last fifteen years, sorption kinetics in natural water systems has begun to receive increasing attention with many studies focussing on examining the equilibration time over a period of days to months. However, in most cases, experimental data assembled in the first few minutes to hours are too few to reflect the adsorption mechanism. Because of a lack of the mechanism knowledge, scant attention was paid, among previous work, to the fact that the reaction volume, stirring status, and even the shape of the container can influence adsorption kinetics. Theoretical modelling of adsorption kinetics is therefore not only of apparent importance for environmental engineering, but also for the design of experiments and the in-depth understanding of sorption processes in natural waters. 6.

A C Co Cp C D D k Kd

APPENDIX: GLOSSARY

P~

adsorbate ion concentration of A in solution phase initial concentration of solute particle concentration (mg/1) concentration of A in the solid phase diffusion coefficient in the liquid film diffusion coefficient in the solid phase rate constant of the surface reaction equilibrium solid-liquid partition coefficient RD kinetic mechanism parameter, Pfp = 3 K d ~ 5

Oc

R2~. kinetic mechanism parameter, Qc = 3--D-

R

radius of the solid particle

760

rc

t teq 0 F

radius of the unreacted solid core time equilibration time adsorption (or exchange) fraction adsorption density (mole/m 2 or mg/g) degree of ion exchange reaction, ~ = rc R relative time, x = ~~teq the kinetic impediment coefficient of surface reaction (inversely proportional to k) thickness of the liquid film

ACKNOWLEDGEMENTS

I thank Z. Zhang from the Department of Chemistry of the Ocean University of Qingdao for helpful advice and C. A. Lewis, J. Maskall and E. D. Stutt for proof reading.

REFERENCES

1. L.S. Balistrieri and J.W. Murray, Geochim. Cosmochim. Acta, 46 (1982) 1253. D. Koeppenkastrop and E.H. De Carlo, Environ. Sci. Technol., 27 (1993) 1796. W. Stumm, C.P. Huang and S.R. Jenkins, Croat. Chem. Acta., 42 (1970) 223. 4. D.E. Yates, S. Levine and T. W. Healy, Chem. Soc. Faraday Trans. I, 70 (1974) 1807. P.W. Schindler, B. Furst, B. Dick and P.U. Wolt, J. Colloid Interface Sci., 55 (1976) 469. J.A. Davis and J.O. Leckie, J. Colloid Interface Sci., 67 (1978) 90. 7. K.F. Hayes and Leckie, J. Colloid Interface Sci., 115 (1987) 564. 8. H.M. Liljestrand and Y.D. Lee, Wat. Sci. Tech., 23 (1991) 447. 9. R.N.J. Comans and D.E. Hockley, Geochim. Cosmochim. Acta, 56 (1992) 1157. I0. H.W. Jannasch, B.D. Honeyman and J.W. Murray, Limnol. Oceanogr., 41 (1996) 82. 1 I. G.E. Boyd, A.W. Adamson and L.S. Myers Jr., J. Amer. Chem. Soc., 69 (I 947) 2836. 12. F.G. Helfferich, Ion Exchange, McGraw-Hill, New York, N.Y., 1962. 13. L.S. Balistrieri and J.M. Murray, Geochim. Cosmochim. Acta, 48 (1984) 921. .

.

.

.

761 14. Y.H. Li, L. Burkhardt, M. Buchholz, P. O'Hara and P. Santschi, Geochim. Cosmochim. Acta, 48 (1984) 2011. 15. U.P. Nyffeler, Y.H. Li and P.H. Santschi, Geochim. Cosmochim. Acta, 48 (1984) 1513. 16. R.N.J. Comans and C.P.J. van Dijk, Nature, 336 (1988) 151. 17. H.W. Jannasch, B.D. Honeyman, L.S. Balistrieri and J.W. Murray, Geochim. Cosmochim. Acta, 52 (1988) 567. 18. C.C. Fuller, J.A. Davis and G.A. Waychunas, Geochim. Cosmochim. Acta, 57 (1993) 2271. 19. D.G. Kinniburgh and M.L. Jackson, in: Adsorption of Inorganics at SolidLiquid Interfaces, M.A. Anderson and A.J. Rubin (eds.), 91, Ann Arbor Science Publishers, 1981. 20. K.F. Hayes and J.O. Leckie, in: Geochemical Processes at Mineral Surfaces, J.A. Davis and K.F. Hayes (eds.), ACS symposium series 323, 114, American Chemical Society, 1986. 21. P.R. Dana and T.D. Wheelock, Ind. Eng. Chem., Fundam., 13 (1974) 20. 22. M. Nativ, S. Goldstein and G. Schmuckler, J. Inorg. Nucl. Chem., 37 (1975) 1951. 23. J. Span, J. Chem. Phys., 52 (1970) 3097. 24. T.C. Huang and K.Y. Li, Ind. Eng. Chem., Fundam., 12 (1973) 50. 25. T.C. Huang and F.N. Tsai, Can. J. Chem. Eng., 55 (1977) 301. 26. F.N. Tsai, J.S. Liu and T.C. Huang, J. Inorg. Nucl. Chem., 43 (1981) 265. 27. F.N. Tsai, J. Phys. Chem., 86 (1982) 2339. 28. J. Hamiltontaylor, M. Kelly, J.G. Titley and D.R. Turner, Geochim. Cosmochim. Acta, 57 (1993) 3367. 29. M. Fuhymann, H. Zhou, J. Neiheisel, M.A.A. Schoonen and R. Dyer, Sci. Total Environ., 202 (1997) 5. 30. U.P. Nyffeler, P.H. Santschi and Y.H. Li, Limnol. Oceanogr., 31 (1986) 277. 31. M.C. Hermosin, P. Martin and J. Cornejo, Environ. Sci. Technol., 27 (1993) 2606. 32. S.D.W. Comber, M.J. Gardner, A.M. Gunn and C. Whalley, Chemosphere, 33 (1996) 1027. 33. W.A. Kornicker and J.W. Morse, Geochim. Cosmochim. Acta, 55 (1991) 2159. 34. G. Pan and P.S. Liss, J. Colloid Interface Sci., 201 (1998) 71. 35. G. Pan and P.S. Liss, J. Colloid Interface Sci., 201 (1998) 77. 36. B.D. Honeyman and P.H. Santschi, Environ. Sci. Technol., 22 (1988) 862. 37. J.P. McKinley and E.A. Jenne, Environ. Sci. Technol., 25 (1991) 2082. 38. G. Benoit, Geochim. Cosmochim. Acta, 59 (1995) 2677. 39. G. Pan, M. Zhang and Z. Zhang, Mar. Chem., 30 (1990) 329. 40. F. Silva and C. Moura, Analyst, 119 (1994) 759. 41. M.L.S. Goncalves, L. Sigg and W. Stumm, Environ. Sci. Technol., 19 (1985) 141. 42. J.A. De Witt and T.G.M. van de Ven, Langmuir, 8 (1992) 788. 43. L.S. Balistrieri and J.W. Murray, ACS Symp. Ser., 93 (1979) 275.

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998Elsevier Science B.V. All rights reserved.

E n v i r o n m e n t a l Life C y c l e A s s e s s m e n t wastewater treatment techniques

763

(LCA) of t w o a d v a n c e d

D. N i j d a m l j. Blom I and J. A. Boere 2 1 Tauw Milieu, P.O. Box 133, 7400 AC Deventer, The Netherlands 2 NORIT Nederland B.V., P.O. Box 105, 300 AC Amersfoort, The Netherlands

Abstract An environmental life cycle assessment (LCA) was applied to two techniques for advanced w a s t e w a t e r treatment: granular activated carbon (GAC) on reactivation basis and advanced oxidation by means of ozone/UV. The LCA was set up for two situations: percolation water from a landfill (heavily contaminated flow) and HCH/chlorobenzenes contaminated groundwater (lightly contaminated flow). 1 m 3 of treated w a t e r was chosen as functional unit. The most i m p o r t a n t conclusions are that GAC has a significantly better environmental profile t h a n ozone/UV for the aspects studied. The environmental impact is dominated by: production and reactivation (GAC), in situ power consumption and the production of pure oxygen (ozone/UV). Key w o r d s life cycle assessment, granular activated carbon, GAC, ozone, ozone/UV, percolation water t r e a t m e n t , g r o u n d w a t e r t r e a t m e n t 1.

INTRODUCTION

Apart f~om technical and financial aspects, environmental considerations play an increasingly i m p o r t a n t role in the selection of water t r e a t m e n t techniques. In practice this choice is often made based an a non-systematic approach. An environmental life cycle assessment (LCA) represents a more objective method of determining the environmental efficiency of a certain operation. A LCA can yield information regarding techniques that are preferable from an environmental point of view. Furthermore, a LCA can serve optimisation of techniques aimed at improval of' environmental efficiency, based on objective and quantitative data. A LCA is described focussed on two techniques for advanced w a s t e w a t e r t r e a t m e n t . These techniques are adsorption on g r a n u l a r activated carbon (GAC), assuming central reactivation of the loaded activated carbon, and advanced chemical oxidation. A generally applied advanced oxidation technique for w a s t e w a t e r t r e a t m e n t was selected, i.e. oxidation with ozone combined with UV irradiation. The LCA was set up for two situations, focussed on the removal of contaminants with low biodegradability.

764 2.

B A C K G R O U N D S OF A L C A

In a LCA a product system is subjected to a total e n v i r o n m e n t a l analysis according to a defined method. This means t h a t for the whole life cycle of a product (from the extraction of raw materials to the waste stage) all environmental impacts are included and are quantified as far as possible. Apart from physical products, like packing material and consumer goods, systems like zoning plans, soil remediation, waste t r e a t m e n t methods and - like in this case - wastewater t r e a t m e n t techniques can be assessed by means of a LCA. Usually, the LCA method is applied in order to gain a better u n d e r s t a n d i n g of products' impact on the environment and for product innovations. LCAs were already executed in the seventies. However, until the early nineties there were no u n a m b i g u o u s guidelines; therefore, the results of different LCAs for the same subject could differ dramatically. Nowadays, generally accepted guidelines are available [1,2]. A LCA includes a survey of all the emissions, raw materials used, waste products released etc. of the production and related processes. Next, these so-called interventions in the environment are t r a n s l a t e d into e n v i r o n m e n t a l impact categories (greenhouse effect, acidification, deterioration of the ozone layer, etc.). Importancy factors are used, or in LCA terms: equivalence factors. Using these equivalence factors the interventions in the environment are t r a n s l a t e d into scores on the various environmental impact categories. The environmental profiles thus obtained, indicate the level of environmental impact for each product system based on the functional unit (the basis of comparison for different products or systems). In a LCA the following environmental aspects are included (Table 1).

Table 1 Survey of environmental aspects (underlined - incorporated in this study) biotic exhaustion abiotic exhaustion (depletion of abiotic resources) energy h u m a n toxicity ecotoxicity greenhouse effect photochemical oxidant formation (smog formation) acidification eutrophication odour, noise victims d a m a g e s to the landscape (final) waste

765 3.

D E S C R I P T I O N OF T H E C A S E S

For certain w a t e r flows g r a n u l a r activated carbon (GAC) and ozone/UV can be regarded as alternative t r e a t m e n t techniques. This, by the way, is a strong contrast to the t r e a t m e n t of drinking water, where ozone and GAC are highly complementary techniques. Both techniques (GAC and ozone/UV) are applicable on a broad range of water flows, e.g.: industrial waste waters, percolation water and groundwater. It often concerns polishing, for example following biological t r e a t m e n t and/or mechanical filtration. Two cases were selected for the LCA study, a relatively light and a heavily contaminated flow; thus, it should give an indication for a wide range of w a s t e w a t e r s containing poorly or nonbiodegradable dissolved organics. For backgrounds of both technologies, see f.i. [3 - 5].

Case I: biologically pre-treated percolation w a t e r E.g. at m a n y G e r m a n landfill sites, the percolate is t r e a t e d by advanced techniques. Figure I indicates a common procedure; as a final step GAC or ozone/UV is used. Based on realistic practical data the composition of the feed (before GAC or ozone/UV, so after pre-treatment) for case I is given in Table 2.

AEROBIC BIOLOGICAL TREATMENT (e.g. ACTIVE SLUDGE)

SAND FILTRATION

GAC

03 / UV

Figure 1. Treatment procedure percolation water (case I).

766 Table 2 Water composition case I flow (m3/h) AOX (mg/1) * feed

1.5

* effluent requirement

_<0.5

COD (mg/1) * feed

1000

* effluent requirement

_<200

Case II" Groundwater treatment Numerous contaminated groundwater sites exist in the (industrialized) world. Case II is based on a site where the groundwater is contaminated with HCHs and chlorobenzenes; this as a result of former production of insecticides, lindane f.i. Further dispersal of the contaminants is avoided by (among other methods) pumping up and treating the groundwater. The treatment procedure is summarized in Figure 2; the water composition for post treatment is given in Table 3.

AERATION

SEDIMENTATION

AEROBIC BIOLOGICAL TREATMENT (BIOROTOR)

SAND FILTRATION

I GAC

03 / UV

Figure 2. Treatment procedure groundwater (case II).

767 Table 3 W a t e r composition case II flow [m3./h)

42

chlorobenzenes (~g/1) * feed

30

* effluent r e q u i r e m e n t

_<1

HCHs (~g/1) * feed

30

* effluent r e q u i r e m e n t

_<0.02

COD (rag/l)

4.

* feed

25

* effluent r e q u i r e m e n t

_<20

D E S C R I P T I O N OF P O S T - T R E A T M E N T P R O C E S S E S

GAC GAC is a porous adsorbent with high internal surface (order of magnitude 1000 m2/g) and a particle size varying from about 0.5 to 3 mm. Organic substances are removed from the water primarily by physical adsorption. In w a s t e w a t e r t r e a t m e n t the GAC filtration mainly takes place in fixed beds, operated in up- or downflow (Figure 3). As soon as the carbon bed breaks through, the GAC is replaced by new or r e g e n e r a t e d product.

I I

I I

I

GAC

I

Figure 3. Schematic representation of GAC filtration.

GAC

768 If the acceptance criteria are met (based on the level of heavy metals f.i.), the loaded GAC can be regenerated by t h e r m a l reactivation. Given the specialistic n a t u r e and the high capital costs involved, the GAC is typically reactivated at central reactivation facilities (Figure 4). The loaded GAC is t r e a t e d at 900-1000~ adsorbed organic substances are pyrolysed and oxidized by m e a n s of s t e a m to CO2, H20 etc. Off-gases are posttreated, including t r e a t m e n t in an afterburner.

OFF GAS (TO AFTERBURNER) SPENT

NATURAL GAS

T

EI -4

(WET)--~:

A _ _,4

I_

- ~

~J.~ .....

AIR

--~--~/

WATER

S-TEAM

QUENCHING--~ WATER

u

REACTIVATED GAC

SPENT

WATER I~:l 1~ u u

STEAM

oo

'

WATER

'

A T/V

D

C ACATE

Figure 4. Example of reactivation kilns. An average loss of about 10 volume percent occurs as a result of attrition during t r a n s p o r t / h a n d l i n g and b u r n off during the t h e r m a l t r e a t m e n t . This loss is compensated by supply with either new GAC or discarded reactivated GAC from food and drinking water industries. Next, the reactivated GAC - including the make up - is reused for w a s t e w a t e r t r e a t m e n t . The most i m p o r t a n t system features for the 2 cases are s u m m a r i z e d in Table 4.

769

Table 4 System features of GAC filtration, case I and II; GAC grade NORIT NRS EA 0.5-1.5 flow (m3/h) GAC filter * volume GAC (m 3) * weight GAC (kg) GAC consumption * average load (m/m %)

Case I 3

Case II 42

18 6300

36 (2 * 18) 12.600

30 (COD)

0.22 (Cl-benzenes)

0.15 (H CHs) * use per m~ w a t e r (kg/m3) in situ power consumption * used pumping power (W) * use per m 3 w a t e r (kWh/m 3) distance to reactivation plant (km)

2.83

0.033

100 0.033

200 0.005

300

300

Ozone/UV Removal of organic substances by means of ozone/UV is based on oxidative degradation. Ozone is a strong oxidant, which is generated in situ from oxygen by means of electric discharge at high voltage (> 3000 V). Both air and pure oxygen can be used, but the latter is more common (costs!). The ozone gas is brought into contact with the water, in order to have ozone dissolved. Surplus, undissolved ozone can be removed from the gas phase; this gas flow can also be re-used. Additionally the w a t e r is exposed to UV light to stimulate the formation of OH radicals. The OH radicals are highly reactive. See Figure 5. oxygen supply , .................................................

-;

I lair

t

9 ,

b

ozone generator

dryer

*, i

ozone destructor

i i

INFLUENT~

...............

~

~

I l reactor

UV lamps

injector

@

Figure 5. Schematic representation of the ozone/UV system.

EFFLUENT ~

770

The relevant system features of ozone/UV for the 2 cases are s u m m a r i z e d in Table 5. The power consumption of the installation is expressed per kg ozone; so a p a r t from generating ozone, the electric energy required for UV lamps, cooling, pumps etc. is also included. Table 5 System features of ozone/UV Flow (m3/h) Ratio 03: COD (m/m) In situ power consumption installation * per kg O3 (kWh/kg 03) * per m 3 w a t e r (kWh/m 3) 02 consumption * per kg 03 (kg O J k g 03) * per m 3 w a t e r (kg O J m 3)

Case I 3 2:1

Case II 42 2.5:1

18 30.6

18 0.45

14.6 24.8

14.6 0.55

5. E N V I R O N M E N T A L A N A L Y S I S For both techniques - GAC and ozone/UV - a LCA was applied according to the method of [1]. In addition to the 'blank' (no treatment), three different options were evaluated: a. GAC, consisting of 100 % reactivated GAC a". GAC, consisting of 90 % reactivated GAC and 10 % new GAC b. ozone/UV c. no t r e a t m e n t (i.e. u n t r e a t e d discharge). 1 m 3 waste w a t e r was chosen as functional u n i t . Thus, the basis of comparison is 1 m 3 of a defined wastewater which is treated to a specific quality level. The basic assumptions used are explained briefly; s u m m a r y , see Table 6. General assumptions The environmental aspects included in the study are listed in Table 1. Components t h a t are not likely to have a significant contribution to the environmental impact were excluded. One single aspect - peat extraction - was approached qualitatively (see later). The production of capital goods was not t a k e n into account (not considered relevant in relation to the total). Ozone/UV The environmental aspects are dominated by the in situ consumption of (electric) energy and the production of liquid oxygen. The effects of t r a n s p o r t for the supply of oxygen were considered negligible. The generation of ozone is reported to require 8-9 kWh/kg. The total in situ energy consumption is determined at 18 kWh/kg ozone; this value represents the energy consumption of the whole installation, including ozone generator, UV lamps, pumps and the cooling system.

771 Table 6 Survey of processes as p a r t of the environmental analysis general - fuel extraction and recycling - generating electricity GAC - peat extraction - peat t r a n s p o r t - peat charcoal production p r i m a r y GAC production - secondary GAC production (reactivation) - use of t r e a t m e n t installation (discharge of t r e a t e d w a t e r included) - t r a n s p o r t of new, reactivated and used GAC ozone/UV system - production of oxygen - use of t r e a t m e n t installation (discharge of t r e a t e d w a t e r included)

T r e a t m e n t with ozone/UV results in the oxidative degradation (transformation) of substances. In this process, undesired by-products may be formed (f.i. the formation of bromate out of bromide). These effects are not included in this study. GAC The e n v i r o n m e n t a l analysis of GAC is more complex as compared to ozone/UV, due to the n u m b e r of process steps; production of new GAC, reactivation of loaded GAC, transport, and a (modest) in situ energy consumption. The release of substances as waste product (e.g. p a r t of the GAC attrition loss) is quantified; the effects of land filling (for example leaching) are not t a k e n into account. Emissions occurring during reactivation are related to the organic and inorganic loads based on the two cases. CO2 formed through oxidation of the organic load is not t a k e n into account; in the case of ozone/UV, CO2 is released in situ and is also not t a k e n into account. Obviously, C02 as a result of combustion of n a t u r a l gas and GAC is included. The production of binders for virgin GAC was not included in this study; however, the caloric value of these binders has been included. It was assumed t h a t peat is the raw m a t e r i a l for virgin GAC. The effects of peat extraction were evaluated only qualitatively. The raw peat is extracted from peat areas in Germany, which are mainly used for agricultural purposes. After the lots have been excavated they regain their agricultural purpose. The final use value of the land does not decrease, also because not all the peat is excavated. Nevertheless, there is a temporary disturbance, due to the excavation activities. Also, to a lesser degree, area's of ecological value are exploited for peat winning. By restoring the exploited areas afterwards, the original n a t u r a l value is m a x i m a l l y retained.

772

6.

EVALUATION

OF THE ENVIRONMENTAL

PROFILES

Environmental profiles The effect scores are summarized graphically in Figures 6 and 7. The higher the score, the higher the environmental impact. The highest score (highest environmental impact) for each environmental aspect is always set at 100 %.

100

100

100 0 I_ 0 0 t,t)

80

(/} 0 r

9r-~

60

0 (L) (~

_=

39

____--

40

=

0 I,._ (D Q.

35

_=

E

=

20

ii-o

0 0

........

/

~

J

t

I

9 la act carbon

I

i

oO+

f

..f

l

+-~

0

,<>o,

/

J

impact category [] la" act carbon

[] Ib ozone/UV

@ Ic no t r e a t m e n t

Figure 6. Environmental profiles case I (percolation water). Untreated discharge scores a low impact level for many environmental aspects. This is quite obvious as no raw materials or energy carriers are used. For the environmental aspects "eutrophication" and "aquatox" the untreated discharge shows a very high score. This is the result of COD, AOX and HCHs/chlorobenzenes emission. In case of ozone/UV the environmental impact is dominated by the electricity consumption and the production of oxygen. The waste is totally resulting from energy generation, most of all by coal-fired power stations (fly ash, bottom ash). Although these residues are largely marketed as secondary materials in The Netherlands, here they are still considered as waste products, because on a European level only a small percentage is marketed as secondary material. The environmental impact of GAC is much lower as compared to ozone/UV. The components related to production and reactivation of GAC dominate. Transports and in situ energy play a minor role. In most cases system a" (supplemented with 10 %

773

100

100

100

100

100

100

100

100 ]

100 0L 0 0

8O

f~=

SO

0

40

3S3~ 2r

2!

k

3s

0 L_ 2O

0

0.30.30.

0

0

0

.......

0

0

0

impact category 9 Ila act carbon

[] Ila" act carbon

[] lib ozone/UV

[] IIc no treatment

Figure 7. Environmental profiles case 11 (groundwater).

new GAC) yields slightly higher scores than system a (100 % reactivated GAC). So reactivation generally results in a lower environmental impact than the production of virgin GAC. Sensitivity to key assumptions Major key parameters in this study are the achievable load on the GAC and the energy consumption of the ozone/UV system: * GAC For the GAC grade involved, the load assumed (Table 4) is realistic; the levels were based on practical experiences. * ozone/UV Again, the electricity consumption (18 kWh/kg ozone) was deduced from practical experiences. However, data provided by equipment suppliers show variations in the range 14-22.5 kWh]kg ozone. The trends regarding the environmental impact of GAC versus ozone/UV are similar through the whole range (Figure 8 and 9).

774

100

100

8 oo

o

o

100

100

100

100

100

100

100 G) t.. 0 0 ee

80

oo e} tO~ "~

6O

0

40 1-

20

0

.

.

i

I

.

.

.

.

.

/

.

.

.

.

.

~'"

".

.

.

.

.

.

.

.

.

.

.

.

.+~176 ~o~ o<.+ /

i

~o

.

.

.

.

t

t i

impact category 9 la act carbon

[] la" act carbon

[] Ib ozone/UV 14 kWh

[] Ib ozone/UV 22.5 kWh

Figure 8. Effect of power consumption level, 14 versus 22.5 kWh/kg ozone; case I.

100

100

100

100

100

100

100

100

100

100

g

~.

_~

~=~

=

66

'

0 p,k',,

,..'."80~,

~,~

.

.0,~.0~.

,.#..e,

00~

@'~ @~

impact category

9 Ila act carbon 9 Ila" act carbon [] lib ozone/UV 14 kWh 9 lib ozone/UV 22.5 kWh

Figure 9. Effect of power consumption level, ] 4 versus 22.5 k W h / k g ozone; case r[.

.o,r .

775 7.

CONCLUSIONS

Untreated discharges yield a very negative score on eutrophication, aquatoxicity and - to some extent - h u m a n toxicity (case I). Obviously for the other impact categories, u n t r e a t e d discharge scores well (except for h u m a n toxicity in case II), because no energy and raw materials are used for treatment. GAC on reactivation basis scores markedly better t h a n ozone/UV. The environmental impact of GAC is dominated by production and reactivation: transport and in situ energy consumption are marginal. The environmental impact of ozone/UV is dominated by the in situ consumption of (electric) energy and the production of pure oxygen. Here also, the transport (of pure oxygen) plays a marginal role. Remarks The conclusions are valid for the cases described; the assumptions used are realistic. Technical developments in both treatment techniques - and combinations with other techniques - can influence such comparisons between techniques. Given the principles and assumptions, Figures 6 and 7 represent the situation for the cases described. Such a comparison, however, should never be taken too static; techniques develop and optimizing is possible through combinations with other treatment processes. Some examples: Ozone/UV * a trend towards lower electricity consumption by the ozone/U-V installation * catalytic ozonization (oxidation by ozone in presence of a catalyst) * ozone/UV, followed by biological treatment (procedure is being developed): organic substances are partially oxidized, then biodegraded. GAC It is possible to optimize GAC types and treatment process. As a result COD loads with a weight percentage of 40 are achieved already (assumption in case I: 30 %) REFERENCES

1. R. Heijungs (ed.), Environmental life cycle analysis of products, CML, Leiden, The Netherlands, 1992. 2. Environmental m a n a g e m e n t - Life cycle assessment- Principles and framework. ISO No. 140440, 1997. 3. B. Langlais, D.A. Reckhow and D.R. Brink (eds.), Ozone in water treatment. Application and engineering, Lewis Publ., USA, 1991. 4. H. Sontheimer, J.C. Crittenden and R.S. Summers (eds.), Activated carbon for water treatment, Univ. Karlsruhe, Germany, 1989. 5. J.A. Boere, Wastewater treatment using granular activated carbon. PAO Course, TU Delft, 1998.

Adsorption and its Applications in Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski(Editor) 9 1998Elsevier Science B.V. All rights reserved.

777

A p p l i c a t i o n of a c t i v a t e d c a r b o n s for t h e e n r i c h m e n t of toxic m e t a l s and their determination by atomic spectroscopy R. Dobrowolski M. Curie-Sktodowska University, Faculty of Chemistry, 20-031 Lublin, Poland

The main parameters influencing preconcentration processes and separation of trace amounts of metals on activated carbon are presented. The factors influencing the effectiveness of sorption of metal ions on activated carbons as well as the possibilities and methods of modification of those factors are discussed. The usefulness of sorption concentration and separation of trace amounts of heavy metals on activated carbons for their further determination by atomic spectrometry methods is evaluated. 1.

INTRODUCTION

Activated carbons are unique adsorbents because of their extended surface area, microporous structure, high adsorption capacity and high degree of surface reactivity. One of their important applications includes the enrichment of toxic metals from the solution for further determination by atomic spectrometry methods. In the case of adsorption of inorganic compounds on the activated carbon from aqueous solutions, the chemical nature of the adsorbent determined by the amount and nature of the surface complexes has, in general, more influence than the surface area and porosity of the adsorbent. Carbons are almost invariably associated with appreciable amounts of oxygen. The oxygen is fixed firmly and comes off only as oxides of carbon with heat treatment in vacuum or in an inert atmosphere. The phenomenon of ion adsorption on activated carbons is caused mainly by the presence of different oxygen surface functional groups. The equilibrium between these groups and the solution depends on the pH value of the solution. Adsorption of ions is affected by the surface charge of the particles of activated carbons and some specific interaction between particular kinds of surface groups and the ions. The most important phenomena which influence the ion adsorption capacity are ion-exchange, nonspecific sorption, surface precipitation redox reactions and formation of surface chelates [1-6]. Contribution of the above phenomena to ion adsorption depends on specific ion properties, kind of surface groups and their concentrations. The ion exchange properties of

778 activated carbons have main contribution to the total sorption capacity. Compared with many other ion-exchange materials, activated carbons are characterized by a number of advantages. Their specific ion-exchange properties may be modified in the desired direction and their ion-exchange capacity may be also controlled. Modification treatment consists in incorporation of different heteroatoms into the surface with simultaneous chemical bonding of these atoms with the carbon lattice. These treatments give rise to stable carbon-nitrogen, carbon-sulfur, carbon-chlorine or carbon-bromine surface compounds. On the other hand, outgassing treatment at high temperature (over 1100~ removes the surface functional groups and changes the anions-exchange capacity considerably. This modification causes change of the surface charge density sign and pH of carbon suspension in the water consistency. This results in the change of ion mechanism adsorption. From the analytical point of view, a preliminary separation and concentration of trace amounts of the elements in the samples for quantitative analysis by the atomic spectrometry method is often necessary. Particularly for the concentration of trace amounts of the elements contained in the concentrated electrolyte solution the use of highly selective adsorbents is necessary. The activated carbons characterized by the vast polydispersive porous structure and high susceptibility to chemical modification of their surface are useful for these purposes. Specific electrochemical and adsorption properties of carbons resulting from their microporous structure and surface chemical properties [2,5] cause that they are irreplaceable in these analytical operations. 2. ACTIVATED CARBON MODIFICATION FOR THE ANALYTICAL PURPOSES Activated carbons usually contain, apart from the crystals, mineral substances occluded in the pores, described often as the ash content. Depending on the type of raw material used for manufacturing activated carbons the ash content is from 0.5 to 20%. The commonly used method of impurities removing is leaching of activated carbons with acids. Owing to the complex composition of mineral additives, the mixture of hydrochloric and hydrofluoric acids is used. Application of activated carbons for a preliminary separation and concentration of trace amounts of the elements and their detrmination by atomic spectroscopy methods involves very fine purification of carbons from mineral additives. In the presented experiments the powder activated carbons No. 1860 manufactured by Merck, Germany (sample M) and Medical Carbon (Carbo Medicinalis) produced by The Dry Destillation of Wood Plant in Hajnowka, Poland (sample MC) were studied. Activated carbon No. 2186 as delivered is often used for preconcentration and separation of different inorganic and organic substances, however, in the case of its application for trace metals analysis by slurry sampling graphite furnace atomic absorption spectrometry technique the fine purification is required because the carbon slury is introduced directly into

779 furnace of an AAS spectrometer. Commercially available carbons were leached by concentrated hydrochloric acid using a Soxhlet a p p a r a t u s made of quartz and by hydrofluoric acid in the Teflon vessel. In Table 1 the m a i n components and the effect of this purification are shown. The m a i n components were leached to the ppm level, however, Fe, Si, A1 are still present at a relatively high level. Table 1 Main components and effects of carbon purification Content (ppm) Carbon

M MN MC MCN

Fe

A1.

K

550 32.5 660 28.4

208 14.3 135 10.2

136 2.5 85 1.8

Si

P

800 120 360 82

136 85 72 48

S 520 220 45 21.8

Ca

Mg

745 242 12.5 8.1 1250 125 15.2. 5.6

Mn

Cu

112 2.8 153 8.2

15.9 0.05 11.3 0.02

The studied carbons were modified using two different methods e.g. oxidation by 13% H202 (sample MO or MCO) or by concentrated nitric acid (sample MN or MCN). The procedure was as follows: 500 cm 3 of perhydrol or concentrated nitric acid were added to 50 grams of the activated carbon and the suspension was h e a t e d at 90~ until dry. The residue was washed with distilled water until conductivity of the w a t e r eluent was close to t h a t of the distilled water. The effects of surface changes were examined by the transmission FTIR spectra of these three samples using a spectrophotometer with a P e r k i n Elmer 1725X Fourier Transform Infrared Spectrophotometer. For each spectrum, 250 scans were made at a resolution of 2 cm -1. No m a t h e m a t i c a l smoothing functions were performed on the spectra; however, the ordinate of each spectrum was normalized to facilitate comparisons between the spectra. The pellet of KBr containing about 0.02% of carbon was used. The pellets were dried overnight at 120~ before the spectra were measured. These activated carbon samples were characterized by N2 adsorption a t - 1 9 6 ~ Specific surface areas were obtained by applying the BET equation to the data of N2 adsorption. Moreover, the static ion-exchange capacity for the studied carbons using 0.05 M HC1 and 0.05 M NaOH was measured. The details of these experiments are given elsewhere [7]. All of these d a t a are s u m m a r i z e d in Table 2. The reaction of activated carbon M with the oxidants changed not only the chemical n a t u r e of its surface but also its texture characteristics. Carbon oxidation by concentrated nitric acid caused the decrease in the surface area to the greatest extent. Some of meso- and macropores were probably destroyed because of the loss of pore walls. On the other hand, oxidation by nitric acid involves creation of different surface nitrate complexes. The m a i n adsorption band of FTIR spectra for the modified Merck carbon is depicted in Figure 1.

780 Table 2 Surface area and ion-exchange capacity of the activated carbon samples Sample MC MCO MCN M MO MN

Surface area m2/g

Anion-exchange capacity mmole/g

Cation-exchange capacity mmole/g

1200 1090 810 1140 910 820

0.29 0.26 0.03 0.29 0.19 0.04

0.37 0.60 2.44 0.34 0.63 2.32

44.5

44.0

3

43.5 43.0 42.5

2

42.0, 41.5 41.0 40.5 40.0 39.5 39.0 38.0 1787.0

I

I

I

I

I

I

I

I

I

1750

1700

1650

1600

1550

1500

1450

1400

1357.0

cm-1

Figure 1. FTIR spectra of activated carbons, 1 represents sample M.; 2 sample MO; 3 sample MN. Identification of the chemical species is not possible but some direct information about the surface chemical n a t u r e of the carbon may be obtained. Thus, if the spectra of the samples are compared, especially in the region 1780-1350 cm 1, it can be seen that on the sample MN a new band at 1577 and 1538 cm -1 is appears with disappearance of the band at 1643 and 1636 cm -1. A new band for the carbon MN shows the presence of nitro groups and surface nitrate complexes [5,8,9]. The spectra of carbon M and MO are very similar but the band at 1717 cm -1 is more intensive and could be assigned to the carboxylic acid groups. Finally, one can conclude t h a t t r e a t m e n t of the sample M with H202 to obtain the sample MO brought about the fixation of a large a m o u n t of oxygen,

781 mainly in the carboxylic acid groups and treatment of the sample M by concentrated nitric acid brought nitro and nitrate aromatic compounds onto the surface of this carbon. Moreover, as the effect of nitric acid oxidation, humic acid is formed on the carbon surface. Comparing the static cation-exchange capacity of the studied carbons one can conclude that by neutralization of humic acids abnormal exchange capacities for the carbons oxidized by nitric acid are obtained. These observations were confirmed experimentally by the UV study of the soluble product of the water carbon suspension at the basic pH values. Identification of the individual humic acids is a very difficult task and will be undertaken later. On the other hand, the oxidation t r e a t m e n t either by H202 or nitric acid causes reverse dependence of the ion-exchange capacity of the modified carbons. An increase in the cation-exchange capacity and a decrease in the anion-exchange capacity were observed. In the case of the oxidation by concentrated nitric acid the anion-exchange capacity is practically equal to zero. Oxidation of activated carbons with different oxidizers leads to formation of carbon preparations characterized by differentiated and significantly exposed ion-sorption properties. The surface of oxidized carbon contains different functional groups (e.g. carboxyl, phenol, carbonyl of quinone type groups, carboxyl anhydride and cyclic peroxides) of acidic nature which can dissociate at different pH values [3]. Conjugation of the functional groups on the carbon surface causes the increase of dissociation degree of some groups to such a degree that ion exchange is possible also in a strongly acidic medium (pH = 1). Organic carboxyl ion-exchangers exhibit their ion-exchange properties at higher (2 units) pH values. Ion-exchange properties of carbon change depending on pH of the solution, dimension of ion and specific properties of the element. An increase of pH value of the solution causes the increase of total ion-exchange capacity of the carbon and makes selectivity of the exchanger worse [10]. Specific properties of the ion may influence significantly the course of ionexchange process on the oxidized carbons. For example the sorption of trivalent chromium ions is significantly lower than that of divalent ions of nickel, copper and even calcium. From the investigations of ion-exchange properties of oxidized carbons it is possible to arrange different ions according to their increasing ability for ion-exchange on the activated carbon, as follows [11]: NH< Na< Rb<< Cs< Mg< Ca, Cd< Ba, Mn< Zn< Co< Ni< Pb<
782 Owing to the ability of oxidized carbons for specific sorption as well as to large differences in the ion-exchange constants of individual ions, the oxidized carbons can be utilized for separation of the ions of similar properties e.g. lithium, sodium, potassium and cesium ions [13]. High numerical values of the ionexchange constants of many ions on the oxidized carbon allow for quantitative isolation of individual metal ions in the presence of large excess of matrix. This makes the use of oxidized carbons for analytical concentration of trace amounts of elements possible. By appropriate changes of chemical structure of the carbon surface the desirable changes in ion-sorption properties of carbon preparations can be attained [4]. The good results may be obtained by building of some elements (oxygen, nitrogen, sulfur, phosphorus and selenium) to the surface layer of carbon and deposition of different organic and inorganic substances on the carbon surface (impregnation). Building of nitrogen atoms into the surface of activated carbons influences, in an interesting way, the ion-sorption properties of this material. Careful partial oxidation of nitrided carbon with hydrogen peroxide or nitric acid diminishes slightly its ion exchange capacity and improves simultaneously its cation exchange properties. A significant improvement of selectivity of the sorption of the ions (especially divalent ions) on this type of carbon is then observed [14]. By deposition of sulfur on the carbon surface the carbon preparations of differentiated ion-exchange properties are obtained. The presence of sulfur on the surface of activated carbons causes the formation of adsorption centers for anions, and in the case of carbon black adsorption centers for ions are formed. Garcia et al. [15] sulfurized the active carbon by t r e a t m e n t in H2S and SO2 atmospheres under different conditions. They employed in the study the commercial activated carbon Merck (1.5 mm, AC) and the samples prepared by heat t r e a t m e n t of AC from 30 to 900~ in N2 (C-N2-900) or H2S (C-H2S-900), or by t r e a t m e n t of the material first at 30~ successively in SO2 and H2S and then at 200~ N2 (C-SO2-H2S-200). Modified carbons were characterized in term of chemical composition based on the elemental and proximate analysis data, and texturally by N2 adsorption a t - 1 9 6 ~ mercury porosimetry and helium density measurements. Some data of their chemical and textural characteristics are shown in Table 3. Table 3 Sulfur content, specific surface area and helium density of the adsorbents (After Garcia et al. [15]. Reproduced with permission from Pergamon Press) Sample AC C-N2-900 C-H2S-900 C-SO2-H2S-200

S wt%

m2g -1

pile g cm-

0.1 0.2 9.9 10.2

921 922 785 764

1.90 2.02 2.04 1.87

aEstimated from N2 adsorption a t - 1 9 6 ~

SBET a

by the BET method, with Am. =16.2/k 2.

783 As shown in Table 3, the sulfur content is very low for AC and C-N2-900 because this carbon material usually posseses a small sulfur content. As for C-H2S-900 and C-SO2-H2S-200, the sulfur content is around 10 wt%. Despite different methods of preparation of the samples the sulfur contents were similar. However, the macroporosity was better developed in the two samples obtained by heat t r e a t m e n t of activated carbons. The FTIR spectra study of sulfurized carbons indicated that the sulfur-carbon complexes formed at high temperature. When the carbon was heated in H2S some atomic groups containing H2S or SHwere observed. Finally, they suggested, concerning sulfurized carbons, that in the t r e a t m e n t of carbons with SO2 and H2S, sulfoxide and thiophenol groups were formed respectively. These modifications of carbons could be very useful for adsorption of soft acid ions. Phosphorization of activated carbons with phosphoric acid or phosphorus trichloride vapours at high temperature leads to formation of complex oxidephosphorus structures, which causes that the modified carbon sample exhibits strong acidic and ion-sorption properties. An impregnation of the activated carbon with different substances creates numerous possibilities of modification of its properties. Selective sorption properties for the sorption from the liquid phase are such a type of carbon, on which some species which can complex the selected ions or bond selected species owing to chemisorption, are adsorbed preliminarily. Carbon preparations with the deposited complexing agents (e.g. dimethylglioxime, o-nitroso-h-naphtol, alkali metal xantogenate or dithizone) selectively adsorb the ions contained in the solutions owing to formation of the complexes between these ions and the substance deposited on the carbon surface. The carbon impregnated with the lead sulphide adsorbed effectively uranium compounds in the solution [16]. The adsorption of uranium from aqueous solution was investigated by Abbasi et al. [17] using conventional commercially available carbon. It was found that treatment with hot nitric acid oxidized the surface carbon and significantly increased the adsorption capacity for uranium in nearneutral and slightly acidic nitrate solutions. In comparison to the conventional organic ion-exchange resins the activated carbons modified in order to improve their ion-sorption properties are characterized by better operational parameters, except the maximum ion exchange capacity. Carbon materials are of relatively low costs and are characterized by high chemical resistance and stable ion exchange capacity in the case of multiple sorption and desorption. Moreover, such materials have a lot of advantages: high radiation resistance and possibility of combustion useful to obtain the maximum concentration of metal ions in the sample.

0

STUDY OF THE TOXIC METAL A D S O R P T I O N ON THE ACTIVATED CARBONS

The adsorption yield of the ion metals on the activated carbons depends on many parameters. From the analytical point of view, the knowledge about the

784 conditions, in which close to 100% of the ion adsorption yield can be reached, is very important, especially when the static method of preconcentration is employed. These p a r a m e t e r s can be divided into the three groups: p a r a m e t e r s describing the nature of carbon surface, specific ion properties and the solution components. Depending on the system studied, the role of each group of p a r a m e t e r s is different. The adsorption centers on the carbon surface are in conjugation with pH values of the bulk solution. The equilibrium between the surface functional groups and the solution depends on the pH value of the solution. Adsorption of ions is affected by the surface charge of the particles of activated carbon and some specific interactions between particular kinds of the surface groups and the ions. This effectplays a significant role in the case of the adsorption of multivalent cations. The strength of interactions between particular surface groups and the ions depends on both the kind and activity of the adsorption center and the ion properties. The most active adsorption centers are occupied firstly by the ions, then the adsorption on the other centers takes place. Exhaustion of the most reactive adsorption centers causes the change of adsorption mechanism. Therefore the studies of the initial part of the ion adsorption isotherms on the activated carbon are very useful. The initial part of the Pb (II) and Cd (II) isotherms on the modified medical carbons at n a t u r a l pH values are shown in Figure 3. In the range of low Pb(II) ion equilibrium

0.1

E

o. ~

5

I'0

15

20

c {mmole/dm'} Figure 2. Adsorption isotherms of Cd(ll) from the aqueous solution on the activated carbons at 25~ Closed circles - the ash free carbon M; open circles - the carbon oxidized with H202; half open circles - the carbon degassed in argon at 1000~ (After Dobrowolski et al. [3]. Reproduced with permission from Pergamon Press).

785

concentration the highest adsorption for the carbon MCN was observed. Practically the most reactive centers for this carbon are occupied by Pb(II) ions up to 0.2 mg/dm 3 of its equilibrium concentration, next the mechanism of adsorption is changed. In the case of Cd(II) adsorption, the linear increase of Cd(II) adsorption in the function of its equilibrium concentration for the carbon MCN was obtained. On the other hand, if the global adsorption isotherm is taken for comparison of the Cd(II) adsorption capacity, the method of carbon modification is essential. Comparing the initial part of the carbon isotherms presented in this paper with those presented previously [3], with respect to the ability of Cd(II) adsorption, one can conclude that the adsorption of Cd(II) was maximum for the carbon modified with concentrated nitric acid and minimum in the case of the untreated sample. The highest adsorption in the case of the carbon oxidized by nitric acid was attributed to both ion-exchange with the surface acidic groups and forming the surface complex chelate type with humic acids present on the carbon surface. Relatively low adsorption in the initial part of the Cd(II) isotherm for the carbon degassed at I I00~ in argon atmosphere (see Figure 2) can be explained by the fact that the conditional constant for Cd(OH)z precipitation must be achieved at first, then precipitation of cadmium hydroxide on the carbon surface OCCURS.

10 {33

t~

E

4

(a)

,_._,

t-

C

.~ 3

.o

t~

Q.

t_

O "ID

6

4

2 4-

4-

%

r

2

13.

0 0

5

10

15

Cd 2+ equilibrium concentration [mg/dm 3 ]

0

1

2

3

4

5

Pb2+ equilibrium concentration [mg/dm 3 ]

Figure 3. Initial part of: (a) Cd(II) and (b) Pb(II) isotherms on the activated carbons. The ionic strength I = 0.1 created by NaC1, (I,) carbon MCN, (o) carbon MCO, (e) carbon MC.

Initial part of Cr(VI) adsorption isotherms on the modified medical carbons are shown in Figure 4.

786

1500 Cr} Cr)

E C

1000

o Q. t_ O U) "ID

500

o 0

400

800

1200

Cr(VI) equilibrium concentration [mg/dm 3 ]

Figure 4. Initial part of Cr(VI) isotherm on the activated carbons. Ionic strength I = 0.1 created by NaC1, (.) carbon MC, (m) carbon MCO, (A) carbon MCN.

These isotherms were measured using the graphite furnace atomic absorption spectrometer (GFAAS) for determination of Cr(VI) at the equilibrium concentration. The study was carried out for the best measuring range of chromium by the GFAAS method. For a very low initial Cr(VI) concentration, the complete Cr(VI) adsorption was observed. The adsorption of Cr(VI) ions is strongly dependent on pH. Figure 5 shows the effect of pH on the adsorption of Cr(VI) from the 15 mmole/dm 3 solution onto the carbon M. 1600 0

E 1200 .m

~-

8oo 400

0 0

0

2

4

6

8

10

12

14

pH Figure 5. Effect of pH on Cr(VI) adsorption on the modified Merck carbon, (e) carbon M, (o) carbon MO, (A) carbon MN.

787 The optimum pH for the adsorption of Cr(VI) onto the carbon M was about 3.5. Depending on the carbon type used for adsorption, the optimum pH for Cr(VI) adsorption changed from 2.5 to 6 [4,15,18-20]. This phenomenon may be explained by changing the surface charge for a different carbon and existing of different chromium species with respect to pH and chromium concentrations [3,15]. The oxidized carbons adsorbed Cr(VI) more effectively compared to the chemisorbed oxygen free carbon [19]. The mechanism of Cr(VI) sorption on the activated carbon is explained as reduction of Cr(VI) to Cr(III) and adsorption of Cr(III) aqua complex onto the carbon surface [4]. Purii and Satija [19] studying the interaction of charcoal with the acidified potassium dichromate solution proposed two stages reaction: K2Cr207 + H2SO4

> K2304 + Cr2Oa + H20 + 3(0)

and Cr203 + H2SO4

Cr2(304)3 + 3H20

The adsorption at lower concentrations of potassium dichromate was found to occur in the first reaction. However, as the concentration of potassium dichromate in the solution increased, the second reaction involving the formation of Cr20~ was observed. Huang and Bowers [20] found that the adsorption and reduction occurred simultaneously and were responsible for removal of Cr(VI) from the bulk solution. These authors also noticed that when the carbon was oxidized with nitric acid and the solution contained chlorinated water with Cr(VI) ions, the removal of Cr(VI) decreased considerably. These observation encouraged us to study the influence of oxidants on Cr(VI) adsorption. In Figure 6 the influence of potassium nitride and potassium perchloride on the Cr(VI) adsorption is shown. The exponential decrease of Cr(VI) adsorption with the concentration increase of the oxidants is observed. In the case of KNO3, for a concentration higher t h a n 0.05 g/dm 3 no adsorption of Cr(VI) was obtained. The slight differences in the interaction of both oxidants on Cr(VI) adsorption into the activated carbon may be caused by the differences in the ion radius of these oxidants. Both oxidants affect redox potential of Cr§ § system. The reduction process of Cr(VI) to Cr(III) is carried out on the carbon surface which possesses negative charge at the acid values of pH. On the other hand, the reduction process is possible when HCrO~ 1 anion is in the acidic bulk solution [18]. Taking into account the analytical application of ion adsorption onto the activated carbon, the influence of matrix on the adsorption capacity is very important. For this reason, the influence of the ion strength on Cr(VI) adsorption was studied. The ion strength was created by proper addition of NaC1 into the Cr(VI) solution. In the first step of the study the influence of NaC1 concentration on Cr(VI) adsorption was determined. In Figure 6 the influence of NaC1 concentration on the Cr(VI) residue in the bulk solution is shown.

788

,--.,

E o

120

(3b

.c. 80-

o~ 300

J

(3r)

=L tO

r~0

40-

.~ >

> O

200

o

._ 00

I

0

0 . ~ -z

0,0

100

0,5

1,0

1,5

2,0

NaCI concentration [mol/dm3 ] Figure 6a. Influence of NaC1 concentration on the Cr(VI) residue in the bulk solution. The system studied: 50 mg of the activated carbon in 50 cm 3 containing 300 ng/cm 3 Cr(VI), (.) denotes carbon MC, (m) carbon MCO, (A) carbon MCN.

0 0,00

0,02

0,04

0,06

Oxidant concentration [g/dm3 ]

Figure 6b. Influence ofKNO3 (o) and KC104 (m) concentration on the Cr(VI) adsorption on carbon MCO. The system studied: 50 mg of the activated carbon in 50 cm 3 containing 300 ng/cm 3 Cr(VI).

The presence of NaC1 does not influence practically Cr(VI) adsorption only for the carbon MCN, in the case of the carbons MC and MCO the Cr(VI) adsorption drastically decreases with the m i n i m u m for 0.025 mol/dm 3 NaC1 concentration, then the adsorption increases asymptotically with the increase of NaC1 concentration. These different behaviors of the studied carbons can be explained by the different adsorption of Cr(III) after the reduction of Cr(VI). For the carbon MCN Cr(III) is adsorbed owing to the complex compounds with humic acids, whereas for the carbons MC and MCO forming the chloride complexes with Cr(III) and the adsorption in this form on the carbon surface seem to take place. Bautista-Toledo et al. [9] studied the presence of NaC1 in the chromium solution on the adsorption capacity of the activated carbon. The studies were carried out on the Merck carbon (sample M) which was modified with HNOa (sample MO). In Figure 7 the adsorption isotherms of Cr(VI) on the activated carbon obtained by Bautista-Toledo [9] are depicted. All adsorption isotherms were obtained out without adding any buffer to control the pH. The final pH of the equilibrium solution was around 6.5 for the sample M and around 3.5 for the sample MO. For the carbon M u p t a k e of Cr(VI) is lower when NaC1 increases. This behavior was explained by the fact t h a t for this carbon the adsorption was nonspecific or through the diffuse layer (competition between Cr024- and CI-) whereas in the case of the carbon MO the increase of Cr(VI) adsorption with increase of NaC1 concentration was observed. This result was a t t r i b u t e d to the fact t h a t an increase in of NaC1 concentration of the solution causes the activity coefficient of the present species to decrease

789

according to the Debye-Hfickel equation which, according to the Nernst equation, causes the reduction potential of the Cr(VI)/Cr(III) system to increase, i.e., the reduction process is more favored which, in turn, causes the a m o u n t of chromium adsorbed as Cr(III) to increase. The oxidation of the carbon by concentrated HNOa, in our case, seems to be more effective and formation of humic acids on the carbon surface strongly influences the mechanism of Cr(VI) adsorption.

9T )

6

<

3

00

1~0

20

30

C r (mg/1)

Figure 7. Adsorption isotherms of Cr(VI) on the activated carbon samples: M.[ NaC1] = 5 . 1 0 .2 M. (0); M.[NaC1] = 5 . 1 0 .4 M. (9 MO.[NaC1] = 5 . 1 0 .2 M. (..); Mo. [NaC1] - 5.10 .4 M. ([]), (After Bautista-Toledo [9]. Reproduced with permission from Pergamon Press).

Apart from the study of Cr(VI) adsorption on the activated carbons the trace mercury enrichment on carbons is very often an analytical aim of consideration [21-28]. Koshima and Onishi [21] studied the quantitative collection of nanogram amounts of mercury and methylmercury in the artificial sea water containing mineral acids as preserving reagents. Activated carbon powder (Merck, GR grade, No. 2186 was used after heat t r e a t m e n t for mercury removal at 350~ for

790 2 hr. Mineral acids contain mercury and cannot be used directly as preserving reagents. Removal of mercury from mineral acids was performed by their t r e a t m e n t with activated carbon. Figure 8 shows t h a t sulfuric acids (<9.8 M) and hydrochloric acid (<5.8 M) can be easily purified by t r e a t i n g t h e m with the activated carbon.

I00

80

o9 > 9 r o9

60

40

20

I

I

I

I

I

I

0

2

4

6

8

10

Acid concentration, M Figure 8. Collection of Hg(II) from different acid solutions. Activated carbon, 100 mg; contact time, 1 hr. 9 H2SO4, 50 cm 3, 20 ng Hg(II); e, HC1, 25 cm 3 40 ng of Hg(II); A, HNO3, 100 cm 3 2 ng of Hg(II), (After Koshima, [22]. Reproduced with permission from Pergamon Press).

It was stated t h a t iodide shows a beneficial effect on the collection of mercury, possibly by formation of HgI24- complex [27] and suggested to add some iodide to the sample as a preservative. Recovery of mercury was practically independent of the a m o u n t of activated carbon used, in the range of 20-250 mg/200cm ~ artificial sea-water. The effect of the mineral acid concentration on the collection of mercury is shown in Figure 9. Good recovery was obtained from the artificial sea-water w h e t h e r unacidified or made up with 0.7 M in sulfuric acid, 1 M hydrochloric acid, or 0.3 M nitric acid. The adsorption of Hg(II) from aqueous solutions using activated carbon was

791

~

100

0

~

0

9

90

v

80 9 o,,)

70

60

50

0

I

I

I

I

I

0.2

0.4

0.6

0.8

1.0

Acid concentration, M Figure 9. Effect of acid concentration on the collection of Hg(II) from the atificial sea-water. Solution, 200 cm3; Hg(II), 3 ng; activated carbon, 100 mg; contact time, 1 hr. o, the artificial sea-water plus H2SO4; o, the artificial sea-water plus HC1; A, the artificial sea-water plus HNO3, (After Koshima, [22]. Reproduced with permission from Pergamon Press).

found to be pH dependence [23,28]. The adsorption of Hg(II) increased as the pH of the solution decreased, the acidic pH range being most suitable. The efficiency of Hg(II) adsorption depends on the nature of the carbon and the nature of the activated t r e a t m e n t that the carbon received [22]. Moreover, carbons also enhanced Hg(II) adsorption by addition of certain chelating agents [24] or by sulfurization of the carbon surface [29]. The large capacity of the sulfurized carbon was attributed to the interaction between the mercury and the sulfur present on the carbon surface. Activated carbons modified by impregnation with the metallic sulfides were also found to be very efficient for selective adsorption

792 from the aqueous solution [4]. Thiem et al.[30] stated that the addition of chelating agents such as tannic acid or EDTA improved the adsorption of mercury and the addition of as little as 0.02 mg/dm 3 chelating agent increased the adsorption of mercury from 10% to 30% depending on the pH of the solution. The calcium ion exhibited an effect on adsorption of mercury similar to that of chelating agents. Yoshida et al. [26] studied the effect of addition of EDTA on Hg(II) and Cd(II) adsorption on carbons prepared from coal by steam activated carbon. By addition to the aqueous solution of toxic metals in the amounts equimolar with metal ions, complex toxic metals were formed and readily adsorbed on carbon under neutral or weak alkaline conditions. The adsorption rate of Cd-EDTA complex on carbons was fast and the equilibrium was reached in an hour. The adsorbed Cd-EDTA complex on the carbon surface was eluted quantitatively with 0.01 M NaOH or 0.1 M HC1. Adams [24] studied the mechanism of Hg(CN)2 and HgC12 adsorption onto the activated carbon. Mercury was found to adsorb onto the activated carbon in the form of Hg(CN)2 from the cyanide solution. The Hg(CN)~- species displays a negligible affinity for carbon. The adsorptive behavior of the mercuric cyanide species is consistent with the distribution profile of the species in solution. The adsorption of mercury from the chloride solution occurs by reduction to the Hg(I) species, with subsequent precipitation of insoluble Hg2C12 within the pores of the carbon. The affinity of Hg(II) and Cd(II) complex for a surface of activated carbon was studied by Yoshida et al. [27]. The strong affinity of Hg(II) and Cd(II) complex with inorganic anion CN-, SCN- and I- for the carbon surface was observed. To improve efficiency of arsenic, selenium and mercury adsorption onto the carbons from the aqueous solution Per~iniemi et al. [31] studied zirconium loaded activated charcoal. They stated that the loading of activated charcoal with zirconium significantly improved the adsorption power of charcoal towards arsenic and selenium. With selenium the state of oxidation was of little significance for adsorption onto ZrC*, but in the case of arsenic, As(V) adsorbed better than As(III). Mercury was adsorbed about equally well by both the unloaded and zirconium-loaded activated charcoal and to some extent Hg(II) adsorbed better than Hg(I) onto both adsorbents. When the pH of the solution of the sample solution was below 10, it had no effect on the adsorption of arsenic, selenium or mercury onto ZrC*. They suggested that adsorption of arsenic and selenium onto ZrC* is attributable to the reaction of these metals with zirconium, rather than with activated charcoal as in the case of mercury. In the end, they concluded that ZrC* is a promising adsorbent for the preconcentration of As(V), Se(IV), and Hg(II) from aqueous solutions for the determination methods in which solid sampling can be used. For this reason, the effects of various anions and cations on the adsorption of the analytes onto the ZrC* were investigated [31]. Figure 10 shows the recoveries of As(V) and Hg(II) as a function of interfering anion concentration. Nitride ion and other halides than fluoride had no notable effect on the adsorption of the analytes onto ZrC*. Mercury, unlike arsenic and selenium did

793

a) ~'100

b)

CI-,Br-, NO~, SO~ "

"

9

-

Br-, NO~, SOl

o~~10C

Se(IV) ~50

F-

o:

.

0

.

1

10 C (rag/l)

100

. PO,'.

0

1000

d) CI-, NO~ -

-

-

.

~~o~-

|

1

0

1o

~oo

~ooo

C (mg/1)

c) ~1o

~I

-

~ f Se(VI) ~~ t

_

I

_

?~oo

el_

_

Br-, N O ~ ,

i-,

Hg(I

o~

~

F

50 PO~

o,

0

.

I

9 ~" ~

I0

C (rag/l)

i00

;P~

I000

o

0

'i

I

~o

ioo

,h,.

ido~

C (mg~)

Figure 10. Effect of anions on the recoveries of (a) As(V) (b) Se(IV), (c) Se(VI), (d) Hg(II) (2.5 mg analyte/1, V-0.1, n=4), (after Perfiniemi [31 ]).

not adsorb quantitatively onto ZrC*, but the adsorption was increased linearly with Hg(II) concentration and this made the mercury determinations possible [32]. Chloride ion increased the adsorption of Hg(II) slightly, possibly due to formation of Hg2C12 on carbon. They concluded that the adsorption of Hg(II) can be improved by adding chloride ions into the samples, and into the calibration standards otherwise over 100% recovery is obtained. Sulphate ion had not great effect on the adsorption of As(V), Se(IV) or Hg(II) onto ZrC*. But it decreased the adsorption of Se(VI). Fluoride ion, and especially phosphate ion, decreased the adsorption of As, Se and Hg possibly due to the competition of these anions and the analytes for the adsorption sites on ZrC*. Studying the effect of various metal ions on the adsorption of the analytes they stated that almost of them at the concentrations level 100 mg/dcm 3 do not interfere with the adsorption of As and Se onto ZrC*. By contrast, Cr(VI), Mn(VII), Mo(VI), Fe(III), Au(II), Sn(II), Sb(III), Te(IV), Bi(III) and Cu(II) were collected in various degree on ZrC* and decreased the adsorption of As and Se when present in high concentration. The authors concluded that these metals ions probably either compete with analytes for the adsorption sides on ZrC* or they may cause the precipitation and loss of analytes on the container walls. In the case of mercury, Per~iniemi assembled the metals into tree classes e.g: the alkaline earth elements, Cr(III), Mn(II), Fe(II) and

794 Te(IV) which increased the adsorption, the second group includes the elements: Mn(II), Co(II), Ni(II), Cu(II), Ag(I) and Pb(II) and have no significant effect on Hg(II) adsorption and the third group which adsorbed themselves at least weakly, decreased the adsorption of Hg(II) probably either by competition for the adsorption sites or by the precipitation and loss of mercury on the container walls. Since arsenic, selenium and mercury often occur in the environmental waters as both inorganic and organic compounds the authors studied the adsorption power of ZrC* towards the organometallic compounds of the analytes at different pH values of the sample solution. It was stated that the organomercury compounds and (C6Hs)AsO(OH)2 behaved in a similar m a n n e r to inorganic compounds of the studied elements. By contrast, the behaviour of (CH3)AsNaO2 and (C6H6)Se2 differed from that of inorganic arsenic and selenium compounds. Figure 11 shows the recoveries of As and Se measured by the energy dispersive x-ray fluorescence (EDXRF) as the function of pH of the solution.

100

8

Ph~Se~.

50

0 0

3

6

9

I 12

9

pH

Figure 11. Recoveries of As and Se as the function of pH of the sample solution (2.0 mg/1, V = 0.1, n=4) (after Per~iniemi [31]).

The most suitable pH range for (CH3)AsNa02 is from 3 to 7; the recovery of the arsenic is then more or less constant. In the case of (C6H6)Se2, the adsorption was constant when pH < 6, but the maximum value was achieved when the pH was between 8 and 9. Optimizing arsenic, selenium and mercury determinations in the aqueous solutions they stated that ZrC* adsorbent is highly selective and the proposed analytical procedure is also suitable for the determination of its organometallic compounds.

795 There are only a few publications [32-37] dealing with the adsorption of arsenic species onto activated charcoal. According to Gupta et al. [34] As(II) does not adsorb onto activated charcoal at all, but As(V) adsorbs to varying degree when the pH of the sample solution is between 2 and 8. Adsorption mechanism of As(III) and As(V) in the activated carbon was studied by Kamegawa et al. [37]. The species of As(V) adsorbed on t h e activated carbon were considered to be H2AsO ~- and HAsO~-, since As(V) was adsorbed in the pH range from 4 to 9 on carbon prepared from coal, coconut shell and wood by the steam activated method. The adsorption ability of activated carbons for As(V) was in the order: activated carbon from coal > from coconut shell > from wood. The authors stated that the interference on As(V) adsorption from coexisting anions was in the following order: C10~ > SO24-> NO3 > CI-. Moreover, they concluded that As(III) was not adsorbed on all the activated carbons tested, since As(III) dissolved in the undissociated form, H3AsO3, below pH 8. However, As(III) was readily oxidized to As(V) with the dissolved oxygen on the activated carbon so that As(V) formed was adsorbed on activated carbon. As a result of this, the activated carbon having a higher activity to oxidize As(III) showed a higher adsorption ability for As. Based on the presented results one can conclude that in the case of Cd(II) and Pb(II) ions poor adsorption on the surface modified carbons is exhibited. For this reason, the organic complexing agents are applied to enhance the uptake of these ions [38-44]. Jevtitch and Bhattacharyya [41] studied the effect of surface and species charge on the separation of metal chelates by the activated carbon. A relationship between the adsorption capacity, surface charge of the activated carbon, and the average species charge for various cadmium ligand systems was presented. Chang and Ku [44] studied in detail the characteristics of metalligand complex formation on the activated carbon and the effect on adsorption behavior. The Freundlich and Langmiur isotherms were used to correlate the parameters of models with the solution pH values. It was found that the highest adsorption capacity was obtained for the Cd-citrate system. Cd-EDTA and CdNTA have similar adsorption capacities on the activated carbon. The experimental results indicated that Cd-EDTA and Cd-NTA systems show no precipitation of cadmium even at pH close to 11. Adsorption equilibrium of these chelates was explained by the species charge and the carbon surface charge characteristics. The species distribution is highly dependent on the solution pH as well as on the type and amount of the chelating agent present. They stated that the effectiveness of adsorption decreases for a metal-ligand anion at higher pH values when the surface of activated carbon becomes negatively charged. With the lower pH values, the opposite effect was observed. This adsorption characteristic was observed in the adsorption of Cd+2-EDTA. However, for the Cd+2-citrate and Cd+2-NTA complexes, a lower adsorption was observed with the decreasing pH values in the pH levels from 6 to 9. This was because at lower pH values and the species concentration, the increasing degree of partial

796

complexation of cadmium with citrate and NTA with decreasing pH values results in the concentration of Cd § cations increase. In fact, the exact mechanism between the chelated toxic metals and the surface groups is very complex and difficult for explanation but the experimental results clearly indicate that the charge barrier between the carbon surface and solute species plays a predominant role during adsorption process. Classification of the systems with respect to the presence of various organic and inorganic compounds, competing metals, varying ionic strength and background electrolyte was introduced by Reed et al. [43]. The authors presented a typical pH-adsorption edge for each system, including one representing a ligand-free system (see Figure 12).

f

I

""-'--

"

(D > 0

E w!

l-0

2

o U_

pH Figure 12. Typical pH-adsorption edges for a ligand free system and for separate groups, (After Reed et al. [43]. Reproduced with permission from Marcel Dekker).

They distinguished four groups and typical pH - adsorption edges for the ligand-free system which are described as follows. In the first system, when the general shape of pH-adsorption edge remains unchanged as in the ligand free system except that the adsorption edges move to a higher pH region compared to the ligand free system. A plateau, defined as the leveling off of the pH-adsorption edge at fraction removed <1.0, may also be observed. This behavior causes that when the adsorbate/adsorbent concentration ratio is increased, decreasing the number of surface sites per unit of metal ion is observed. The secondary

797 compound competes with the solid surface for the metal ion, decreasing the amount metal ions available for interaction with the solid surface. The secondary compound may be adsorbed by the solid; decreasing the number of surface sites available for metal removal. Increasing the ionic strength alters the metal solution chemistry and the electric double layer. The second system includes the metal-ligand complexes adsorbed by a ligandsolid equivalent interaction. The shape of the pH-adsorption edge is radically different from that observed in the ligand-free system. Adsorption increases at lower pH values and decreases at higher pH-values compared with the ligandfree case. Typically, at low pH values the surface of carbons is positively charged while at high pH values the surface is negatively charged [3]. Metal-organic complexes are often negatively charged over a wide pH range (e.g. EDTA-metal complex). In a ligand free system, the electrostatic force is usually repulsive at low pH values, decreasing the adsorption energy. For a negatively charged metalligand complex the electrostatic force is attractive, increasing the adsorption energy. At higher pH values the negatively charged carbon surface and metalligand complex produce a repulsive electrostatic force, which can decrease adsorption compared to the ligand-free system. The third system refers to the unique complexation between the metal, ligand, and carbon surface and causes an increase of ion adsorption at all pH values. The general shape of the pH - adsorption edge is similar to the ligand free system. This phenomenon is rarely observed. This behavior occurs at the time when at low pH values the carbon surface carries a positive charge and the metal ligandcomplex is negatively charged while at higher pH values the signs of the charges are reversed. Thus, the electrostatic force is attractive over a wide pH range. The fourth system can be distinguished when adsorption is suppressed at all pH values. The presence of a strong metal complexer (e.g. EDTA) or a large amount of a weak complexing agent may cause a decrease in adsorption.

0

TRACE A N A L Y S I S BY ATOMIC S P E C T R O M E T R Y WITH THE E N R I C H M E N T OF THE TOXIC METALS ON A C T I V A T E D C A R B O N S

Purity and fine powdering of activated carbons used in trace analysis followed by further analysis using spectral methods are the main criteria for the choice of these carbons. Commercial carbons contain mineral impurities, which make their direct analytical application impossible. Mineral impurities of activated carbons are removed most often by the Korver method [45] consisting in the extraction of the impurities with hydrochloric and hydrofluoric acids. This method is, however, time-consuming and is used in this connection mainly for purification of greater amounts of the adsorbent. A relatively simple method of removal of mineral impurities from activated carbons was described by Dobrowolski [46]. At first commercial carbons were leached by concentrated hydrochloric acid using a Soxhlet apparatus made of quartz and by hydrofluoric acid in the Teflon vessel as described earlier. Then the

798 chromatographic elution with the concentrated acids and continuous control of the removal process by AAS method was applied. Toxic metals were leached to the ppb level (see Table 4) except for chromium at carbon M which is still present at the ppm. Table 4 Results of toxic metals determination in the commercial and fine purified activated carbon Content (ppm) Carbon

Cd

Pb

Cr

Co

Hg

As

Se

M

0.950

0.515

5.85

1.01

0.0035

< LOD

0.035

MN

0.017

0.073

1.05

0.005

0.0013

< LOD

0.006

MC

0.080

0.120

0.45

0.04

0.0020

< LOD

0.020

MCN

0.003

0.010

0.007

0.002

0.0012

< LOD

0.005

LOD denote the limit of detection.

An exact method of purification of activated carbons from mineral impurities was described by Karyakin and Gribovskaya [47]. Using the multiple extraction with hydrochloric acid they have obtained the carbon preparation (extraction process was repeated 7-8 times at the temperature of 80-90~ containing ca 10-5% of impurities, mainly silicon. However, this method of carbon purification is time-consuming and is not applied widely in the analytical practice. On the other hand, high purity carbons can be ashed after the adsorption process and trace amounts of metals may be determined by the flameless AAS method. Such a procedure was used by Koshima and Onishi [23] to determine silver, gold and platinum after their previous adsorption on carbon adsorbents. The hitherto existing methods of trace analysis utilizing activated carbons consist mainly in complexation of the metals and in adsorption of formed complexes on the activated carbon by a static or dynamic method. Complex species adsorb easily on the surface of the activated carbon which allows for significant concentration of metal trace amounts. Different substances e.g. xylenol orange [48], 2,2-dipyridylketono-2-pyridylhydrazone (DPPH) [49], 0,0-diethylether of ditiophosphoric acid [50,51], potassium xantogenate [52], potassium ethyl xanthate [53], sodium diethyldithiocarbamate [ 5 4 ] and 8-hydroxyquinoline [55] were used as chelating agents. Recently, the simple method of nickel determination in urine by the flame atomic absorption spectrometry (FAAS) was described by Aydemir and Gficer [56]. They preconcentrated nickel from urine with ammonium pyrrolidine dithiocarbamate on the activated carbon. They applied a static rhethod of

799 preconcentration at the constant pH value, adjusted to 6.3. The carbon suspension was stirred mechanically for one hour and filtered through a filter paper. After the addition of concentrated nitric acid to the residue and evaporation to the dryness and centrifugation they determined nickel by injecting microvolumes to the flame of AAS spectrometer. An interesting method for atomic absorption spectrometric determination of lead copper, cadmium and nickel in the drinking water samples after preconcentration by sorbing the l-(2-pyrodyzlazo) 2-naphthol (PAN) complex of these metals on an activated carbon column was established by Soylak et al. [57]. The metal/PAN complexes were quantitatively retained on the activated carbon in the pH range 6-8. Metals retained on the activated carbon column were completely eluted with 2 M HC1 in acetone. This method is characterized by good recovery (> 95%) and a small relative standard deviation (< 7%, with relative error < 3%) of analysis. Dithizone has been used as a chelating agent for the combined flow atomic absorption spectrometric system to develop an efficient on-line preconcentrationsolvent elution procedure for determination of trace amounts of cadmium [58]. The authors described a new method whereby the trace amounts of cadmium are preconcentrated as a metal chelate using the chelatingdye dithezone on an activated carbon mini-column included in the flow injection (FI) system. The chelate is eluted with a small volume of a water-immiscible solvent and the analyte does not disperse on transfer to the flame atomic absorption spectrometry (FAAS) instrument, which increases the preconcentration factors. The system was used successfully for determination of cadmium in biological reference material. The enrichment method was developed by Gficer and Yaman [59] for the determination of cadmium and lead in vegetable matter by FAAS after preconcentration with 8-hydoxyquinoline or cupferron on the activated carbon. The enrichment factor of up to 100 was achieved by using both complexling agents. The optimum lowest pH values were found as 4.8 and 4.4 with cupferron and 5.3 and 5.8 with oxine, for cadmium and lead. The described procedure allows to detect 0.06 and 0.48 ng of Cd and Pb respectively, in biological material. The relative standard deviations were found to be 2% for cadmium and 3% for lead. Owing to the complexity of sorption processes, concentration of the ions characterized by different chemical nature is possible. Desorption of metal ions from activated carbons before their further determination by the AAS method (assuming an appropriate choice of a chelating agent) usually permits to attain a high degree of recovery (above 95%). In the analytical applications a special attention was paid to adsorption of mercury on the activated carbon. Koshima and Onishi [23] concentrated the nanogram amounts of mercury and methylmercury contained in the sea water on the activated carbon, then desorbed the concentrated compounds and determined by the flameless AAS after previous deposition on the gold support. Using this method they determined less than 4 ng of mercury contained in 200-300 cm ~ of the sample of artificial sea water using only 100 mg of the activated carbon.

800 During the tests of different carbon adsorbents [23] they concluded that the adsorption process is most effective in the case of the activated carbon. Matsueda [60] determined mercury (II) contained in the river water, after the previous concentration of this metal on the activated carbon. Recently activated carbon fibers were also used for adsorption of mercury (II) from aqueous solutions [61]. A majority of the publications relating to the utilization of activated carbons in the trace analysis by the AAS method describe modification of sorption conditions connected mainly with the changes of chelating agent, eluent and physicochemical parameters of the system e.g. pH. Less attention is paid to physicochemical nature of the surface of activated carbons. Storozhuk and Ivanova [62] paid the attention to significant contribution of surface phenomena occurring on the surface of activated carbons which can be used successfully for separation and concentration of trace amounts of elements. They stated that during the elution of adsorbed silver (I) with different eluents more than 90% of silver adsorbed on the activated carbon was reduced to metallic silver. Thus, the elution of adsorbed species with the eluents of different chemical nature allowed for analytical separation of adsorbed ions depending on their chemical nature. Some interesting surface properties are also found in basic carbons for which mainly physical adsorption of metal ions and surface concentration of these ions are observed [3]. Such properties ensure the increase of selective concentration and this may be used in the process of separation of metal ions on the activated carbon in a wide range of the matrix concentration. The examples of practical application of sorption on the activated carbon and of concentration of trace amounts of the elements from different matrices before their determination by the AAS method are given monographs Bachmann [63] and Mizuike [64]. Flameless atomization [e.g. 46] and flame atomization with injection of an analyte using a micropipette [e.g. 50,54] are especially useful for determination of the metals because they allow to determine a few components contained in the concentrate of a small volume. This allows to obtain very good detection limits (nanogram amounts of element per 1 g of matrix). Apart from the above method consisting in concentration of trace amounts of metals on activated carbons followed by determination using the AAS method one can find the papers suggesting the possibilities of utilization of activated carbons for the concentration of trace amounts of metals and determination of these metals by the atomic emission spectrometry (AES) method in which arc is used as an excitation source. After the preliminary concentration the carbon was washed and dried and then appropriate samples were collected from this material. These samples were usually mixed with appropriate spectral buffer e.g. with NaC1 and then placed in the craters of the carbon auxiliary electrodes. Such a procedure allows for direct excitation of emission spectra of the samples in the direct and alternating current arc. The optimal weight of carbon sample is 25-50 mg. Such an amount of the sample may be transferred completely into the crater of the electrode. The procedure consisting in a direct transfer of carbon samples to the electrodes and excitation of the emission spectra in the arc allows

801 for multielemental determination of the metals deposited on the carbon in the concentration range of 10~-105% with relatively good accuracy (sufficient for practical purposes) and at relatively small consumption of the reagents. This method has been used for determination of impurities in high purity zinc (II) and nickel (II) nitrates [65] and for determination of trace amounts of metals in sea water [66]. The main reasons for toxic ion metals preconcentration on activated carbons are to obtain a better limit of detection and separation of the determined element from the matrix. Toxic metals should be determined at ppb or even ppt levels in different environmental samples. Using modern spectroscopy methods and by preconcentration of metal ions on activated carbon these analytical goals can be reached. Okutani et al. [67] presented a rapid and simple preconcentration method by selective adsorption using carbon as an absorbent and acetylacetone as a complexing agent for beryllium determination by the graphite furnace atomic absorption spectrometry (GFAAS). Beryllium and its compounds are very toxic, especially for the lung, skin, and eyes, and the higher concentration can cause death (the toxic concentration of Be for a person is 0.1mg/m3). Consequently, a reliable method for the detection of beryllium in environmental specimens is required. The authors proposed the method based on adsorption of the beryllium-acetylacetonate complex on the activated carbon from the rainwater and sea-water samples, separation and dispersion in pure water. The resulting suspension was introduced directly into the graphite furnace atomizer. The detection limit was 0.6 ng/dm~, and the relative standard deviation (RSD) at 0.25 ~g/dm 3 was 3.0-4.0% ( n = 4 ) . A similar method was applied for determination of the total selenium content in sediments and natural water [68]. A trace level of Se was collected on the activated carbon (AC) as the Se(IV)-3phenyl-5-mercapto-l,3,4-thiadiazole-2(3H)-thione (Bismuthiol II) complex. The activated carbon was separated from the aqueous phase through a membrane filter (8 pm. pore size). The carbon phase on the filter was then dispersed in water containing the Pd modifier using an ultrasonic device for 30 s. The activated carbon suspension was directly introduced into the graphite tube atomizer and the Se concentration was determined by the GFAAS method. Naganuma and Okutani [69] adopted this method to determine trace amounts of bismuth in the sediments. They preconcentrated bismuth as the bismuthpotassium o-ethyldithiocarbamate (potassium xanthogenate, KetX) complex on the activated carbon and then introduced directly as a carbon suspension into the carbon tube atomizer of AAS spectrometer. The detection limit for the proposed method was 0.005 pg/100 cm 3, and RSD was 2.0% (n=8). The method was applied for determination of bismuth in the river-bottom and submarine sediments. Application of the activated carbon for preconcentration of Cd, Pb, Cr and Co of micro amounts from the reference material, drinking and sea water samples was comprehensively presented by the author [70]. The carbon MN, additionally impregnated with the 1% water so,ution 1 of ammonium pyrrolidinedithiocarbamate (APDC) was used as an adsorbent. Preconcentration of toxic ion

802 metals was carried out by the static method. 50 mg of the carbon was added to a 500 cm 3 sample, pH of the solution was natural and then the solution was mixed mechanically for two hours. The activated carbon was separated from the aqueous phase using an 8 ~m. membrane filter (25 mm in diameter). The carbon was removed from the filter and dispersed in 5 cm 3 of water with the addition of 0.005% Trition X-100 solution, using the ultrasonic device just before injection into the furnace. In the case of Cd and Pb analysis, the palladium nitride was added to the standard and carbon slurry (500 ~g/5 cm 3) due to the fact that carbon impurities (Si and P) influence the atomization mechanism of these elements. The results of toxic metal ions determination in different water samples are shown in Table 5.

Table 5 Determination of toxic ion metals sampling GFAAS technique

Element

Sample

SLRS-3

riverine water

Drinking

Cd

NASS-2 Seawater SLRS-3

NASS-2 Seawater SLRS-3

NASS-2 Seawater SLRS-3

NASS-2 Seawater

12 + 1

sea water

27 + 2

from the Baltic

32 +1.7

riverine water

65 + 3.2

water

83 + 3.3

Reference value (ppt) 13 -I-2a

29 + 4 68+7 b -

slurry

Recovery % 92 93 96 -

39+6

92

180

+7.2

-

-

riverine water

304

+ 24

300 + 40 a

sea water

water

sea water from the Baltic riverine water water

sea water from the Baltic

36 +3.5

using the carbon

from the Baltic

Drinking

Co

Slurry sampling* (ppt)

18 +0.9

Drinking

Cr

samples

water

Drinking

Pb.

in the water

56 + 4.5 173

+ 8.5

192

+ 9.6

25

+ 1.5

II +0.8 5 + 0.6 82 +5.1

175 + I0

I01 99

27 + 3 e

93

-

-

4 + 1 -

125 -

the mean of seven individual determination a - the direct determination by isotope dilution inductively coupled plasma mass spectrometry (ICP MS) b - the acid digestion isotope dilution inductively coupled plasma mass spectrometry (ADIDICP MS) e - the concentration by evaporation, GFAAS *

-

803 The results of determination obtained by the application of the preconcentration procedure and the slurry sampling technique are comparable to the certified values. The standard deviations for these determinations are lower compared to those obtained by different methods used during certification. CONCLUSIONS

Summing up the above considerations we can conclude that: I. Activated carbons are used successfully in the process of concentration and separation of toxic metals before their determination by the atomic spectroscopy methods. Description of adsorption of the electrolytes on the activated carbon is a very difficult task because of complexity of many processes occurring simultaneously such as ion exchange, nonspecific sorption, redox surface reactions, surface precipitation as well as formation of surface complexes of the chelate type. 2. In comparison to the synthetic ion exchangers carbon adsorbents allow for determination of concentrated impurities without the necessity of their isolation from the sorbent. These impurities can be determined directly after their introduction to the electrothermal atomizer e.g. in the form of the suspension (slurry sampling technique) of the activated carbon in diluted acids (carbon powder does not cause any interferences in the case of AAS method). Moreover the inductively coupled plasma atomic emission spectrometry (ICP-AES) [71] or the direct current plasma atomic emission spectrometry (DCP-AES) slurry sampling technique could be used successfully. The samples may be introduced also to the craters of auxiliary electrodes and then determined by the AES method in the arc. 3. Application of activated carbons is recommended especially for the concentration of mercury or methylmercury before their determination. 4. Concentration methods utilizing the carbon adsorbents create the possibilities of the analysis of small volumes of solutions. 5. Concentration and separation of the substances on the activated carbon adsorbents are often of lower costs and simpler than the multielemental extraction method. REFERENCES

I. H.P. Boehm, Advan. Catalysis, 16, (1966) 179. 2. H. Jankowska, A. Swiatkowski and J. Choma, Activated Carbon, WNT, Warsaw, 1985. 3. R. Dobrowolski, M. Jaroniec and M. Kosmulski, Carbon, 24 (1986) 15. 4. R.C. Bansal, J.B. Donnet and F. Stoeckli, Active Carbon, Marcel Dekker Inc., New York, Basel, 1988. 5. A. Capelle and F. De Vooys, Activated carbon...a fascinating material, Norit N.V., The Netherlands, Amersfoot, 1983.

804 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

H. von Kienle and E. Bader, Aktivkohle und ihre industrielle Anwendung, Ferdinand Enke Verlag, Stuttgart, 1980. A. Gierak, Adsorption Science & Technology, 14 (1996) 47. J. Zawadzki, Carbon, 19 (1991) 19. I. Bautista-Toledo, J. Rivera-Utrilla, M. A. Ferro-Garcia and C. Morebo-Castilla, Carbon, 32 (1994) 93. A. N. Mironov and V.P. Tauskanov, Adsorpcija i Adsorbenty, 2 (1974) 32. D.G. Hager, Chem. Eng. Progr., 72 (1976) 57. I.A. Tarkovskaya, Okislennyj Ugol, Kijev, Naukovaya Dumka, 1981. B.K. Strasko, I.A. Kuzin and A.M. Semusin, Zh. Prikl. Chim., 39 (1966) 2044. V.T. Nikolaev and V.V. Strelko, Gemosorpcija na aktivirovannych uglach, Naukovaya Dumka, Kijev, 1979. A. Macias-Garcia, C. Valenzuela-Calahorro, V. Gomez-Serrano and A. Espinosa-Mansilla, Carbon, 31 (1993) 1249. J.P. Novikov, V.B. Gliva and V.M. Komarevskii, Radiochimija, 19 (1977) 653. W.A. Abbasi and M. Streat, Separation Science and Technology, 29(9) (1994) 1217. H. Yoshida, K. Kamegawa and S. Arita, Nippon Kagaku Kaishi, 3 (1977) 387 B. R. Puri and B. R. Satija, J. Indian Chem. Soc., 45 (1968) 298. R. Leyva Ramos, A. Juarez and R. M. Guerrero Coronado, Wat. Sci. Tech., 30 (1994) 191. C.P. Huang and A.R. Bowers, I0 th Int. Conf. Water Pollution Res., 1977. H. Koshima and H. Onishi, Talanta, 27 (1980) 795. H. Koshima and H. Onishi, Bunseki Kagaku, 31 (1982) 421. M. D. Adams, Hydrometallurgy, 26 (1991) 201. H. Yoshida, K. Kamegawa and S. Arita, Bunseki Kagaku, I0 (1976) 1596. H. Yoshida, K. Kamegawa and S. Arita, The Report of the Government Industrial Research Institute, Kyushu, No. 22, 1979. H. Yoshida, K. Kamegawa and S. Arita, Nippon Kagaku Kaishi, 8 (1977) 1231. Y. Nagasaki, Japan Kokai, 76(122) (1974) 477. R. K. Sinha and P. L. Walker, Carbon, I0 (1972) 754. L. Thiem, D. Badorek and J. T. O'Connor, J. Am. Water Works Ass., 68 (I 976) 447. S. Perfiniemi, M. Ahlgren, Anal. Chem. Acta, 302 (1995) 89. S. Perfiniemi, S. Hannonen, H. Mustalahti and M. Ahlgren, Fresenius Z. Anal. Chem., 349 (1994) 510. A.T. Ellis, D. E. Leyden, W. Wegscheider, B. B. Jablonski and W. B. Bodnar, Anal. Chem. Acta, 142 (1982) 73. S. K. Gupta and K. Y. Chen, J. Water Polution Control Fed., 50 (1978) 493. E. Kobayashi, M. Sugai, I. Satoshi, Nippon Kagaku Kaishi, 23 (1984) 655. L. V. Rajkovic, Sep. Sci. Technol., 27 (1992) 1423. K. Kamegawa, H. Yoshida and S. Arita, Bunseki Kagaku, I0 (1979) 1365.

805 38. Y. Lu, K. S. Subramanian and L. Chuni, J. Environ. Sci. Health, A28(I) (1993) 113. 39. C. P. Huang and F. B. Ostrovic, J. Env. Eng. Div. Proc., ASCE, 104 (EE5) (1978) 863. 40. M. M. Jevtitch and D. Bhattacharyya, Chem. Eng. Commun., 23 (1983) 191. 41. D. Bhattacharyya and C. Y. Cheng, Environ. Progress, 6(2) (1987) II0. 42. O. M. Corapcoiglu and C. P. Huang, Water Research, 21(9) (1987) 1031. 43. B. E. Reed and S. K. Nonavinakere, Sep. Sci. Tech., 27(14) (1992) 1985. 44. C. Chang and Y. Ku, J. Hazardous Materials, 38 (1994) 439. 45. J.A. Korver, Chem. Weekblad, 46 (1950) 301. 46. R. Dobrowolski, Ph.D. Thesis, M. Curie-Sklodowska University, Lublin,1988. 47. A.V. Karyakin, U. Gribovskaya, Metody opticeskoj spektroskopii i luminescencji w analizie prirodnych i stocnych vod, Chimija, Moskva, 1987. 48. E. Jackwerth, Fresenius Z. Anal. Chem., 271 (1974) 120. 49. D.J. Hutchinson and A.A. Schilt, Anal. Chim. Acta, 154 (1983) 159. 50. H. Berndt and J. Messerschmidt, Fresenius Z. Anal. Chem., 308 (1981) 104. 51. H. Berndt, J. Messerschmidt and E. Reiter, Fresenius Z. Anal. Chem., 310 (1982) 230. 52. H. Berndt, E. Jackwerth and M. Kimura, Anal. Chim. Acta, 93 (1977) 45. 53. P. Rama Devi, G. Rama Krishna Naibu, Analyst, 115 (1990) 1469. 54. E. Jackwerth, H. Berndt, Anal. Chim. Acta, 74 (1975) 299. 55. B.M. Vanderberght, R.E. van Grieken, Anal. Chem., 49 311 (1977) 56. T. Aydemir, S. Gticer, Anal. Lett., 29(3) (1996) 351. 57. M. Soylak, I. Narin, M. Dogan, Anal. Lett., 30(I 5) (1997) 28 I0. 58. Y. P. de Pena, M. Gallego and M. Valcarcel, J. Anal. At. Spectrom., 9 (I 994) 69 I. 59. S Gficer and M. Yaman, J. Anal. Spectrom., 7 (1992) 179. 60. T. Matsueda, Bunseki Kagaku, 29 (1980) 110. 61. K. Kaneko, Carbon, 26 (1988) 903. 62. R.K. Storozhuk and C.S. Ivanova, Adsorpcija i Adsorbenty, 7 (1979) 19. 63. K. Bachmann, CRC Critical Reviews in Anal. Chem., 12(1) (1981)1. 64. A. Mizuike, Enrichment Techniques for Inorganic Trace Analysis, Springer Verlag, M~nchen, 1983. 65. K.I. Lazebnik, M.I Obruckii, A.N. Tomasevskaya and I.A. Tarkovskaya, Adsorpcija i Adsorbenty, 3 (1974) 57. 66. Z.P. Suranova, O.J. Grabcuk and A.N. Tomasevskaya, Adsorpcija i Adsorbenty, 2 (1974) 55. 67. T. Okutani, Y. Tsuruta and A. Sakuragawa, Anal. Chem., 65 (1993) 1273. 68. T. Kubota, K. Suzuki and T. Okutani, Talanta, 42 (1995) 955. 69. A. Naganuma and T. Okutani, Anal. Sci., 6 (1990) 77. 70. R. Dobrowolski, The 23rd Annual Conference of the FACSS, Kansas City, USA, 1996. 71. K.D. Ohls, J. Flock and H. Loepp, Winter Conference on Plasma Spectrochemistry, San Diego, USA, 1992.

Adsorption and its Applications in Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski(Editor) 9 1998Elsevier Science B.V. All rights reserved.

807

Air pollution control by adsorption

W.M.T.M. Reimerink, D. v.d. Kleut NORIT Nederland B.V., P.O. Box 105, 3800 AC Amersfoort, The Netherlands 1. I N T R O D U C T I O N The pressure on industry to decrease the emission of pollutants to the air is increasing. The importance to industry is to keep the costs as low as possible. A broad spectrum of techniques is available and is developed to control air pollution. The choice of a technique is determined by the type of pollution and the process conditions. In relation to price/performance, physical adsorption is one of the most important techniques to control air pollution. Both organic and inorganic molecules can be removed from a gas stream by physical adsorption. However, the adsorption affinity increases as the molecules become larger. As a consequence the adsorption capacity of an adsorbent is higher for large molecules t h a n for small molecules. For this reason physical adsorption is extremely suited for adsorption of organic compounds from gas, air, water and liquid streams. 2. A D S O R B E N T S In actual practice only the following adsorbents are applied: Activated carbon Carbon molecular sieves Polymers Silica Alumina - Zeolites Activated carbon can adsorb a broad range of pollutants with varying dimensions by its broad pore distribution of micro- and small meso pores. Activated carbon can adsorb a large a m o u n t of pollutants due to its large pore volume. Due to its hydrophobic character adsorption takes place at high relative humidity. The usability of activated carbon for air pollution control is limited by the risk of ignition at high temperature. Much research has been done to develop polymer adsorbents.In the market polymers as polyad [1,2] are applied for adsorption of high boiling compounds. At -

808 high relative h u m i d i t y these adsorbents have a tendency to swell. The polymers cannot be used at high t e m p e r a t u r e s due to deformation. A great disadvantage of polymer adsorbents is the low adsorption capacity on a volume basis. Alumina and silica [1,2] are meso porous and are not suited for adsorption of small organic molecules by physical adsorption. Unmodified silica and a l u m i n a are hydrophilic, so a high relative humidity disturbs the adsorption of organic pollutants to a large extent. Silica is suitable as adsorbent of water by chemisorption. Zeolites [1,2] and carbon molecular sieves [3-7] have a narrow pore distribution. The pore distribution determines which adsorbate adsorbs well and which adsorbate adsorbs to a less extent. These adsorbents can be applied in purifications of well defined gas s t r e a m s such as are present in gas separation applications. Compared with activated carbon the usability of these adsorbents is limited. Zeolites can be made hydrophobic by increasing the A1 content. These hydrophobic zeolites are suited to purify gas s t r e a m s at high relative h u m i d i t y and high t e m p e r a t u r e . Of the above mentioned adsorbents activated carbon is the most convenient for air pollution control for a broad range of compounds and for a large variation in process conditions. Compared to carbon molecular sieves and hydrophobic zeolites, activated carbon is a relatively cheap adsorbent. 3. A C T I V A T E D C A R B O N Activated carbons are micro porous carbonaceous materials. The activated carbons available in the m a r k e t differ in pore distribution, in form and in chemical composition. To decrease the emission an optimal carbon and system should be chosen dependent of the kind of molecules to be removed and the process conditions. The differences between activated carbons types are a consequence of the choice of activation process, the activation conditions and to some extend the choice of raw material. Activated carbons are produced from raw materials such as peat, wood, lignite, anthracite, fruit pits and shells. The raw materials are converted in activated carbon by steam or chemical activation With steam activation [8] the raw material is carbonised and/or oxidized depending on the carbonisation degree. Activation takes place above 900~ with steam. Process variations as residence time in the kiln, the activation temperature, the type of kiln and other conditions, allow carbonised materials to develops small micro pores which are enlarged up to large micro pores or small meso pores. Activated carbons suitable for gas and air purification are micro porous. When the gas stream contains a low concentration of pollutants, lesser activated carbons with a large a m o u n t of small micro pores exhibiting a high adsorption capacity at low relative pressure are applied. These carbons are produced by steam activation in a rotary kiln after a relatively short residence time. In the case which the gas stream contains a high concentration of pollutants higher activated carbons are applied. These carbons show a higher adsorption capacity at high relative pressures t h a n lower activated carbons. At high concentrations the service time of a filter is relatively short. As a consequence the activated carbon has to be regenerated insitu,

809 otherwise the carbon consumption will be too high. Higher activated carbons have a greater proportion of larger micro pores and small meso pores. These larger pores easily desorb their adsorbate. High activated carbons are also produced by steam activation in a rotary kiln. Only the residence time is longer than for the production of carbons with small micro pores. With chemical activation[9] an activating chemical, normally phosphoric acid, is mixed with a young carbonaceous vegetable material, carbonised at about 500~ followed by recovery of the activation chemical by water washing. The activated carbons produced on this way have less micro pores and more meso pore compared to steam activated carbons and are suited for adsorption of larger molecules such as are present in decolorization steps in the chemical, pharmaceutical and food industries. However special types of chemical activated carbons are suited for a small part of gas phase applications with insitu regeneration. By the choice of the raw material and by modifications in the activation process, more small meso pores and large micro pores can be produced. In Figure 1 the benzene adsorption isotherm of steam activated carbons suitable for gas phase purification with insitu regeneration (SORBONORIT 3) and for gas purification on throw away basis (NORIT R 2030) are given. In this figure a chemical activated carbon for gas phase purification with insitu regeneration has also been involved (NORIT GF 45).The adsorption capacity is given as weight of adsorbate per unit volume of activated carbon since in most applications, adsorbents are compared on performance in an existing filter (fixed volume). IBenzene

o

adsorption

isotherms

I

........i ....iiitil .........tttitii'!t ........tti titil .........i .....iitifl .........i'.i~ ""ti "'_ i i i iiiiii i i i iiiiii i i i iiiiii i i i ii~ill ,.T ~

r

......... i..... iTTIiYYi ......... YYYIYiIIY ........

i !i.~"i:'" 9: ,.,~i'Yiiiil .: ... i iiliii

ii ii iiiiiii ii iiiii

ii ii iili!i i iiiii

......... i ..... !...~...!..~.i .~.~. :: i i iiiiiii

ii ii iiilili i iiiii:

"

i

i

-

~ ii.~!~ii

.~),~D...~I":

i i iilii~

+." 9 i ii~L~"

i

i iiliiii

i i iiiliii

i iiiiiii

i i iiliiii

i i i iiiiii : .~It] iiiiii i i i iiiiii i i ii iiili : : : :::::: .y.,V": :::::: : : : :::::: : : : :::::: ......... i ..... !"" .~"i'~. "i~ .~ .......... ~'" .~"i" .~'i'i'i ~. ........ .~'"" .~'"i'" .~'!" .~ ~. i.~ ......... i ..... ........ i...i..6.66661

0.00001

0.0001

0.001 rel. p r e s s u r e

0.01

i iiiiii

i i iiiiii

i i i iiili : : : ::::: ......... i ..... i...6..i.6.i.i6

0.1

1

(%)

1~ SORBONORIT 3 4,. NORIT R 2 0 3 0

-T. NORIT GF 45

Figure 1. Adsorption isotherm of different carbon types for gas phase applications.

810 With the help of the Kelvin equation the relative pressure can be converted into pores dimensions. Thus Figure 1 shows the indirect relationship between adsorption capacity and pore distributions of different carbon types. The following carbon physical forms are utilised dependent upon the application: - extrudated carbons - granular carbon - powdered carbons - fibres. Extrudates and granular carbons are mostly applied in fixed bed systems. The particle size is chosen dependent on the allowed pressure drop. Especially in gas phase recovery systems, activated carbon is exposed to large pressure differences for long times and is transported on a regular basis to sieve the carbon. In that case the hardness and attrition are important qualities Extrudates especially are extremely hard. Recently in contrast to fixed bed systems, powder injection systems have been applied in gas/air purification. In recovery systems with a large flow and a relatively low concentration loosely woven fibre systems are used. Activated carbons ignite at high temperature in air. The ignition depends of the activation process and the purity of the activated carbon and varies from 200~ for a chemical activated carbon to 500~ for steam activated carbon produced from peat without additives as potassium. In most gas/air applications activated carbons are used at low temperatures (up to 200~ In this case danger of ignition does not exist. Without modification, activated carbon can show chemical interaction with adsorbates or can be catalytically active. This chemical interaction and the catalytic activity can be desired or not. For example, in the recovery of ketones the catalytic activity is undesired. The chemical interaction and the catalytic activity are connected to the presence of functional groups and ash components. Steam activated carbons contain as a consequence of exposing to air after activation, a limited number of varying functional groups, which give the carbon basic qualities. Chemically activated carbons possess by virtue of the production process a much larger number of varying functional groups, which give the carbon acid qualities. The ash content is dictated by the used raw material and can be diminished by washing. For very small molecules, the physical adsorption capacity can be low. Thus for a gas stream with a mixture of very small and larger molecules the low adsorption capacity for the small molecules can dictate the performance of the filter to a large extent. In that case the activated carbon can be modified to increase the removal efficiency of the small molecules by chemisorption or catalytic conversion. For this reason activated carbons are modified.

811 4. GAS P H A S E A D S O R P T I O N ISOTHERMS 4.1. I n t r o d u c t i o n Gas phase adsorption isotherms describe the relationship between the relative pressure (or the concentration) of a component in the gas phase and the maximum loading capacity, that is the loading capacity at equilibrium. A great number of equations have been developed to describe the equilibrium adsorption. For activated carbon as adsorbent none of these equations describe the measured isotherm for all the concentration ranges. For gas phase adsorption the following equations are used in practice [10] - equations based on the theory of Dubinin - the Langmuir equation - the Freundlich equation - the Henry equation - the BET equation. For gas and air purification isotherms based on the equation of Dubinin yield good results and give the most possibilities to predict the adsorption capacity for different compounds and temperatures [11]. 4.2. The a d s o r p t i o n i s o t h e r m of D u b i n i n and R a d u s h k e v i c h The adsorption isotherms of Dubinin and Radushkevich are based on the potential theory of Polanyi and assume filling of the pore volume by means of liquefaction of the gas by physical adsorption. The equation has been modified by a large number of investigators. Investigations carried out by Van Soelen [11] show that these modified equations hardly show an improvement for predicting the adsorption capacity. The equation of Dubinin-Radushkevich, has the following form

in (Av)- ln(W. d ) - B.

0//nl /

9log

(1)

Av

Equilibrium adsorption in terms of weight per volume unit of activated carbon (g/cm 3) W and B Carbon constants d Density of the adsorbate(g/cm 3) T Temperature (K) b Affinity constant of the adsorbate p/po Relative pressure n Exponent, varies from 1 to 3 For n=2 the equation has been suited to micro porous activated carbons. J. Reussien [12] shows that the equation with n=l can be applied for the meso porous part of the pores structure. The carbon constants W and B can be calculated from the intercept and the slope by plotting ln(Av) against [T/b'log(p/p0)] n. To use the

812 above equations, the temperature must be below the critical temperature. For temperatures above the critical temperature an adjusted equation must be applied [12]. Activated carbon is used for the adsorption of a broad range of adsorbates. The advantage of the use of the theory of Dubinin and Radushkevich is the possibility to predicts the adsorption capacity of all kinds of adsorbates on a carbon type using the carbon constants W and B calculated from the adsorption isotherm of a standard adsorbate. The adsorption isotherm of other adsorbates can then be calculated by substituting the liquid density of the adsorbate at the adsorption temperature and the affinity constant. The affinity constant can be calculated from the parachor or from the surface tension as given in standard tables. For gas streams with more than 1 component the adsorption capacities of the combined components are calculated by combining the isotherms of the pure components [13].

5. T H E F I X E D B E D A D S O R P T I O N 5.1.

PROCESS

Introduction

In a fixed bed system a polluted gas stream is passed through a bed of activated carbon. After the start of the adsorption process the activated carbon at the inlet side is loaded. Only after a certain time does the inlet side of the bed reach equilibrium because adsorption in pores does not takes place directly and is subject to transport limitations. Within the bed a mass transfer zone develops (MTZ). After a certain time the MTZ boundary reaches the outlet side of the bed and the emission concentration increases up to the allowed value, when the adsorption process is stopped. For adsorbent - adsorbate systems with a convex adsorption isotherm the length of the MTZ is constant and independent of the bed height [14]. In Figure 2 the course of the MTZ through various bed heights is given. The adsorption capacity of a filter is determined by the equilibrium adsorption and the length of the MTZ. A large equilibrium adsorption capacity and a small MTZ means a long service time of the filter and a low carbon consumption. In the most gas phase applications the MTZ is relatively small, certainly at low relative humidity. Thus the adsorption capacity of a filter is largely determined by the equilibrium adsorption. NORIT has developed an empirical model to calculate service time and carbon consumption on the basis of a standard carbon analysis and process conditions. The carbon analysis used are: the adsorption isotherm of a standard adsorbate the particle size. The process conditions used are: temperature concentration -

flow

-

adsorbate qualities.

813

Mass Transfer zone in a carbon bed 1.2

oo ~

0.8

o~

0.6

start loading

8 N 0

end loading 0.4

0.2

0 0

0.2

0.4

0.6

0.8

1

1.2

bed height Figure 2. The MTZ as a function of the bed height.

5.2. Adsorption kinetics Adsorption of gas molecules does not takes place instantaneously. The t r a n s p o r t is limited by: - axial dispersion - external transport - internal transport I n t e r n a l t r a n s p o r t characteristics are affected by: - pore diffusion - K n u d s e n diffusion a limited adsorption velocity surface diffusion. D e p e n d i n g of the process conditions, the type of a d s o r b a t e a n d the adsorbent qualities one or more steps are dominant. For exact d e t e r m i n a t i o n of the MTZ a set of differential equations h a v e to be solved. To e s t i m a t e the MTZ an empirical equation can be used which is derived from m e a s u r e d d a t a w i t h i n the m a t r i x of process conditions, which exist in practical situations. For activated carbon systems u n d e r relative dry conditions this equation is:

814 MTZ : e S T . (F)"0"054 9(Ci)0"133 9(Dp)1"549 9log Ci-Co Co with MTZ Mass transfer zone (cm) constant CST F Flow (cma/min) inlet concentration (g/cm 3) Ci outlet concentration (g/cm 3) Co particle diameter (cm) Dp

(2)

In solvent recovery applications when the activated carbon is wet after regeneration and cannot be dried during adsorption, the MTZ is 3 times larger.

5.3. The service time and the c a r b o n c o n s u m p t i o n Figure 2 shows that the MTZ curve is symmetrical. Thus half of the MTZ part of the bed can be considered to be in equilibrium with the inlet concentration and half of the MTZ part of the bed can be considered to be completely empty. For a fixed bed with an cross section area S (cm2)and a bed height L(cm) the dynamic loading Aa (g) A d - A v ( L - - - M T Z / $ 2"

(3)

At a flow F and an inlet concentration Ci the service time tb(min) of a bed with volume L'S is equal to tb-

Ad Ci.F

(4)

The carbon consumption CS (cm3/min) is CS- Ci'F'L'S Ad

(5)

5.4. F i x e d bed in insitu r e g e n e r a t i o n s y s t e m s 5.4.1. I n t r o d u c t i o n Insitu regeneration can be applied to gas streams with a high component concentration. In this case, the carbon consumption is too high for throw away basis operation. In insitu-regeneration systems the carbon is loaded in the same way as in fixed bed adsorption system up to break through. After break through the adsorbate is desorbed. Desorption may be followed by a (partly) drying step. During insitu regeneration the adsorbate is desorbed by pressure swing or temperature swing action. With pressure swing, desorption takes place at lower pressure than is present during adsorption. Pressure swing is applied for gas

815 separation. In t e m p e r a t u r e swing, desorption takes place at a higher t e m p e r a t u r e t h a n is present during adsorption.Temperature swing is mostly applied for solvent recovery. S t e a m or inert gas such as nitrogen is used as carrier gas. The benefit of the use of s t e a m is t h a t the installation, including the activated carbon bed,is w a r m e d up very quickly. Inert gas regeneration is applied for components which desorb at relatively high t e m p e r a t u r e and for components which decompose by oxidation with the activated carbon acting as a catalyst. In the last case steam activation is only possible by the use of activated carbon with a low catalytic activity such as the SORBONORIT K4. The motive to recover solvents can be the value of the recovered solvent. Recovery of solvents can also be a method to fulfill emission requirements. In some cases, recovery of a mixture of solvents can be used to effectively concentrate emissions to allow incineration.

5.4.2. The c a l c u l a t i o n of the s t e a m c o n s u m p t i o n With s t e a m regeneration, the solvent is recovered at high t e m p e r a t u r e with steam as the carrier gas. The rest loading on the carbon Ar as a function of the steam consumption is m e a s u r e d to determine the universal steam curve. For calculation of the adsorption capacity of an adsorbent in solvent recovery the effective loading Aef is an i m p o r t a n t factor. The effective loading is the difference between the dynamic loading Ad and the rest loading on the carbon after regeneration. In Figure 3 a steam curve has been given. Figure 3 shows that the recovery of the same a m o u n t of solvent costs much more in steam when starting from a lower dynamic

I

5

Steam curve

i

...........................................................................................................................................................................................................................

••••••••••••

5

0 0

200

! 400 steam

Figure 3. The steam curve.

I 600 volume

800

1000

816 loading t h a n starting from a higher dynamic loading. In most cases, desorption is stopped before all the adsorbates have been desorbed. By using activated carbons with special pore distribution, desorption can be made more effective.

5.5. F i x e d b e d i n s t a l l a t i o n s In designing an installation the m a x i m u m linear velocity (cm/sec) should be about 100 times the particle diameter, thus preventing fluidisation. The m i n i m u m linear velocity is a few cm/sec preventing axial dispersion. Thus the cross section of an installation is mainly determined by these conditions. The bedheight is mainly determined by the desired service time and the allowed pressure drop. Absorbers with a small bed height are applied for the removal of low concentrations of pollutants (< 1 rag/m3). The bed height is normally about 2 to 5 cm. The contact time is the order of 0.05 to 0.2 sec. Examples of this kind of filter are cylinders and thin rectangle boxes, divided into compartments. In this type of application carbon bounded in sheets such as N O R I T H E N E can be applied. I m p o r t a n t applications for these kind of filters are: - air conditioning - concentration peak smoothing. For higher concentrations (1 mg/m 3 up to 1 g/m 3) larger absorbers with a bed height of 25 up to 50 cm are used. The contact time in this kind of filter is about 0.2 to 2 seconds. Examples of this type of filter are simple steel drums provided with an inlet, an outlet and a base (aeropure filter), rectangular carbon absorbers and vertical as horizontal cylindrical absorbers. Important applications for these kind of filters are: - emission prevention in the chemical and food industries - paint spray installations - sewage air purification. For still higher concentrations (1 up to 50 g/m 3) recovery installations are applied with a bed height of 50 up to 150 cm. The contact time is about 2-4 seconds. Such recovery installations are of m i n i m u m 2 absorbers, one in loading and one in regeneration. Most installations comprise of a large number of absorbers. I m p o r t a n t applications of these kind of filters are: - solvent recovery in printing industry - dry cleaning. 6. I N S T A L L A T I O N S WITH P O W D E R S In a recent development air pollution control systems with injection of powder carbon in the gas stream can be applied. To keep a system in equilibrium, sufficient carbon m u s t be dosed t h a t emission concentration is in equilibrium with the equilibrium adsorption of the carbon. So in gas streams with a pollutant inlet concentration Ci and an emission concentration Co, Z g/cm 3 has to be adsorbed on to the activated carbon.

817 Z = C i - Co At the emission concentration of Co the equilibrium adsorption of the activated carbon Av can be calculated on basis of the theory of Dubinin as shown in 4.2. The carbon consumption CS is CS -

Z.F

(6)

Av The m a x i m u m loading of an activated carbon in such a system may be low compared to the m a x i m u m loading of an activated carbon in a fixed bed system, but by the use of very small particles, the kinetic effect is much faster, an a d v a n t a g e in processes determined by kinetics. Powder injection is applied in gas streams with a high debit and it can be built into a purification train. The system can be relatively cheap compared to fixed bed systems and is flexible concerning carbon dosage. Powder injection systems are used on a large scale at the purification of the flue gas of waste incineration plants [15]. In this way dioxins, dibenzofurans and heavy metals are removed from the flue gas. Typical dosing rates are 50 up to 200 g/m 3. Recently, impregnated powders have also been applied for special applications such as the removal of high concentrations of mercury. Powdered activated carbons have been tested excessively on explosion risks and are considered safe for flue gas conditions.

7. MODIFIED ACTIVATED CARBON The physical adsorption capacity for very small molecules can be low. Thus for a gas stream with a mixture of very small and larger molecules, the low adsorption capacity for the small molecules can dictate the performance of the filter to large extent. In this case the activated carbon m a y be modified to increase the removal efficiency of the small molecules by chemisorption or catalytic conversion. All possible impregnations with metal salts and with organic molecules as well as the modification of the functional groups are mentioned in the literature. To reduce air pollution only a few types of impregnations and modifications are of commercial interest. These i m p r e g n a n t s and the application are given in Table 1.

818 Table 1 Impregnation/modification commercial available activated carbons Component

Impregnant

Application

H2S, methyl mercaptan

- KI - Fe(OH)3 - complexes of transition metals (i.e:Cu,Cr) KOH

sewage air chemical industry

KOH - Na(OH) - KeCO3

sewage air chemical industry air conditioning

- CuO ZnSO4

chemical industry

Hg

-S -KI

purification methane prod. of batteries waste incineration

COS

complexes of transition metals (i.e:Cu,Cr)

chemical industry

HCN, C1CN

complexes of Cu,Cr,Zn and TEDA

gasmasks chemical industry

-

SO2

-

NH3

-

ASH3, PH3

Cu and Cr complexes

radioactive iodide

R

E

F

E

R

E

N

C

E

- TEDA -KI

chemical industry nuclear power plants

S

1. TNO (IMET), Alternatieve Adsorbentia voor het reinigen van koolwaterstoffen bevattende luchtstromen, (1991) (Dutch). 2. Y. Cohen (ed.), Novel adsorbents and their environmental applications, American Institute of Chemical Engineers, (1995). 3. Carbon containing molecular sieves, US Patent No. 3801513 (1971/1974). 4. Carbon containing molecular sieves, US Patent No. 3979330 (1974/1976). 5. Kohlenstoffhaltige Molekularsiebe, German Offenlegungschrift 2305435 (1973/1974). 6. Verfahren zur Gewinning von Stickstoffreichen Gasen aus neber Ne wenigstens 02 enthaltenden Gasen, wie zB. Luft, German Offenlegungsschrift 2441447 (1974/1976).

819 7. Kohlenstoffhaltige Adsorptionsmittel mit einstellbarem unterschiedlichen Porensystem, German Offenlegungsschrift 2624663 (1976/1977). 8. T. Wigmans, Fundamentals and practical implications of activated carbon production by partial gasification of carbonaceous material, NATO ASI series E 105, 559. 9. A. Cameron and J.D. MacDowall, The pore structure of wood based activated carbons, from: Principles and applications of pore structural characterization. Proc. R.I.L.E.M./C.N.R. Symp., Milan, Italy, (1983). 10. R.C. Bansal, J.P. Donnet and F. Stoeckli, Active carbon, New York and Basel

(~988). 1 I. A.C.D. van Soelen, Gas-fysisorptie-isothermen van aktieve kool. RU Utrecht (1991). 12. J.G.J. Reussien, De standaard-benzeenadsorptieisotherm aan actieve kool als een basisgegeven voor de berekening van terugwinnings- en luchtzuiveringsinstallaties, NORIT N.V., (1973). 13. D.M. Ruthven, Principles of adsorption and adsorption processes, New York, (1984). 14. Kel'tsev, Translation chapter 6 and 8 (Dutch), (1984). 15. B.v.d. Akker, D.v.d. Kleut and W.M.T.M. Reimerink, 16th Symposium on Chlorinated dioxins and related compounds, DIOXIN 96, Amsterdam (1996).

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

New developments devices (ELCD)

821

of a c t i v a t e d c a r b o n s for e v a p o r a t i v e loss c o n t r o l

W.M.T.M Reimerink a, J.D. MacDowall, b D. v.d. Kleut a aNORIT Nederland B.V., P.O Box 105, 3800 AC Amersfoort, The Netherlands bNORIT (U.K) Ltd., Clydesmill Place, Cambuslang Industrial Estate. Glasgow G32 8RF, Scotland 1.

INTRODUCTION

Automobiles emit gasoline vapour from their fuel tank and (for old type of automobiles) from their carburettor, when the car is parked and the temperature under the motor hood is still high. Further a car parked at high out-door temperatures can emit a large amount of gasoline vapour by the fuel tank heating up. In addition to emission during parking, gasoline is also emitted during filling the fuel tank. The automobile contributes substantially to man-made hydrocarbon emissions. For example in West Europe the car evaporative emissions amount to about 10%. At refuelling another 3% is emitted. The diagram in Figure 1 shows the contributions to man-made hydrocarbon emissions. In the sixties in California USA legislation was passed to reduce the emission of gasoline from cars. Automobiles had to meet the so-named SHED emission test. Later on in the USA this legislation became more general. In the nineties the legislation in the USA was tightened. Emission of gasoline during refuelling had to decrease. In 1992 legislation in Europe was passed to reduce the emissions from cars by the introduction of an European SHED emission test. Producers of automobiles made arrangements to fulfill to the legal requirements. One of the measurements is the siting of an activated carbon hydrocarbon storage canister under the motor hood. During parking the filter is in contact with the fuel vapour from the tank and in contact with vapour from the carburettor for adsorption of the gasoline. During driving, air is sucked through the filter to desorb the gasoline which is fed to the engine for burning. These filters are described as Evaporative Loss Control Devices (ELCD filters). The extent and stringency of the legislation demanded that new types of canisters and optimized types of activated carbons were developed. In 1990 NORIT started [1-3] the development of an new type of activated carbon for the application in ELCD filters. Important requirements for the nineties and later are a high recovery of gasoline vapour and a low pressure drop. In the future

822

the composition of gasoline will change. ELCD carbons must be able to recover gasoline vapour with more alcoholic substances [4].

I

Contributions to man-made hydrocarbon emission Western E urope (24.8~)

I Car evap. D Car refuelling

(99%)

EI Car exh austs E3 Gasoline distr. I So Ivents

D Ref iner ies c3g.C

[ ] Other

s.8%) .

.

.

.

.

.

.

Figure 1. M a n - made hydrocarbons emissions. 2.

THE S H E D TEST, THE BWC A N D THE GWC TEST

In the SHED [5-8] test an entire vehicle is allowed to emit a certain amount of hydrocarbons during a prescribed test procedure in a test chamber after the car has been driven in. In the test chamber the car is placed on a roller bank. Situations of driving and parking are simulated. The total hydrocarbon emission from all parts of the car are measured, including tail pipe emissions of hydrocarbons. For refuelling special tests are developed. The SHED test is a very expensive and time consuming test. For testing evaporative emissions in the SHED test the activated carbon type has to be approved. For approval the activated carbon has to fulfill requirements on Butane Working Capacity (BWC). A generally useful BWC test is described by ASTM. After approval of a carbon type the BWC test is used as a test for production control. Besides the BWC extended test cycles can be carried out with gasoline to determine the Gasoline Working Capacity (GWC). This is necessary to get approval for a new carbon type. In the ASTM [9] BWC test pure butane is allowed to flow through a tube, filled with activated carbon at a rate of 250 ml per minute for 25 minutes. After

823 completion of the loading step, the activated carbon is purged with air for 40 minutes at 300 ml per minute. Both loading and purging steps are carried out at 25~ The difference between the weights of the carbon tube after loading and after purging divided by the carbon volume in the tube times 100 is the BWC. The BWC is given in g/100 ml. No international standard test method is available for carrying out gasoline tests. Every canister and activated carbon producer has developed his own test which is discussed and adjusted in consultation with the customers (filter and car producers) for correlation with the SHED test. The NORIT GWC test [2,3] is carried out using a carbon bed with diameter 36 mm and bed height of 150 mm. The activated carbon is loaded with a synthetic gasoline vapour, consisting of 50% air, 33% butane and 17% of higher boiling hydrocarbons (RVP(psi)=8.7). Synthetic gasoline vapour is used to obtain optimal reproducibility of the test and to observe small differences in activated carbon performance. The temperature during loading is 30~ and the flow is 0.121 L per minute. The loading is continued up to a breakthrough of 0.5 vol % of the ingoing hydrocarbons. Purging is carried with dry air at the same temperature and at a flow of 2.857 L/min during 16 minutes, which equates to 300 bed volumes. The flow direction during purging is opposite to the direction during the loading step. The GWC is the weight of butane and higher hydrocarbons adsorbed during the subsequent loading step per litre of activated carbon. The test is repeated up to 75 times to measure the decrease in GWC caused by building up a heel of hydrocarbons. In Figure 2 a scheme of the test equipment is shown. Some tests have been carried out with ethanol in the gasoline. In that case the synthetic gasoline vapour composition changed in 15.3% air, 71.1% butane 2.0% ethanol and 11.6% higher hydrocarbons.

Furnace <-

~

~-

Purge air in

Breakthrough

I-- Air in 4/['- Butane in g

Purge air out

Figure 2. Diagram of the GCW equipment.

Hydrocarbons in

~ Vapouriser

824 3.

THE ACTIVATED CARBON PROFILE

To fulfill the requirements of the test the activated carbons m u s t have a high adsorption capacity at a relative pressure of butane of 0.42 %. According the Kelvin equation pores up to 1.9 nm are filled with butane. For desorption pores must be as large as possible. So at a low relative pressure (P/Po <0.001) the adsorption capacity must be low. Pores filled at a P/P0> 0.42 and pores filled at a P/Po< 0.001 does not contribute to the BWC and have to be avoided. So the optimum pores distribution for ELCD is as follows: - m i n i m u m quantity of small micro pores to reduce retention level - m a x i m u m quantity of large micro pores and small meso pores for a high adsorption capacity and a fast desorption - minimum quantity of macro pores. Macro pores decrease the working capacity via the density because these pores are only used for transport and not for storage of gasoline. The desired performance of the grade CNR 115 developed by NORIT is: - BWC : 12 g/100 cc -

GWC:

6 0 g/L.

In Figure 3 the adsorption isotherm of the ideal activated carbon is given. The BWC of this carbon is 12.0 g/100 ml.

Butane adsorption isotherm ideal ELCD carbon

12

: : : : : : : : : : : : : : : : .: . . : . . : . .::::: .

: : :.

:

:

:

:

: : : : :

: :

: :

: : : : : : : : : : . : : :

: : : : : : : : : : : : : : : : .: . . : . . :::::: . . .

:

:

: : : : :

:

. . :. . .: . :.

:

:

: : : : :

.......... - ...... i....-...~..~..i.4.~" ........... : ...... .-:....i...i...-:.i..~.i.~ ........... .- ...... i....~-...-"..i.4.i . . . . . . . . .

: : :.

. . :. . :.

: :

: :

::: ::: :::

:

:

:

:::

,~--~IZl-.~..-.i.-.

10

~

6

<

4

2

0

0.0001

,

. . . . . .

~

0.001

~

-

~

-t.zi-l.zll.zl

.

0.01

.

.

.

.

.

.

.

.

.

0.1

.

.

.

.

.

.

1

Rel. Pressure

Figure 3. The adsorption isotherm of an ideal activated carbon for the ELCD application.

825 By using an activated carbon with a high BWC smaller filters can be used in comparison to activated carbons with a relatively low BWC. Filters to capture refuelling emissions must be much larger even when an optimal carbon is used. For larger filters a low pressure drop becomes an important quality of the carbon. A low pressure drop is reached by increasing the size of the activated carbon particles. However for larger particles, equilibrium adsorption is not reached in a larger part of the carbon bed in comparison to smaller particles because of adsorption kinetics resulting in a lower working capacity. The optimum pressure drop of around 350 mm WC/m carbon layer at an air velocity of 0.25 m per second is obtained if the desired activated carbon has particles with a diameter of about 2 mm or a d50 of 2.7 mm. Activated carbons for ELCD has to be packed in a filter and the carbons remains for years in a canister exposed to vibration by driving. So the activated carbon particles have to be sufficiently hard. The ball-pan hardness must be in the order of 75. Next to working capacity and pressure drop the production cost of a new type of activated carbon is important. 4.

THE P R O D U C T D E V E L O P M E N T OF THE CNR 115

Activated carbons are produced by 2 different kind of processes steam activation chemical activation. With steam activation the raw material is carbonised and/or oxidized depending on the carbonisation degree. Activation takes place above 900~ with steam. Process variations of residence time in the kiln, the activation temperature and other conditions allow activated carbons to develop small micro pores which are enlarged up to large micro pores and small meso pores. For ELCD applications many micro pores have to enlarged to get the right pores distribution. With chemical activation an activating chemical, normally phosphoric acid, is mixed with a young carbonaceous vegetable material, carbonised at temperatures of about 500~ followed by the recovery of the activation chemical by water washing. In this method the phosphoric acid is not volatilised to any degree at the carbonisation temperature, alters the mechanism of charring resulting in higher yields and supports the evolving carbon matrix during carbonisation and coincident volatile removal. The resulting product contains a well developed meso porosity which is attributed to the voids separating the carbonised micro fibrils of the cellulose which make up the majority of the plant tissue. The most raw materials used for chemical activation have a too large macro pores volume which is inadequate for ELCD applications. In addition the developed meso pores are too large The activated carbons produced by chemical activation have less micro pores compared to steam activated carbons. Of these two processes, chemical activation is the preferred route to the correct porosity for the ELCD application. To enlarge small micro pores to get the desired pores distribution via steam activation is a too expensive process.

826 To develop an ELCD carbon by chemical activation the following problems h a d to be solved. - Choice of raw material * For chemical activation the raw material must be of a cellulosic nature. * The raw material must have the right pore structure. For a t t a i n i n g a high adsorption capacity on a volume basis the macro porosity m u s t be low. * The raw material must be available in sufficient quantities. * The price of the raw material must be sufficiently low. - Choice of form Most ideal form is the extrudate. E x t r u d a t e s show a low attrition by a small external surface. Filling of canisters with extrudates is easier t h a n with broken carbons due to better flow properties. - Impregnation With chemical activation of large particles with less macro pores, the problem is to impregnate the raw material with sufficient phosphoric acid, necessary for activation - The costs To lower the costs the n u m b e r of unit operations must be limited The problems were solved as follows[I] : - The most suited raw materials as far as the plant structure concerned are nut and shell byproducts. For a combination of reasons including quality, cost and availability broken olive stones, produced in large quantities in the Mediterranean as a by-product of olive oil extraction were chosen as the preferred raw material. - To impregnate the raw material the olive stones are milled. The mixture of milled olive stones and phosphoric acid presents a plastic mass of ideal consistency for low pressure forming during pelletising to produce pellets between 2 mm and 4 mm in diameter. These ,,green" pellets however are still very soft and easily deformed so are not suitable as such to be activated in the normal m a n n e r directly in a rotary kiln. A gently drying step prior to carbonisation is required to give s t r e n g t h and some rigidity to the product. Carbonisation followed by recovery of the phosphoric acid by water washing results in a relatively dense pelletised activated carbon with the desired properties.The self binding capability of the raw material under activation conditions is due to the presence of higher concentrations of lignin, the n a t u r a l binding agent in plant structures present in the raw material. The above mentioned product has been brought in to the market under the name of CNR 115.

827 5.

THE PERFORMANCE MARKET

OF ELCD

CARBONS

AVAILABLE

IN T H E

In Table 1 the activated carbons available commercially in the ELCD market are given. Table 1 Carbon type

BWC

Diameter(mm) or Ds0 (~t)

Form

NORIT CNR 115 1 1 . 5 - 13.0 extrudate

Hardness w/w%

Attrition mg/min

2.2 (ram)

75

55

WV-All00

1 1 . 6 - 12.2

granular

950 (~t)

41

270

BAX 1100

1 0 . 7 - 11.9 extrudate

2.1 (mm)

80

75

750 (~t)

58

189

PICA

9 . 9 - 10.4

granular

In Figure 4 the butane adsorption isotherms on volume basis are given. The results in show a higher adsorption capacity for the CNR compared to the WV-A 1100, the BAX 1100 and the PICA carbon and a lower heel for the WV-Al100, the BAX 1100 compared to the PICA carbon and the CNR 115. Thus the lower adsorption capacity of the WV-Al100 and the BAX 1100 are compensated more or less by a lower heel. For the PICA carbon also the lower adsorption capacity as the higher heel results in relatively low BWC.

butane adsorption isotherm a

,4

(-~

E

lO

,~1

!

0 > E 0

r" .,, fJ~

0 0.0001

I

I 0.001

0.01

0.1

rel. pressure

I

4;~ NORIT CNR 1 15 4 PICA FX1 135

.~. WV-A1100

Figure 4. Butane adsorption isotherm of different carbon types.

4E: BAX 1100

I

828 The pressure drop as a function of the air velocity for these types of carbon is given in Figure 5. The results in table 1 and figure 5 show that only the NORIT CNR 115 and the BAX 1100 combine a high BWC with a low pressure drop.

I Pressure drop in air I

L..

.................

(D >. (0 ..Q O

| ...............

4 ...............

~ ...............

| ...............

~ ...............

4 ...............

| ................

6

o

. . . . . . . . . . . . . . . . . ". . . . . . . . . . . . . . . - . . . . . . . . . . . . . . . -" . . . . . . . . . . . . . . . " . . . . . . . . . . . . . . . - . . . . . . . . ~" . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lit O

~

E ~

b~4

W

1 1 !5

O

.P

"~

E

C N R

O WV-A1100 -

,

13_

NORIT

<~

~

........................................................... ~'"! ................................................................ ~ 2 -

,'iO

BAX

1100

O

PIC F X 1 0 3 5

i

9

m

..~" <>

I ; i i _J;] 1................ i..... ~.,.r~- ............... -"~.,4-~~......'.-

0

0.2

0.4

IA ................................

0.6

0.8

air velocity (m/s) Figure 5. The pressure drop for different carbon type.

In Figure 6 the GWC is given as a function of the cycle number for NORIT CNR 115 and BAX 1100. Figure 6 shows a higher GWC for the CNR after a large number of cycles. These results agree with a higher BWC for the NORIT CNR 115 compared to the BAX 1100. The GWC of the NORIT CNR 115 is higher than 60 g/L, even after 75 cycles. In Figure 7 the emission measured during cycle 10 and 75 is given. Figure 7 shows a low rest emission up to breakthrough. The breakthrough curve is equal for both carbons and is not dependent of the cycle number. So it is allowed to measure the emission up to the prescribed breakthrough concentration while in the SHED test the emission is integrated in time. The breakthrough of the CNR starts later t h a n the breakthrough of the BAX 1100, resulting in a higher GWC. In Figure 8 the influence of ethanol is given. The graph shows that using ethanol in the gasoline results in an increase of GWC for all the cycles. This changing the composition of gasoline resulting in more butane and in ethanol the vapour has a positive effect on GWC.

829

I Gasoline Working Capacity I 110 100

90 13)

80

o (9

70

. 60

~

.......................................... _ _ ........i................................................................. _

-'' ...... ~""-"'---~;-------; :--::-.:-."~:-.-.:.-~:.-.i..-;.s.....-~.S~

...........

i

50 0

10

20

I

30

40 50 cycle number

_CNR115

60

..BAX1100

70

80

I

Figure 6. The GWC as a function of the number of cycles.

FID S I G N A L 500 '

I

| 400

..................................................................................................................................

~ ............................................

i

i

~

, 300 ................................................................................................................................. ! .............................................

200 ...............................................................................................................................................................................

'/

a

'

1 O0

.............................................................................................................................. , ...........................................

[

,,' ./1

0 0

20

40

60

80

time (min)

--NORIT CNR 115 (10) --NORIT CNR 115 (75) .... BAX 1100 (10)

. , B A X 1100 (75)

Figure 7. The breakthrough concentration during cycle 10 and cycle 75.

830

Gasoline Working Capacity NORIT CNR 1 15 Influence of ethanol 12

! I I

i

o L .................................................................................................................................................................................................... !

10

A

i~

E

I 0

O O

-~

/... lb%

8

~-. |

(J

~

----

.

..................

.

. __o

.......

. o .....

.__.

ANDARD

........

(.9 6

....................................................................................................................................................................................................

4

0

I

I

I

20

40

60

80

cycle number Figure 8. The influence of ethanol in the gasoline. 6.

CONCLUSIONS

The NORIT CNR 115 is an activated carbon which meets the requirements for the future where a high GWC/BWC has to be combined with a low pressure drop. To produce this carbon some technical problems had to be solved. Changing the composition of gasoline by introducing ethanol has a positive effect on GWC.

REFERENCES

1. J.D. MacDowall, A method of producing granular activated carbon, EP Patent No. 423967 A2 (basic) 1991. 2. J.D. MacDowall, P.D.A. McCrae and W.M.T.M. Reimerink, The development of an effective activated carbon for evaporative emission control. Paper presented 25 th ISATA Silver Jubilee Meeting, Florence, l-5th June 1992. 3. W.M.T.M. Reimerink, J.D. MacDowall, P.D.A. McCrae and S.A. Fris, The development of an activated carbon for evaporative loss control device (E.L.C.D), Carbon Conference, Essen, June 1992. 4. Volatile organic compound emissions in western europe: control options and their cost-effectiveness for gasoline vehicles, distribution and refining. CONCAWE, 1987.

831 5. California evaporative emission standards and test procedures for 1978 and subsequent models. 6. Federal Register EPA CFR part 86. 7. Publication E.U. No. L242187. 8. Federal register, Vol. 55, No. 13, 1990. 9. Determination of butane working capacity. ASTM 5228-92.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

R e m o v a l of m i c r o o r g a n i s m s / p a r t i c u l a t e s

833

from indoor air

T.K. Ghosh and A.L. Hines Particulate Systems Research Center and Nuclear Engineering Program University of Missouri-Columbia, Columbia, Missouri 65211, USA 1. I N T R O D U C T I O N Particulates and airborne microorganisms contribute significantly to the indoor air pollution problem. Particulates and bioaerosols can have an immediate impact on public health, such as headaches, nausea, respiratory infections, allergies, humidifier fever, or Legionnaires' disease. The loss of billions of dollars per year in medical expenses and lost productivity has been attributed to these pollutants. A majority of bacteria and viruses found in indoor environments originate from h u m a n s and pets through sneezing, coughing, dander and saliva, which increases with an increase in inhabitant population. A variety of microorganisms also are present in the outdoor air. Once microorganisms are introduced into an indoor environment, they can settle in amplification sites where they thrive and grow. Amplification sites include any site with the proper pH, temperature, and moisture content. Common growth sites include moist wood and cloth, damp floors, shower heads, and refrigerator drain pans. A number of fungi and bacteria also arise from humidifiers, air conditioning systems, cooling towers, and ventilation and heating systems. A list of microorganisms that are commonly found in indoor environments, their sizes and pathogenic properties is given in Table 1. Although outdoor air is a major source of indoor particulates, various indoor activities such as the burning of wood and tobacco smoking can elevate significantly their indoor concentrations. Tobacco smoke is also a source of both toxic chemical vapors and particles in indoor environments [ 1 ] . Recent laboratory studies have identified more than one hundred chemical compounds in the vapor and particulate phases of tobacco smoke [2]. A number of these compounds are considered toxic and carcinogenic to humans. The mean level of the mass of respirable suspended particulates (RSP) in smoking environments is generally higher t h a n that found outdoors [3-6]. The concentration of RSP in homes with one or two smokers could be two to four times higher than that in homes without a smoker. Tobacco smoke has a wide distribution of particle size. Particles in the size range of 0.1 to 0.5 ~m from tobacco smoke, which are most difficult to remove since they are too small to capture by interception or

834

,..o

,..o o o,.-i

c~

0 ,_0 ,I,.4 9 r...4

4~

0

9

0

9

n~

~ o

9

b.0

b,l

,i-,i

0

~

I~

0

0

~

o

9

~

o

~

9 ,..4

o ~

0

00

~D

~I

0

~ ~

~o 0

~

,,,

0

b.O

00

~

"~0

~

9 ~,,,i

o~

"~

o

4~

0

00

0

9 r,,.i

0

0

0

0 ,.O

0 ,.O

.i-t

4~

r.~ o~..4 4~

9 r-.,I

o or-I

El o

~i,-.I

8

c5

r

~o

9 i,,.~

c5

I--.4

1.o

or-I

r

~o

9 ,--~

~

"~ r.~ or-I

o

4-~

~..~ 4-~

~ ~ g

o

....

c6

dl

.~

o

c5

Ixl G~

l:::q "

o6

&

Ixl 0

0

~

I~ oO

r

dl 0

~.~

Ixl

Ixl b.-

c5 06

0

.,.~

d

,4 ob c5

~

c5 ~ c5

~

,4 ~ c5

~

c5 l_~ c5

,-q

o

o;..~ ~o

~

~

o

o o

.,.-4

o

m o m

o m o

o

~g

o

.~,-I

9

o or]

o

b~ o

OQ

c~

o

[;]

o .m

9

o

0

0

o

-~

.r-t

,~

~

-~

o

.,~

o

9r-I

~ "~

.'m o ",-~

~o

"~

o co ,~

b.0

~

"'-'

c~

~

~

o ~

m

~

0

m

6 ,-4

6

0

o

~

~b d

~

"~

~ ~

o .~ .~

o~

~4 d

o

d~ d

..~ ~

~g

~

o

o

o

~

o

o "~

..~

c~

0 0

co

o

o~

b.0 .~.~ ~

~

c~

o

c~

.b

co c~

c~ .,.-i

o o

.,-i

.~,,t

co

o

9

835

9

r~

0

00

r~

0

r~

r.~

CD 0

c~ r-.-i

<

L~

c~

c~

0

c~

0

i

CO

i

d

0 C~1

o'~ ~9 ~

0

~,~

~. .~

~

.~~

~o

.,.~ ~

~

i

d

o

d

o0

d

0

9,,.,,~

c~

0~,,~

-,~

0',1

~

0

~

,~

d

c~

~

0

~

d

d

~

o.

c~

~

CD

CD C~

C~

9

~3

0

9

9

.f-.i

9

r~ 9

d

o ~

c~

c~ 0

C~ b~ 0 C~

c~

9,,~

o

o

r.~ C~

9

4-~

9 r~

0

9

9

9 ~.-.i

9

0~.-i

0

b~

0

~D

~ ~

0

r~

.,.,

.r-.~

836

~

837 impaction and too big for removal by the diffusion process, constitute a significant mass-fraction of indoor respirable particle matters. These particles have a high probability of penetrating to the pulmonary compartment of lungs causing adverse health effects. Particulate matters can also be a carrier of a number of toxic substances by adsorption on its surface. Carbon particles, such as those emitted during combustion Of the paper portion of a cigarette, are efficient adsorbents for a number of organic and inorganic compounds and can carry them into lungs. This can have a potentially damaging effect on the h u m a n body. Particles that contain an attached toxic substance can increase a person's physiological response to that substance above the normal tolerance level of the same individual. The relation between the size ranges of selected particulates and the potential for lung damage is shown in Figure 1.

GREATEST LUNG DAMAGE ! CEMENT DUST

TOBACCO SMOKE

Diameter o f "

Antropogenic

I

COAL DUST

Pollutant

OIL SMOKE

Particles

I

I

FLY ASH

I

AUTO EXHAUST

!

VISIBLE LIGHT

RAIN

VIRUSES

HUMAN H A I R T

Important SMALL

Lengths

BACTERIA

MOLECULES

and C o m m o n Particle

SEA SALT PARTICLES

Diameter

0.0001

0.001

0.01

0.1

BEACH SAND

1 2

5 10

100

1000

10,000

Microns Figure 1. Size range of various particulate matters (adapted from Indoor Air: Quality & Control by A.L. Hines, T.K. Ghosh, S.K. Loyalka and R.C. Warder, Prentice Hall, New Jersey, 1993, p. 117).

838 2. C O N T R O L S T R A T E G I E S

Various preventive methods such as maintenance and cleaning, increased ventilation (air exchange rates), humidity control, physical methods of sterilization, application of biocides, and the use of mechanical filters have been suggested as a means of controlling microbial growth indoors [7]. Preventive maintenance and cleaning are probably the easiest methods of controlling microorganisms in buildings and residences. Regular maintenance of air handling systems and fan-coil units, cleaning of drain pans, and the periodic replacement of filters have been found to reduce microbial concentrations in indoor air. Cleaning also reduces the likelihood of microbial amplification in filters, cooling towers, and humidification systems. Biocides are primarily used to control microbial growth in water such as in humidifier water reservoirs, cooling tower water, and in any place with stagnant water. The most frequently used biocides include hypochlorites, phenolic compounds, and glutaraldehyde [8]. Biocides kill microorganisms by physically attacking their structure and inhibiting their ability to reproduce. The selection of a proper biocide is dependent not only on its microbial inhibiting factor, but also on its dosage and frequency of use, and on its impact on the environment [9]. Microorganisms can develop a resistance to biocides after a certain period of time [10]. This resistance arises from the ability of organisms to mutate and change their biochemical behavior. Several studies have shown that the use of multiple biocides with scheduled treatment plans can reduce microbial resistance. However, most of these biocides are toxic to h u m a n s [8]. A particulate filter combined with an activated carbon filter is most frequently used for removal of gaseous constituents and particulates resulting from tobacco smoke, since ventilation alone is not sufficient to remove or to dilute Environmental Tobacco Smoke (ETS) concentrations indoors below levels recommended by the USEPA. Various types of particulate filters, such as metal prefilters, V-bag filters, HEPA filters and electrostatic precipitators have been used. Turk et al. [11] studied 38 commercial buildings in the Pacific Northwest region of the USA whose air exchange rates averaged 70% higher than that recommended in the ASHRAE 62-1989 standard. The RSP levels in buildings with smokers were on average 40% higher t h a n the USEPA standard. Olander et al. [12] examined a number of air cleaners for removing tobacco smoke including electrostatic precipitators, electret fiber filters, fiber filters, and ionizers. Some of these were also equipped with a second cleaning unit such as activated carbon beds, impregnated alumina beds, an ionizing lamp, and an electron generator. In addition to monitoring particulates in the size range of 0.01 to 7.5 microns, the cleaning rates of carbon monoxide, ammonia, formaldehyde, nitric oxide, nitrogen dioxide, hydrocarbons, hydrogen cyanide, and ozone were also measured. The removal efficiencies for particulates were 90 to 95%, but those for gaseous constituents were below 50%. Sandberg and Mellin

839 [13] used an electrostatic field and an ion generator to remove tobacco smoke and mineral dust from a 3.8 m 3 Plexiglas-room. The reduction in mass concentration was about 60% w h e n both the electric field and the ion generator were in operation. The increased ventilation rate h a d little effect on the removal efficiency. Similar results were obtained when only the electric field was in operation. M a n n and Airah [14] combined an efficient particulate filter with an activated carbon filter to control odors and to remove various gaseous and vapor constituents and particulates of tobacco smoke from working environments. A self contained recirculating air purifier t h a t incorporated a particulate filter and a carbon adsorption unit was found to be most effective. A detailed review of the removal of ETS from indoors using various methods can be found in Hines et al. [7]. Pierce et al. [6] evaluated four air cleaning systems for reducing ETS components in offices. Air cleaning devices used in their study was a combined system of a particulate filter and a gas-solid adsorption unit for removal of particles and vapor components of ETS, respectively. Results of their study showed t h a t such air cleaning devices operating concurrently with dilution ventilation can be effective in reducing the levels of nicotine and RSP in a designated smoking area. Devices equipped with a H E P A filter were most effective in reducing RSP level, whereas a carbon media was found to be the best for removal of nicotine. The adsorption capacity, or the lifetime, of a carbon filter depended on a n u m b e r of factors, such as a m o u n t s and types of carbon used in a filter, concentrations of ETS in air, operating t e m p e r a t u r e , and relative humidity of air.

3. A D S O R B E N T B A S E D REMOVAL S Y S T E M S G r a n u l a r bed filters can be attractive alternatives to m a n y existing systems for particulate removal from air. Beds packed with inert solids, such as sand and glass beads, have been used in the past to remove particulates from industrial effluents, but with limited success. A review of the subject has been provided by Zenz and Krockta [15]. A typical application is the t r e a t m e n t of a particle-laden flue gas, which is introduced in a g r a n u l a r bed from either the top or bottom. When the resistance to flow increases to a p r e d e t e r m i n e d level, as indicated by the pressure drop, compressed air is blown in a direction opposite to the process air flow to clean the bed. Jackson and Calvert [16] employed Raschig rings, spheres, and spherical packings of 0.5, 1, and 1.5 inch in size to remove oil mist. Particle collection efficiencies up to 50% were obtained for 1 micron particles. P a r t e s k y et al. [17] used a sand bed to remove fly ash and sulfur dioxide. Another example is the study by Gebhart et al. [18] who used a glass bead m e d i u m to remove 0.1 to 2 ~m diameter particles. Kalen and Zenz [19] developed a trial unit to capture particles from a fluid cracking unit in the petroleum industry. Due to a variety of mechanical and economical problems, however, these units are no longer commercially available [20]. To improve the

840 performance of these granular bed filters, it has been suggested to augment the filters with an electrostatic field. Such a bed handles high temperatures and high gas flow rates at low pressure drops with reasonable efficiency [21,22]. However, these systems were developed mainly for use in industrial processes. Therefore, they are not suitable for use in other indoor settings and consequently were not tested with tobacco smoke, other respirable particulates, or microorganisms. However, if desiccant materials are used as the filtering media, these type of units may be used for cleaning indoor air. Desiccant based air conditioning systems are finding increased applications over the traditional vapor compression refrigeration systems due to their more efficient handling of the latent heat load and lower operating and maintenance costs. As a result, there is renewed interest in studying the indoor air quality enhancement capability of desiccant systems. In desiccant based systems, moist air is passed through a bed containing desiccants where moisture is removed from air by adsorption. The dry air from the bed is further conditioned to adjust its temperature and humidity. Silica gel, molecular sieves, and activated alumina are primarily used as desiccants because of their excellent water adsorption capacity. Among these materials, silica gel and molecular sieves are used most frequently as desiccants to dehumidify the air in desiccant based air-conditioning systems. The adsorbent based filter bed can also capture particulates and remove or kill various microorganisms from air. They can even adsorb gaseous constituents of tobacco smoke. However, these materials are still being investigated to evaluate their effectiveness for removing respirable particles, especially tobacco smoke from indoor air, as well as their effectiveness in killing microorganisms. The primary objective of the use of solid desiccant systems is to dehumidify air. However, their water adsorption characteristics can change dramatically if particulates are present in air that is being processed. Therefore, most of the research are related to studying the moisture adsorption characteristic of desiccant materials in the presence of particles and tobacco smoke. Pesaran and co-workers [23-25] exposed several desiccants to cigarette smoke in a chamber, which they described as a Desiccant Contamination Test Facility. The chamber was designed to hold 100 sample test tubes, each of which was three inch in length and 0.32 inch in diameter. The sample tubes were packed with various types of desiccant materials such as silica gel, molecular sieve-13X, activated alumina, activated carbon, silica gel on a tape, and lithium chloride, and were exposed together to smoke in the chamber. Resulting data showed that the moisture adsorption capacity of all materials decreased due to exposure to smoke. Farouk et al. [26] reported that the rate of moisture adsorption of molecular sieve, silica gel, and activated alumina decreased by 50% in the presence of 5% agricultural dust. However, they found that the equilibrium moisture adsorption capacities of these desiccants were not affected by dust particles having a 10 to 20 pm diameter. This suggests that dust particles blocked the pores, adding extra resistance to the diffusion process, thus decreasing the units efficiency in dust removal.

841 4. F U N D A M E N T A L S A N D M E C H A N I S M S

Particulate removal systems work on the principle that as a gas stream containing particles flows through a filtration device the particles are acted on by various external forces that cause their separation from the gas stream. Either individually or in a combination, the following mechanisms generally cause the separation of the particles from the gas stream: 1) Sedimentation: Particles present in gas streams are allowed to settle under gravity to the floor of a chamber. 2) Electrostatic precipitation: Particles are electrically charged and then subjected to an electric field for their removal from the gas stream. 3) Inertial deposition: Suspended particles tend to keep moving in their original direction even when the flow path is changed. Ultimately, particles loose their inertia and deposit on the filter surface. 4) Brownian diffusion: Particles suspended in a gas are always in Brownian motion. When the gas stream flows around solid surfaces, the random motion of the particles will bring them in contact with the surface, where they adhere and collected. Devices designed based on Brownian diffusion are more effective in removing smaller particles. Design of particulate removal devices, therefore, requires the knowledge of the motion of particles under various environmental conditions. The factors that will affect the motion of the particle and thereby the removal efficiency of the device include: a) particulate concentration in the stream to be cleaned, b) the size distribution of the particles in the stream, and c) the gas flow rate. The design of these devices is described in detail in various publications and text books [27]. The mechanisms by which microorganisms and particulates adsorb to and desorb from adsorbent surfaces involve many other phenomena. The adhesion of microorganisms and particulates onto a solid surface may occur in three stages; a) deposition of microorganisms, b) attachment, adhesion, or adsorption, and/or c) growth or killing on the solid surface. The deposition of microorganisms onto solid surfaces depends on the initial movement of microbial cells through a suspending fluid toward and from a sorbent surface. It can involve nonsorptive charge behavior and diffusional, gravitational, and convective transports. The subsequent attachment involves a number of forces acting between the microbial cell and solid surface. A list of these forces are given in Table 2. The attachment and any subsequent colonization of microorganisms depend on the nature of the solid surface. Solid surfaces may be porous (such as that of silica gel, molecular sieves, and activated carbons), or nonporous (such as that used in filters). Although pore diameters of these adsorbents are in the same order of magnitude as that of the microorganism, no study is available to show

842 Table 2 Forces of attraction and repulsion between microbial cells and adsorbent surfaces Forces of Attraction Hydrogen, thio, amide, and ester bonds Ion-pair formation (NHa + ...-OOC-) Interparticle bridging (polyelectrolytes) Electrostatic attraction between surfaces of similar charge Electrostatic attraction due to image forces Surface tension van der Waals forces of attraction Electromagnetic forces Hydrodynamic forces Diffusional forces Gravitational forces Forces of Repulsion Charge repulsion between surfaces of similar charge van der Waals forces of repulsion Steric exclusion (hindrance) Negative chemotaxis (cellular mobility) Adapted from Adsorption of Microorganisms to Surfaces, G. Bitton and K. C. Marshall (Eds.), John Wiley & Sons, New York, 1980, p. 22.

that microorganisms are in fact adsorbed in the pores. It is more likely that the microorganisms either deposit or adhere to outer surfaces of the solid. The initial interaction between a microbial cell and solid surface, which is generally called the adsorption stage, depends mainly on the characteristics of the solid surface. As shown in Figure 2, the mechanism of adhesion of a microorganism onto a surface depends on the adsorbent particle size. If the adsorbent particle is significantly larger than microorganisms, a number of microbial cells can become attached to a single large particle. When the size of an adsorbent particle and a cell is in the same order of magnitude, they can mutually interact. In the case of adsorbent particles smaller t h a n microbial cells, a single cell can attach itself to a number of adsorbent particles. The mechanisms for adsorption, rate of transport, and equilibrium capacities can differ in all three cases. In the case of desiccant materials, once microbial cells attach themselves to the solid surface, the desiccant may adsorb water and fluids from the cell causing their death due to desiccation or cell rupture.

843

Adsorbent particle

Microbe a) Adsorbent particles larger than microbial cells ~

~~,..~

)~~)'~~'~--

Adsorbent particle

Microbe

b) Adsorbent particles and microbial cells of equal size . "~"

.~

Adsorbentparticle

2~. ~ ~ ~ . . . ~ Microbe c) Adsorbent particles smaller than microbial cells Figure 2. Schematicrepresentations of microbial adsorption (adapted from Adsorption of Microorganisms to Surfaces, G. Bitton and K.C. Marshall (eds.), John Willey & Sons, New York, 1980,p. 13). When using solid adsorbents to remove microorganisms and/or particulates, the material will be either packed in a bed or coated in a monolithic honeycomb. Considering the size ranges of respirable suspended particles and microorganisms in indoor air, Brownian diffusion is expected to be the controlling mechanism in bringing the particles into contact with the surface. Although the forces that are mainly responsible for adhesion of microorganisms to adsorbent

844 surface may include van der Waals and electrostatic forces, chemical bonding through covalent, ionic, and hydrogen bonds, and hydrophobic bonding, interfacial reactions between microorganisms and adsorbent surfaces are also very important in particle removal. Microorganisms are capable of producing exudates which can condition the surface to which they attach making them more or less sticky. According to Baier [28], this interfacial film makes the solid surface suitable for biological attachment. Although a wealth of literature concerning the attachment of microorganisms at various solid-liquid interfaces are available, only limited information is available on the attachment of microorganisms in solid-air interfaces. The kinetics of diffusional processes have been applied successfully in developing rate equation for both adsorption and desorption of microorganisms for solid-liquid systems. The Langmuir and Freundlich adsorption isotherms adequately describe the equilibrium adsorption of microorganisms. However, further experimental data are required to verify the application of these adsorption isotherms to solid-air systems. 5. L A B O R A T O R Y AND F I E L D T E S T R E S U L T S 5.1. P a r t i c u l a t e s A packed bed can provide dual action as a typical filter for airborne particulates and as an adsorbent for other indoor air pollutants. Laboratory scale experiments [29,30] were conducted using each of three adsorbents; silica gel, molecular sieve 13X and activated carbon in order to assess their tobacco smoke removal capability from air. The filtration capability was investigated by passing room air through an adsorbent bed. Approximately 25 g of the adsorbent samples were packed into a 0.5 inch diameter glass tube. The flow rate through the bed was maintained at 5 L/min. These experiments were conducted at room temperature (about 296 K) and relative humidity ranging from 30% to 60%. The total number of particles decreased from the original 4.8-6.0 x 103 particles to 1.3-3.0 x 103 particles/cm 3 of air at the initial period of the experiment. There was no significant difference in the reduction of removal efficiency (27% to 50%) when the relative humidity of the room was varied. Also, a similar removal efficiency was noted for each of the three adsorbents, which had similar particle sizes: 6x12 mesh for silica gel, 4x8 mesh bead for molecular sieve, and 6x16 for activated carbon. This gave similar voidages in all three beds, and, therefore, caused the filtration efficiency to be in the same range. To study the removal of environmental tobacco smoke from indoor air, a filamatic piston pump was used to duplicate the puffing of a cigarette. The design of the smoking machine is given in details by Lee et al. [30]. A lighted cigarette was attached to the smoking apparatus by a plastic tube. Downward motion of the piston created a vacuum in the pump chamber and opened an intake valve, thus drawing about 35 cm 3 of air through a burning cigarette into

845 the p u m p chamber. The movement of the piston was controlled by a computer and the time required to intake 35 cm 3 of air was adjusted to about 2 second. This was approximately equivalent to taking a puff on a cigarette by a person. After about 2 seconds, the piston moved u p w a r d which closed the intake valve, and at the same time opened another valve to vent the smoke trapped in the v a c u u m c h a m b e r to the atmosphere. The process was repeated at one minute intervals. This essentially duplicated the smoking p a t t e r n of a person. During the one m i n u t e cycle, the cigarette burned idly and generated the sidestream smoke in a Plexiglas tube, which was further diluted, allowed to age, and its relative h u m i d i t y was adjusted to 50% in a 20 feet long 2" I.D. Plexiglas tube to obtain environmental tobacco smoke (ETS). A schematic diagram of the experimental set-up is shown in Figure 3.

~

0

0

0

0

0

f

0

Vacuum Transducer

w~,..__~

io

o

/

o o o 1

Sampling Port Thermocouple "

Humidity Probe To Gas Chromatograph

a~r

Smoking Machine

l 'ii

Thermocouples

I b-

in l

Total Hydrocarbon Analyzer

Filter Holder

To Gas Chromatograph

'

Humidity Probe

Air Blower

Figure 3. A schematic diagram of the smoking apparatus.

The adsorption bed was made of a stainless steel tube t h a t was 30 cm in length and 4 cm in i n t e r n a l diameter. A stainless steel screen was placed at the bottom of the column to support the samples. A thermocouple was placed at the middle part of the bed to m e a s u r e the t e m p e r a t u r e and was placed in the laboratory where the t e m p e r a t u r e was approximately 296 K. The variation of t e m p e r a t u r e

846 in the laboratory during an experiment was within _+ 1 K. The gas mixture flowed through the column from the top to the bottom. A detail description of the column can be obtained in [29]. A Condensation Nucleus Counter (CNC) was used to count the number of particles in the air. The numbers are maintained in the ETS level by adjusting the dry air flow rate t h a t was used to dilute the smoke in the first tube. A combined humidity and t e m p e r a t u r e indicator was used to measure relative humidity and temperature of the mixture. A filter paper was used to collect particles at the inlet and outlet of the bed. A Hydrocarbon Analyzer was used to measure the total hydrocarbon concentration, which was reported as total methane. Following exposure to tobacco smoke, the equilibrium water adsorption capacity of exposed samples was measured gravimetrically in an all glass a p p a r a t u s using a Cahn C-2000 electrobalance. A description of the adsorption system and the experimental procedure can be obtained in [31]. The electrobalance was capable of weighing up to 3.5 g with a sensitivity of 1 pg. A vacuum of 10 .4 m m H g was obtained in the system prior to initiating an adsorption run. Approximately 50 mg of samples were sufficient for obtaining the equilibrium data. The samples are placed in a small a l u m i n u m bucket t h a t is suspended from the electrobalance. Samples can be regenerated, in situ, under various conditions. The conditions used in this study for regenerating the samples have been described later in this section. Following regeneration, water vapor was introduced into the system in steps and the equilibrium adsorption data were obtained at 298 K. After equilibrium was reached, as indicated by a constant sample weight, the pressure and weight gain were recorded. Subsequent points were obtained by introducing more water vapor into the system and allowing it to equilibrate. About 25 grams of samples were used in an experiment. Prior to each adsorption run, the bed was regenerated to remove any moisture or impurities that might have been adsorbed during the weighing and packing of the bed. This was done by heating the bed at 473 K for 12 hours with a stream of nitrogen flowing through the bed at a rate of 500 cm3/min. Following regeneration, the bed was cooled to the room temperature and a cigarette was lit manually in the smoking machine. The cigarette was withdrawn from the chamber at the end of a 10-minute combustion period. On average, ten puffs were taken from a cigarette. The flow rate of air containing ETS was approximately 8600 cm3/min in the last section of the Plexiglas tube. About 4000 cm3/min of the mixture was passed through the bed, while t h a t through the inlet particulate filter was 4000 cm3/min. A flow rate of about 600 cm3/min was required for the CNC for counting the number of particles. The number of particles, relative humidity, temperature, and hydrocarbon concentrations of the gas mixture at the outlet of the bed were measured periodically using appropriate instruments. An adsorption run was continued for 6 to 10 hours depending on the type of material used in the experiment. Following which the regeneration of the bed was started

847 by heating it to 473 K, while a stream of nitrogen flowed through the bed at a rate of 500 cma/min. The regeneration of the bed continued until the hydrocarbon concentration of the exhaust air was not detectable by the hydrocarbon analyzer. An adsorption-regeneration cycle was repeated for 15 times with the same sample. After each adsorption run, the filter papers were collected, and their weights were determined as soon as possible. After 15 adsorption-regeneration cycles, adsorbents were collected from the top, middle, and bottom sections of the bed. The remaining adsorbent was regenerated, and a sample was collected for analysis. The BET surface area was determined for all four samples. The concentration of ETS in the air was controlled by monitoring the number of particles. The number of particles in the air was controlled between 30,000 and 35,000 particles/ cm a of air, however, it increased sharply for a very short period of time following a puff by the machine. The number of particles then decreased steadily before taking another puff. This variation in the number of particles between puffs for a ten-minute period is shown in Figure 4. The methanol extracts of the particulate phase of the sidestream smoke and the ETS

* Time elapsed before a puff: 60 sec 4.0 -

3.5

X

E ro ~ ~

3.0 Z

2.5 0

120

240

360

480

Time (sec) Figure 4. Variation in the number of particles in the gas mixture during a ten minute combustion period.

600

848

were analyzed by the GC/MS spectrometer. A number of compounds were identified including nicotine in the sidestream smoke, however, only a few of them could be detected in the ETS. Although the number of particles were controlled between 30,000 and 35,000 particles/cm a of air in all the experiments, the mass concentration ranged from 600 to 1000 mg/m a. This was also an indication that the particulate phase did not have a uniform composition. The concentration profiles for water vapor from the three adsorbent beds for the 15th cycle are compared in Figure 5. Although silica gel has a greater preference for water than for hydrocarbons because of the presence of hydroxyl groups on the surface, the bed could not be saturated with water even after 10 hours of adsorption. In contrast, the molecular sieve 13X bed became saturated with water vapor in about 6.7 h. As expected, water broke through the activated carbon column quickly. Activated carbon does not adsorb a significant amount of water at relative humidities below 40%. A sharp rise in the adsorption capacity for water occurs at humidities around 50%, which is mainly due to the pore filling [32]. A temperature rise above the set adsorption temperature (298 K) was observed in both silica gel and molecular sieve beds due to release of heats during adsorption process. Studies by Babu [33] and Leu [34] have shown that water alone can cause the temperature of the bed to rise by 20 to 50 K when silica gel or molecular sieve was used as the adsorbent. The tailing in the breakthrough

1.0 {jp'~

o -,..4

.,,.-

9 w

t

" 9

0.8

9

9

9 I

J

w

,

~

~

i

Molecular sieve 13X 9

A

9

9

A

a

A

9

9

9

9

9

9

9

9

9

Activated carbon

9

T,-a

_,,..~" Silica gel 9 ~ 9149

o

0 . 6 ~ 9 1 4 99

),--I

o

0.4-

o9 o9

.4.a

0 ::S O O

"

0.2 i F

~'-"" *** *9

0.0

, 0

Relative humidity: 50% Adsorbent weight: 25 g Flow rate: 0.0126 g-mole/(min.cm') Number of particles at the inlet: 30000-35000 particles/cm 3 Inlet concentration of hydrocarbons: 3.5-4 ppm Temperature: 298 K

, 500

1000

Time (min) Figure 5. A comparison of breakthrough characteristics of water vapor from different adsorbent beds.

1500

849 curve for water vapor from silica gel bed may have been due to this temperature rise in the bed. More water was adsorbed by silica gel as the bed cooled causing the tailing in the breakthrough curve. However, the shape of breakthrough curves for water vapor from the both silica gel and molecular sieve beds remained nearly the same in all 15 cycles and water broke through the bed almost at the same time. This suggested that adsorption mechanisms did not change from cycle to cycle. The equilibrium adsorption data of water vapor obtained under static conditions are shown in Figures 6 through 8. The data were obtained on four types of samples (Samples A, B, C, and D). Two samples, one each from the top (Sample A) and bottom (Sample B) sections of the bed, were collected after the 15th adsorption cycle. These two samples were kept in the test chamber under vacuum for 12 hours (regenerated under vacuum only) before starting the adsorption of water vapor. One of the four samples was not exposed to tobacco smoke at any time (fresh sample, Sample C), and it was regenerated by heating at 473 K under vacuum for 12 hours. The last sample was collected from the bed after the 15th adsorption cycle, but all the adsorbents in the bed were mixed thoroughly before taking the sample (Sample D). This particular sample was regenerated in the test chamber by heating at 473 K under vacuum for 12 hours.

0.4

'

'

'

'

I

'

'

,"

'i

"I

'

'

'

'

I

'

'

'

'

I"

"

'

"

'

I

'

'

'

'

&

0.3 A _0 i

@ 9

9

9

[]

0.2

9 a,~

0.1

9

9

9

t ,

0.0

0

Fresh sample, never exposed to tobacco smoke Exposed to tobacco smoke and regenerated at 473 K under vaccum Sample from top of the bed after exposure to tobacco smoke Sample from bottom of the bed after exposure to tobacco smoke

9

9

,

I

I

I

5

,

I

i

,

I

10

,

L,

I

,

I

15

.

.

.

.

I

20

.

.

.

.

I

25

.

.

.

.

30

Pressure (mmHg) Figure 6. Comparison of the isotherm data for water vapor at 298 K on samples of molecular sieve exposed to tobacco smoke with that of an unexposed sample.

850 0.5

'

9 9

0.4

I

'

'

'

'

m

,

,

,

,

m

,

'

Fresh sample, never exposed to Lobacco smoke Exposed to tobacco smoke and regenerated at 473 K under vacuum Sample from top of the bed after exposure to tobacco smoke Sample from bottom of the bed after exposure to tobacco smoke

9 m

0o

'

0.3

~

dko

.

,

,

,

i

t

t

I

9

0 Oak

-'-'

9

0.2

9

& 4~

0.1

i

ni

9

m9

4,

9

9

mb

4~ m

.m

9

me,

t

0.0

i

J

0

t

m

i

i

i

i

5

I

I

|

i

10

i

i

i

i

|

!

15

[

20

25

Pressure (mmHg) F i g u r e 7. C o m p a r i s o n o f the i s o t h e r m data for w a t e r v a p o r at 298 K on s a m p l e s o f m o l e c u l a r sieve e x p o s e d to t o b a c c o s m o k e w i t h o f an u n e x p o s e d sample. 0.4 9 9 9 9

Fresh sample, never exposed to tobacco smoke Exposed to tobacco smoke and regenerated at 473 K under vacuum Sample from top of the bed after exposure to tobacco smoke Sample from bottom of the bed after exposure to tobacco smoke

, 49.

011

0.3 A

r R 0.2 "

A~

0.1 E9

0.0

&

0

,

5

.

,

10

9

,

15

20

Pressure (mmHg) Figure 8. Comparison of the isotherm data for water vapor at 298 K on samples of activated carbon obtained from different sections of the adsorption bed with that of fresh and regenerated samples.

851 Silica gel collected from the top of the bed was brown in color, while that in the bottom was light brown. The brown color of silica gel was due to the adsorption of tar and nicotine from the smoke. Since most of the tar and nicotine were removed in the top section of the bed, the color of the silica gel in the subsequent layers were lighter. It should be noted that the air stream containing tobacco smoke was fed from the top of the bed. Although the scanning electron micrographs showed solid-like deposition on the outer surface of silica gel samples, they did not become sticky after exposure to tobacco smoke. Probably most of the nicotine and tar were adsorbed on the pores, not on the outer surface of silica gel. A similar analysis could not be done for molecular sieve due to its dark brown color. As expected, silica gel samples collected from the top of the bed (sample A) had the lowest water vapor adsorption capacity. The silica gel exposed to tobacco smoke could not be regenerated completely even after 12 hours of heating at 473 K. Silica gel has significant number of pores with diameters less t h a n 10/k, and these pores became plugged or covered with contaminants. This was also evident from the color of the regenerated samples, which had a dark brown appearance even after regeneration in the test chamber. The water isotherms on molecular sieve 13X retained type II characteristics on all the samples exposed to the tobacco smoke. The sample from the top of the bed (Sample A) and the sample from the bottom of the bed (Sample B) had almost the same water vapor adsorption capacity. When the samples were regenerated by applying both heat (473 K) and vacuum (Sample C), they had the same adsorption capacity as that of the fresh sample (Sample D). The molecular sieve 13X had an average pore diameter of 10/k, and it appears that most of the tars and nicotine could be removed from the pores under these regenerating conditions. The water adsorption data of carbon samples obtained from the top and bottom of the bed are compared in Figure 8 with that of a fresh sample and with the one that had been exposed to tobacco smoke and further regenerated at 473 K for 12 hours. As can be seen from the figure, there was no difference in the water adsorption capacity among these samples and with a fresh activated carbon. It appears that activated carbon could be regenerated by applying vacuum only. This indicates that tobacco smoke was adsorbed in larger diameter pores. The surface areas of these samples measured by using nitrogen as the adsorbate at 77 K supported the earlier observation. The surface areas of these samples are given in Table 3. Results obtained by Pesaran et al. [25] supported the data for silica gel, however, their data contradicted Lee et. al's [30] findings for molecular sieve and activated carbon (see Table 4). It may be noted that Pesaran et al. exposed their samples to a much higher concentration of tobacco smoke than used by Lee et. al. Since the adsorption capacity of a solid is dependent on the gas phase concentration of the adsorbate, at higher concentrations, tobacco smoke constituents may have been adsorbed in smaller pores from which desorption was difficult.

852 Table 3 Available surface area of adsorbents after exposure to tobacco smoke Samples

S i l i c a Gel Unexposed silica gel a

Available surface area (m2/F)

672

Silica gel exposed to tobacco smoke and then regenerated b

550

Sample t a k e n from the top of the bed following exposed to tobacco smoke c

527

Sample t a k e n from the bottom of the bed following exposed to tobacco smoke d

542

Molecular Sieve Unexposed molecular sieve a

395

Molecular sieve exposed to tobacco smoke and then regenerated b

350

Sample t a k e n from the top of the bed following exposed to tobacco smoke c

40.5

Sample t a k e n from the bottom of the bed following exposed to tobacco smoke d

40.4

Activated carbon Unexposed activated carbon a

1052

Activated carbon exposed sample to tobacco smoke and then regeneration b

936

Sample from the top of the bed c

815

Sample from the bottom of the bed d

916

a Sample was not exposed to tobacco smoke at any time and was regenerated at 473 K for 12 hours under vacuum before the test. b Sample was exposed to tobacco smoke for 15 adsorption and regeneration cycles and was regenerated after the 15th cycle at 473 K for 12 hours under vacuum. c Sample was collected from the top of the bed after 15 adsorption and regeneration cycles and was regenerated under vacuum only for 12 hours before the test. d Sample was collected from the bottom of the bed after 15 adsorption and regeneration cycles and was regenerated under vacuum only for 12 hours before the test.

853

Table 4 Comparison of moisture capacities virgin, clean, and adsorbent samples exposed to tobacco smoke Capacity

Adsorbent Silica gel Silica gel on

Loss

Ambient (Control)

Exposed to tobacco smoke

Difference in capacity

Most loss occurs

0-30%

30%-80%

20%-50%

In 1 month

20%-60%

70%-95%

30%-60%

In 1 month

30%-70%

30%-70%

In 1 month

tape

Molecular sieve

Activated 5%-20% 15%-50% 10%-40% In 1 month carbon Source: A. A. Pesaran and T. J. Dresler, Report No. SERI/PR-254-3677, 1990, p. 47.

The breakthrough characteristics of hydrocarbons present in tobacco smoke from the silica gel, molecular sieve, and activated carbon beds are compared in Figure 9. The total hydrocarbons measured as methane equivalent has been

2.5

,

r ....

,-

"--T

|

|

|

|

[

.

.

.

.

I

'

'

|

'

I

, ....

|

'

I

Inlet concentration of hydrocarbons - 35-40 ppm 2.0

0 .4-a

0

G~

9 9 9

Silica Gel Molecular Sieve 13X Activated Carbon

1.5

1.0

9

0000000000000000

9

9

9

9

9 N,~ ow, e~

0 9149

o.5 ~

O

9

t t

9

kk&k&&&&kkkkAkkkAkkkk

Z

0.0 0

100

200

300

400

500

Time (min)

Figure 9. Comparison of breakthrough characteristics of hydrocarbons present in tobacco smoke from different adsorbent beds.

854 reported in this figure. In a silica gel bed, the total hydrocarbon concentration decreased slowly reaching a minimum value t h a t ranged from 35% to 75% of the inlet concentration during the first 10 to 15 minutes of an adsorption cycle. However, the bed became s a t u r a t e d with hydrocarbons after about 150 minutes. The concentration of hydrocarbons in the vapor phase was in the range of 35 to 40 ppm compared to about 20,000 ppm for water vapor. Since silica gel has a greater affinity for water vapor, hydrocarbons were not able to compete with water for an adsorption site. In a molecular sieve bed, the outlet concentration of hydrocarbons decreased to almost zero as soon as the air-tobacco smoke mixture was introduced into the bed. However, as the experiment continued, the outlet concentration rose above their inlet concentration. This type of behavior was not observed in a silica gel bed. In the initial period of a run, hydrocarbon molecules were able to compete with water molecule and occupied the vacant cages in the molecular sieve crystal lattice. However, water eventually displaced the hydrocarbons from the pores, because of its stronger affinity for the surface. As shown in Figure 9, hydrocarbon concentrations at the bed outlet from the activated carbon bed remained at about 13 to 16 ppm after six hours of continuous operation. The total hydrocarbon concentration in the inlet gas stream was about 35 to 40 ppm. It may be noted that water vapor concentration of the same air stream was between 20,000 and 30,000 ppm. Initially water occupied most of the pores before being displaced by hydrocarbons. It appears t h a t as the experiment progressed, hydrocarbons started to displace water and occupied these sites. As a result, the hydrocarbon concentration at the outlet stream remained the same for a while. Since the hydrocarbon concentration was very low compared to t h a t of water vapor, the displacement of water vapor by hydrocarbons could not be detected. It also appears t h a t certain types of hydrocarbons present in ETS could not be adsorbed by the carbon. In the present study the total amounts of hydrocarbons were determined. Therefore, it was not possible to identify these compounds. Both silica gel and molecular sieve are desiccant materials having strong affinity for water t h a n hydrocarbons. Therefore, both these materials became s a t u r a t e d with ETS-hydrocarbons r a t h e r quickly compared to activated carbon. The ratios of the number of particles at the outlet to inlet of the bed for these adsorbents for the 15th cycle are compared in Figure 10. At the beginning of an adsorption cycle, a significant number of particles were captured in the bed, however, the efficiency decreased as the experiment continued, and a significant number of particles escaped from the bed. As can be seen from the figure, more t h a n 95% of the particles were removed by the bed initially. During this time, the pressure drop in the bed also increased, though marginally. As the experiment was continued, more particles started escaping from the bed. In all three beds the same amounts of adsorbent material were used. However, due to the different bulk density, the bed heights were different. This may have contributed to the difference in the initial capture efficiency. All three beds lost

855 1.2

I

9

1.0

u ~ O

kkA9

9

'

it A

'

9

'

'

I

9

~

II 9

0.8

0

0.6

2: O

o.4

9 9 9

Silica Gel M o l e c u l a r Sieve 13X Activated Carbon

Q~ . w...4

'~ ~

0.2

~00 9 0000 O

9 N u m b e r of particles at the inlet = 30000-35000 p a r t i c l e s / c m 3 9

0.0

0

100

200

300

,

,

I

400

,

i

500

600

T i m e (min)

Figure 10. Comparison of particle capture characteristics of different adsorbent beds. 7000

6000

5000 p~

4000 3000 o 2000

1000

0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Cycle Number * Experiment

was continued

f o r 24: h o u r .

Figure 11. Mass of particles retained by different adsorbent beds after different adsorption cycles.

856 its particle capture capability after a certain period of time. This type of behavior may be due to the different surface charge of these materials. It is also interesting to note that the captured particles were removed from the surface during regeneration by heat. This may be due to the thermophoretic effect, since the material surface was hot due to direct application of heat to the adsorbents through the bed-wall whereas a relatively cool air was flowing near the surface. The regeneration air was not heated prior to introduction to the bed. The particle capture efficiency of a bed for different cycles was determined by performing a mass balance, and the data for three adsorbents are shown in Figure 11. The amount of particles remaining on the bed after an adsorption cycle was determined from the weight change of filter papers. The capture efficiency remained approximately the same in all 15 cycles.

5.2. Removal of m i c r o o r g a n i s m s The microorganism removal capability of an adsorbent bed was evaluated by passing room air that contained about 270 colony forming units (CFU) per cubic meter of air through a 2.5 cm diameter bed packed with 25 g of the adsorbent. The relative humidity of the room air was about 30%. The room air was drawn through the bed at a rate of 10 L/min for a period of 10 minutes, and the air samples were collected by using an Andersen impactor. A schematic diagram of the apparatus is shown in Figure 12. An Andersen sampler is an impactor type device consisting of six stages through which air is drawn consecutively. Each stage contains approximately 400 holes, whose diameters decrease in the subsequent stages. The jet velocity is uniform for each stage but increases for each succeeding stage. Immediately below each stage is a specially designed petri dish that contains the nutrient to facilitate microbial growth. When the velocity imparted to the particle is sufficient, its inertia overcomes the aerodynamic drag and the particle leaves the stream of air and deposits on the nutrient medium. Otherwise, the particle continues to travel to subsequent stages. Each succeeding stage removes the largest remaining particles, with the last stage collecting those that are left. The petri dishes containing the nutrients were prepared at least 24 hours prior to their use and were stored at room temperature. They were checked periodically for contamination. Following collection of microorganisms, the plates were incubated for the appropriate period of time (24 hours for bacteria and 4872 hours for fungi). Incubation temperatures used were 306-310 K for bacteria and 298 K for fungi. Following incubation, the individual colonies on each plate were counted on a Quebec colony counter. Various types of nutrient media were used to isolate each type of colony (fungi or bacteria) based on their morphology or color. The media used to identify the organisms included nutrient agar, potato dextrose agar, and luria agar. Sabouraud dextrose agar and rose bengal agar were used to isolate the Actinomyces and fungi. Fungi were characterized by their shape, growth characteristic in the broth and on the agar, and the mode of reproduction.

857

Outdoor air Ca, 0

Humidity probe

I

"1

Valve

Q

Humidity probe

Jacketed adsorption column Andersen impactor

Pump

Flow m e t e r

[itrogen br use during :generation of Lebed

Vent Three way valve

Figure 12. A schematic diagram of the experimental system for studying removal of microorganisms.

Bacteria were isolated on the basis of biochemical tests and differential media. Biochemical tests included gram-stains, endospore stains, and sugar fermentation tubes. Several other tests such as casein hydrolysis, catalase production, starch hydrolysis, nitrate reduction, litmus milk reaction, urea hydrolysis, growth at 5% NaC1, and citrate utilization were performed on specific bacteria to further differentiate the genus and species. Bergey's Manual of Determinative Bacteriology [35] was used to identify the biochemical responses. The organisms collected at various locations of the laboratory were found to have the same colony morphology, gram reaction, and cellular morphology. The organisms identified during the fall of 1988 included Staphylococcus aureus, Staphylococcus saprophyticus, Bacillus pasteur, and Bacillus circulans. Three

858 different varieties of fungi were also isolated. The genera identified in the engineering laboratory include Actinomyces (bacteria) and Aspergillus (fungi). Staphylococcal bacteria were also observed in several samples. Of the 22 isolates, nine were fungi and thirteen were bacteria. Although silica gel and molecular sieve 13X beds were able to remove statistically significant numbers of microorganisms from room air, the activated carbon bed was considerably less effective. Silica gel and molecular sieve 13X adsorbed all the moisture from air causing death to the microorganisms through desiccation. The water adsorption capacity of activated carbon from air whose relative humidity is below about 42% at room temperature is extremely small. Consequently, most microorganisms survived in the activated carbon bed. Results from the three beds are shown in Figure 13.

300

"~

250

"O

'x~

o

xco

L)

200

E ~

150

I~ 0D

o 6~IX

=

a:

o

o~

~"~

,~

100

=

50

o

Z

E ~ e~

e~

.~

0

Air Analyzed Figure 13. Microorganism removal capabilities of different adsorbents.

859

Kovak et al. [36] i n v e s t i g a t e d the microbial r e m o v a l capability of desiccant b a s e d air-conditioning units in t h r e e field locations a n d in a l a b o r a t o r y setting using specific bacteria. Two of the field units were installed in p a t i e n t a r e a s in two different hospitals a n d the t h i r d one was located in a commons room in a n u r s i n g home. All t h r e e units h a d s t a n d a r d 30% filters at the inlet. Rooms without a desiccant b a s e d air-conditi0ning unit at each facility were used as controls. The reduction in b a c t e r i a a n d fungi counts in these t h r e e facilities with respect to control rooms counts are shown in Figure 14. An a v e r a g e reduction of 39% was r e p o r t e d for Site 1. Nine out of ten s a m p l e s showed decrease in the microbial count, with a h i g h of about 73%. Only one s a m p l e showed no reduction. At Site 2, all ten s a m p l e s exhibited reduction in microbial counts, with an average of 64%. The t h i r d site also h a d an a v e r a g e reduction of 64% in the microbial counts.

80

Site 3 70-

Site 2

"iili

ii ' i '"~all

60@

I'!:l~,iii!}i'~l!,i~,~, ,',:1 50-

Site I ~

40-

~

30-

,!!!t,t,t!, Itl~tltil~lliii!lJt il II, li,!il lili|i ; ;, ":11

t!i: i:il

i

IJltl II ;Ill ,~:, ,',',~t II,,!]ilIl!ll ll/!~, i,,]: I

irft~tiit~,,i,:, , i

i,,lilitl!t lil [,t]ftifl,lti,i,! f,iliii[ill, t! IIllIt'i!ili't"ittti'i!l 20-

J~ll ~,lf ~',, ,,, Iiii / '!i', :

10-

Figure 14. Reduction in bacteria and fungi counts in three field tests (adapted from Kovak et al., ASHRAE J., 1997, p. 62).

860 L a b o r a t o r y tests were carried out using seven m i c r o o r g a n i s m s t h a t included

Staphylococcus aureus, Pseudomonus aeruginosa, Escherichia coli, Staphylococcus epidermidis, Enterococcus faecalis, Candida albicans, and Mycobacterium fortuitum. Except for Mycobacterium fortuitum, an overall reduction was observed for other six microorganisms. The a v e r a g e percent reductions for individual m i c r o o r g a n i s m s are shown in Figure 15. Tests were conducted b e t w e e n 5 to 20 times for each o r g a n i s m s a n d the a v e r a g e percent reduction is r e p o r t e d in this figure. The six microorganisms showed in the figure are opportunistic p a t h o g e n s a n d are responsible for various disease a n d infection in h e a l t h care facilities. W h e n all the d a t a are considered, an a v e r a g e reduction of 38% was observed for these seven microorganisms. A reduction for Mycobacterium fortuitum was not observed m a y be due to their r e s i s t a n c e to desiccation a n d heat.

o

r~ 9 I-..I

opU .,,.-q

.P,,I

l~176 1 8O -!

60

013

~

I-i

r,~

O

~

.~.ql

9 wt

fd~

,,a

~I

m

40 o

20

0

Type of Microorganisms Figure 15. Reduction in counts for selected microbes in laboratory test data (adapted from Kovak et al., ASHRAE J., April, 1997, p. 63).

861 Both the laboratory and field tests data support that the desiccants are capable of reducing airborne microorganisms, although the exact mechanism has not been identified. However, these studies do suggest that this technology may be helpful in reducing the number of bioaerosols in indoor air both directly through desiccation and killing, and indirectly by providing dry air. REFERENCES 1. C.J. Proctor, A Comparison of the Volatile Organic Compounds Present in the Air of Real-World Environments With and Without Environmental Tobacco Smoke, 82rid Annual Meeting of the Air and Waste Management Association, Anaheim, California; Paper 89-80.4, 1989. 2. B.S. Hulka, Measuring Exposure and Assessing Health Effects of Environmental Tobacco Smoke, in: Indoor Air Quality, H. Kasuga (ed.), Springer-Verlag, Germany, 1990, 16. 3. J.D. Spengler, Atmospheric Environment, vol. 15 (1981) 23. 4. D.J. Eatough, F. M. Caka, J. Crawford, S. Braithwaite, L. D. Hansen and E. A. Lewis, Atmospheric Environment, 26A(12) (1992) 2211. 5. L. Wallace, J. Air & Waste Management Assoc., 46 (1996) 98. 6. W.M. Pierce, J. N. Janczewski, B. Roethlisberger, M. Pelton and K. Kunstel, ASHRAE J., November (1996) 51. 7. A.L. Hines, T. K. Ghosh, S. K. Loyalka and R. C. Warder, Investigation of Co-Sorption of Gases and Vapors as a Means to Enhance Indoor Quality, Report No.GRI-90/0194, p. 268, NTIS No. PB91-178806, 1990. 8. ACGIH (American Conference of Governmental and Industrial Hygienists) Guidelines for the Assessment of Bioaerosols in the Indoor Environment, The Conference, Cincinnati, Ohio, 1989. 9. G.A. Cappeline, J. G. Carroll and S. D. Strauss, Power, 56 (1977). 10. S. Shair and H. F. Dorrington, Power Engineering, April, (1976) 103. 11. B. H. Turk, D. T. Grimsrud, J. T. Brown, K. L. Geisling-Sobotka, J. Harrison and R. J. Prill, ASHRAE Transactions, 95 (1989) 422. 12. L. Olander, J. Johansson and R. Johansson, Air Cleaners for Tobacco Smoke, Proceedings of the Fourth International Conference on Indoor Air Quality and Climate, West Germany, 2 (1987) 39. 13. M. Sandberg and A. Mellin, Artificial Electrostatic Fields as Air Cleaning Systems in Room, Proceedings of the Fourth International Conference on Indoor Air Quality and Climate, West Germany, 1 (1987) 231. 14. J. L. Mann and M. Airah, Australian Refrigeration, Air conditioning and Heating, 41(8) (1987) 21, 23, 28. 15. F.A. Zenz and H. Krockta, Brit. Chem. Eng. Process Technol., 17 (1972) 224. 16. S. Jackson and S. Calvert, AIChE J., 12 (6) (1966) 1075. 17. L. Partesky, L. Theodore, R. Pfeffer and A. M. Squires, Journal of the Air Pollution Control Association, 21(4) (1971) 204. 18. J. Gebhart, C. Roth and W. Stahlhofen, J. of Aerosol Science, 4 (5) (1973) 355. 19. B. Kalen and F. A. Zenz, Chemical Engineering Progress, 69 (5) (1973) 67.

862 20. H. P. Dibbs and P. Marier, AIChE Symposium Series, No. 147 (1975) 59. 21. S. C. Saxena, R. F. Henry and W. F. Podolski, Advances in Transport Processes, 4 (1986) 465. 22. M. Shapiro, G. Laufer and C. Gutfinger, Aerosol Science and Technology, 5(1) (1986) 39. 23. A. A. Pesaran and B. K. Parsons, Desiccant Materials Contamination Research: Status Report, SERI, December, 1987. 24. A. A. Pesaran and C. E. Bingham, Desiccant Contamination Research: Report on the Desiccant Contamination Test Facility, Report No. SERI/PR-254-3457, 1988. 25. A. A. Pesaran and T. J. Dresler, Desiccant Contamination Experiments: Preliminary Results, Report No. SERI/PR-254-3677, Solar Energy Research Institute, Golden, CO, 1990. 26. S. M. Farouk, G. H. Brusewitz and P. D. Bloome, Desiccant Moisture Sorption as Altered by Dust, ASAE Paper No. 80-3084 (1980). 27. R. C. Flagan and J. H. Seinfeld, Fundamentals of Air Pollution Engineering, Prentice Hall, NJ, 1988. 28. R. E. Baier, Advances in Chemistry Series, 145 (1975) 1. 29. S-Y. Lee, Removal of Environmental Tobacco Smoke from Indoor Air by Solid Adsorbents, M. S. Thesis, University of Missouri-Columbia, 1993. 30. S-Y. Lee, T. K. Ghosh, A. L. Hines and D. Novosel, Gas Separation and Purification, 9(4) (1995) 285. 31. A. L. Hines and T. K. Ghosh, Water Vapor Uptake and Removal of Chemical Pollutants by Solid Adsorbents, Gas Research Institute, Chicago, IL, Report No. GRI- 92/0157.2, p. 234, NTIS No. PB95-104691, 1992. 32. N. M. Hassan, T. K. Ghosh, A. L. Hines and S. K. Loyalka, Carbon, 29(4) (1991) 681. 33. J. C. Babu, Adsorption of H20 and Indoor Pollutants in a Fixed Bed, M. S. Thesis, University of Missouri-Columbia, 1991. 34. S-W. Leu, Co-adsorption of Indoor Air Pollutants in the Presence of Water Vapor on Solid Adsorbents, M. S. Thesis, University of Missouri-Columbia, 1993. 35. D. H. Bergey, in: Bergey's Manual of Determinative Bacteriology, R. E. Buchanan and N. E. Gibbons (eds.), Williams and Wilkins Co., Baltimore, MD, 1989. 36. B. Kovak, P. R. Heimann and J. Hammel, ASHRAE J., April, (1997) 60.

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski(Editor) 9 1998Elsevier Science B.V. All rights reserved.

863

S t u d i e s of m i x e d a d s o r p t i o n l a y e r s f o r m e d b y p o t e n t i a l c o r r o s i o n inhibitors K. Sykut, J. Saba Faculty of Chemistry, M. Curie - Sktodowska University, 20-031 Lublin, Poland

1. I N T R O D U C T I O N The issues of corrosion of metals deserve particular attention due to the wide use of metals in all fields of technology as well as the increasing corrosive aggresiveness of the n a t u r a l environment caused by air and water pollution. Due to their high electric conduction, the corrosion of metal most frequently is of the electrochemical character. The mechanism of the electro-chemical corrosion is based on the emergence of a variety of galvanic macro- and micro-cells: concentration oxygen cells, cells of various electrolyte concentration, activising and passivising cells and electrolytic cells. The basic corrosion research concerns the kinetics of the electrode processes due to the fact t h a t in wet environments corrosion is caused by these very processes. Anti-corrosive measures involve either the modification of the corrosion environment or the inhibition of spontaneous electrode processes. The former group of anti-corrosive measures includes: elimination of oxygen from water, increase of the solution's pH, desalting of water, dehumidification of air, t e m p e r a t u r e increase in order to prevent steam condensation, elimination of solid pollutants from air or water and others. The use of corrosion inhibitors is supposed to inhibit the anodic and/or cathodic processes by means of an increase in overvoltage or the blocking of the active metal surface. Anodic inhibitors, most frequently anions, cause the increase in the anodic polarisation potential, which results in the corrosive potential being shifted towards the positive values. In order to achieve the sufficient effectiveness, the inhibitors should be used in relatively high concentrations, e.g. l g/dm 3. Cathodic inhibitors are usually cations which, as they deposit on the cathode, block the surface of metal. As § and Sb § cations may serve as an example which inhibit dissolution of iron in acids or polyphosphates which in water solutions, in particular in the presence of Ca § , produce large colloidal cations. The efficiency of inhibitors may be improved by applying their appropriate m i x t u r e s , e.g. a cathodic inhibitor mixed with an anodic inhibitor.

864 Organic corrosion inhibitors belong to the same type of compounds as inhibitors of metal etching in acids or inhibitors of the development of crystals, the so-called brighteners, used to deposit glossy electroplated coating. The majority of these compounds contain nitrogen or sulphur or both the elements. Due to strong adsorption across the entire surface of metal they may at the same time inhibit anodic and cathodic processes. The inhibitor develops a mono or polymolecular coating whose attributes depend on the energy of its bonding as well as its density. Thiourea belongs to popular inhibitors of iron corrosion in the acid environment. By using the electrochemical impedance spectroscopy method [1] it was found t h a t in a solution of H2804 the inhibition with thiourea takes place in the concentration below 0.1mM. The rate determining the step for the anodic dissolution reaction of iron in the low polarisation region was a nucleophilic additive reaction, first order with respect to thiourea. The rate determining step for the cathodic process a Heyrovsky reaction, 0.5 order with respect to thiourea and 1.5 order with respect to H § ion. Using the above research method it was found that that the maximum inhibitor effectiveness of thiourea on the corrosive behaviour of ARMCO iron in deaerated 0.5M H2SO4 appears at the concentration of some 0.1 mM [2]. The adsorptive bahaviour of thiourea on the electrode surface up to its peaks follows a Frumkin-type isotherm with lateral repulsion (the attraction constant is a= - 4.4), where the molecules are vertically adsorbed on the Fe via the S atom. Thiourea acts as a mixed inhibitor up to the critical concentration. It decreases the dissolution of Fe and the H2 evolution reaction by blocking the electrode surface. Thiourea acts also as an inhibitor for copper corrosion at acidic pH values exclusively, but enhances copper electrodissolution through the anodic film in alkaline solutions [ 3 ] . Electrochemical data recorded for copper electrodes immersed in a solution containing thiourea indicates that thiouera is adsorbed at potentials more negative t h a n the onset of copper oxide formation, and the passive layer is built upon a layer of adsorbed thiourea. Thus, oxidation of the specimens leads to the establishment of a competition between the formation of an anodic layer containing thiourea and the oxides layer, and the electrodissolution of the metal through the anodic film. Studies of phenylthiourea as a corrosion inhibitor for mild steel prove that the addition of 250 ppm of phenylthiourea was sufficient to inhibit corrosion of Fe 0.25C - 0 . 4 7 M n - 0.0089%Si steel in acid solvents containing 1000 ppm sulfite, 5000 ppm chloride and 500 ppm fluoride ions [4]. The inhibition efficiency was 94% at 50~ Potentiodynamic polarisation studies on carbon steel in HC1 solutions indicate that in the presence of phenylthiourea, the polarisation of both partial reactions is increased leading to a marked decrease in corrosion rates and debasing of the corrosion potential. This was demonstrated by significant reductions for potentials taken not far from the corrosion potentials [5]. Amino acids are among popular inhibitors of steel corrosion [6,7] and a l u m i n i u m corrosion [8, 9] in acidic environments. In general, amino acids are weak inhibitors of mild steel in 3M H2SO4 solution, with the exception of

865 tryptophan. On the other hand, S-containing amino acids (cysteine, cystine and methionine) acts as corrosion inhibitors with an inhibition efficiency of 76+4% at 10-2M concentration. Combination of either cystine or cysteine with methionine gave higher inhibition efficiency than that in the presence of methionine alone. Inhibition was predominantly anodic and their adsorption conformed with the Temkin isotherm. The increase in the inhibition efficiency in the rise in temperature and the resulting Temkin isotherm indicated that inhibition occurring through chemisorption. Aspartic acid is a mixed-type inhibitor for aluminium in acid chlorine solutions. The Flory-Huggins adsorption isotherm describes well the experimental data for aluminium in the presence of different concentrations of aspartic acid. The large negative value of free adsorption energy indicates a strong interaction between the aspartic acid molecule and the corroding aluminium surface. The presence of Cl ions in a certain concentration range has a synergistic effect on the inhibition efficiency of aspartic acids. The reasearch on the inhibiting impact of selected amino acids demonstrates that they inhibit not only iron corrosion in 0.5M H2SO4 but also the corrosion of iron-chromium alloy but not the corrosion of chromium [I0]. The isomerism of the location of substituents in aromatic compounds exerts a specific influence on the corrosion inhibiting potential of these substances. At 5mM chlorosubstituted anilines concentration in 0.1M sulfamic acid, the efficiency of inhibitors for zinc decreases in order: aniline (79%)>pchloroaniline>m-chloroaniline>o-chloroaniline (65%) [II]. The mode of inhibition action appears to be chemisorption because the plot of log(|174 vs. log c obtained as straight line suggests that the inhibitors cover both the anodic and cathodic regions through general adsorption following Langmuir isotherm. In turn the studies of the inhibiting impact of isomers of aminobenzoic acid indicate that all the three isomers inhibit the corrosion of mild steel both in HCI and H2SO4 in order: orto-> meta-> para- [12]. The inhibition is greater in HC1 t h a n in H2SO4. The p r e d o m i n a n t behaviour is in the cathodic inhibitor mode. Some of the surfactants belong to the group of efficient corrosion inhibitors. The studies of the inhibition effect of some non-ionic surfactants on the corrosion of iron in acid chloride solutions showed t h a t inhibition efficiencies increased with increasing surfactant concentrations and a t t a i n e d a m a x i m u m around their critical micelar concentration [13]. A comparative study indicated 3.7-dimethyl1,2,3,6,7-oktanepentol was the best inhibitor, a cathodic type-inhibitor and acted on the cathodic reaction without modifying the m e c h a n i s m of the hydrogen evolution reaction. The impact of cation surfactants on steel corrosion in acid solutions is complex in n a t u r e affecting the dissolution of the metal and the cathodic reaction of H2 evolution [14]. The inhibitory effect of the surfactants increased with alkyl chain length and at concentrations above their critical micele concentration. The corrosion potential of the steel was unaffected by the presence of surfactants, however, the corrosion current decreased with increasing concentration of surfactants. Among the high-molecular compounds, polioxyethyl

866 ether d e m o n s t r a t e s the potential to inhibit the corrosion of zinc and its alloy with indium [15]. The use of a mixture of corrosion inhibitors oftentimes improves their efficiency resulting in synergistic effects, however, an antagonistic effect is also possible [16]. The interactions between inhibitors were t a k e n into account by second order equations. Examination of regression coefficient and their significance indicated t h a t synergism may exist if concentrations of inhibitor compounds are 30 mg/dm 3 and below, but if the inhibitor compounds are applied in concentrations of 100mg/dm ~ antagonism r a t h e r t h a n synergism takes place. The synergistic inhibition effect of sodium octylmerkaptopropionate and 8quinolinol was observed in the studies of iron corrosion in an a e r a t e d 0.5M Na2SO4 solvent [17]. A high inhibition efficiency 98.2% was obtained by a mixture of 3-10-4M octylmerkaptopropionate and 5 9 10-4M 8-quinolinol. It was concluded t h a t the high synergism of the mixture is mostly a t t r i b u t e d to coverage of the iron surface with a precipitate layer composed of iron oxide and a chelate of 8-quinolinol with Fe § - o c t y l m e r c a p t o p r o p i o n a t e complexes precipitated at defects preferentially. The research of the inhibiting impact on the corrosion of mild steel and 1M H2SO4 of some thiols: 2-mercaptobenzothiazol, 2-mercaptobenzimidazol, 2mercaptobenzoxazol indicate t h a t they are one mixed type inhibitors affecting both the anodic and cathodic processes by simple blocking of the active sites of metal [18]. The corrosion process is controlled by a charge-transfer proces. The presence of an electron cloud on the aromatic ring, the electronegative atoms (N, S and O) and the easily polarisable hetero atoms is likely to induce greater adsorption of the thiol molecule on the surface of the mild steel which can lead to an effective inhibition. The surface coverage area (0) of mild steel by the adsorption of different thiols was calculated from the uninhibited and inhibited corrosion currents. The addition of halide ions to the corrosive medium containing thiols, similarly as in the case of the application of aspartic acid [9], was found to increase the inhibition efficiency. The t h e r m o d y n a m i c values indicate t h a t the presence of inhibitors increases the activation energy. Given the experimental data for adsorption of surface-active substances on different metals, it appears likely t h a t the mechanism of creation of mixed adsorption layers on mercury will be subject to similar rules. The selection of mercury as the electrode material is determined above all be a very good reproducibility of m e a s u r e m e n t s resulting from the homogeneity and purity of the mercury/solution interface. Adsorption surfaces were studied which are produced either by substances which are popular corrosion inhibitors or which in t e r m s of their chemical composition are very similar to the corrosion inhibitors applied. Mixed adsorption surfaces were produced by two organic substances with opposite impact upon the kinetics of Zn (II) ion reduction. In selecting these substances the m u t u a l influence of the molecules of the adsorbants applied and the function of their size and structure was also t a k e n into consideration. Zn (II) ion electroreduction inhibitors: n-butanol (BU) and ethylene polyglycols with

867 m e a n molecular mass 400 (PEG 400) and 10000 (PEG 10000) were used as components of mixed adsorption layers, with tiourea (TU) [19] and toluidine isomers: m-toluidine (mT) and p-toluidine (pT) [20] - being used as accelerating agents. The adsorption equilibrium within the area of the adsorption potential similar to the potential of the zero charge (pzc) were examined using the classical t h e r m o d y n a m i c methodology: the adsorption p a r a m e t e r s of the double layer was determined using the F r u m k i n isotherm and the virial isotherm. Based on the potential drop across the inner layer ~M-2, the electrostatic parameters of the double layer were determined. The influence of mixed adsorption layers on the Zn (II) ion reduction kinetics constitutes a new, original research methodology which enables a description of the properties of adsorption layers at potentials which are distant from pzc, and by the same taken close to the value of corrosion potentials of non-precious metals. Obtaining insight into the laws governing the formation of mixed adsprption layers and their properties may constitute a further step on the way towards their fuller application.

2. A D S O R P T I O N M E A S U R E M E N T S 2.1. D i f f e r e n t i a l c a p a c i t y c u r v e s of d o u b l e layer The study of adsorption relating to the formation of mixed adsorption layers was conducted in the following system: constant concentration of the Zn (II) ion electroreduction inhibitor- increasing concentration of accelerating agents of the electroreduction. Water solution IM NaCIO4 was used as the basic electrolyte. The measurements were conducted using a three-electrode system: mercury electrode- saturated calomel electrode as a reference electrode- platinum spiral as an auxiliary electrode. The double layer capacity was measured using the ac impedance technique at a frequency of 800 Hz. A few measurements were also carried out at 200 to 1500 Hz in order to check the frequency dependence of the results [21]. In the potential range studied no dispersion of the capacitance was observed. Figures I, 2 and 3 present the differential capacity curves of 0.55M BU with increasing amount of TU, pT and mT, respectively. Comparison of curve shapes for differential capacity presented in the above Figures demonstrates that the introduction of TU or toluidine isomers to the solution containing BU always results in increased differential capacity. The extent of the increase as well as the area of potential at which it occurs depends on the type of substance as well as on the mutual concentrations. In the BU-TU system with a constant concentration of BU, the height of the BU desorption peak clearly depends on the concentration of TU, with slight changes in the peak potential (Figure I). In BUtoluidine isomers solutions, the desorption peak undoubtedly constitutes the resultant of the BU and toluidine desorption peak and its parameters change with the changes in BU concentration, with a constant toludine concentration, and in the opposite situation. The maximum

868

v E

J

080 btclld If 0.60

e

0.40

0.20

a,,,

0.00

I

I

,

0.2

0.4

i

0.6

I

i

0.8

I

I

,

1.0

i

I

1.2

I

,

1.4

1.6 -E/V

Figure 1. Differential capacity curves ofHg/1M NaC104 + 0.55M BU for different contents of TU: a) 0M, b) 0.005M, c) 0.033M, d) 0.088M, e) 0.33M, f) 0.55M.

t o l u i d i n e i s o m e r s c o n c e n t r a t i o n of O.05M r e s u l t s from its solubility in the solutions u n d e r e x a m i n a t i o n .

e

0.80

0.60 ]

0.40

0.20

0.00

I

0.2

,

I

0.4

I

I

0.6

z

I

0.8

,

I

1.0

,

1

1.2

~

I

1.4

,

I

1.6 -E/V

Figure 2. Differential capacity curves ofHg/1M NaC104 + 0.55M BU for different contents of pT" a) 0M, b) 0.0015M, c) 0.008M, d) 0.03M, e) 0.05M.

869

0.80

D e

G 0.60

^

0.4O

0.20

0.00 0.2

i

I

,

I

0.4

I

I

0.6

I

I

0.8

I

I

1.0

,

I

1.2

,

I

1.4

1.6 -E/V

Figure 3. Differential capacity curves of Hg/1M NaC104 + 0.55M BU for different contents of mT as indicated in Figure 2.

The differential capacity curves obtained for solutions comprising a constant q u a n t i t y of 10-4M PEG 400 or PEG 10000 and increasing quantities of TU or pT were presented in Figures 4 to 7. For mT, the shape of the curves is similar to the t h a t of curves presented in Figures 6 and 7.

0.80 E

G

0.60

0.40 -e * - - e

~

\

0.20

0.00

I

0.2

t

I

0.4

i

I

0.6

i

I

0.8

i

I

1.0

J

I

1.2

i

I

I

1.4

I

1.6 -E/V

Figure 4. Differential capacity curves of Hg/1M NaC104 + 10-4M PEG 400 for different contents of TU" a) 0M, b) 0.0055M, c) 0.011M, d) 0.055M, e) 0.11M, f) 0.33M, g) 0.55M

[22].

870

o.ao r

o

AA

C 0.60 0.40

b

0.20

_

0.00

,

0.4

I

_~ ....

0.6

I

_

0.8

1.0

i

,

k

1.2

....

1.4

d

1.6

-E/V

Figure 5. Differential capacity curves of Hg/1M NaC104 + 10-4M PEG 10000 for different contents of TU as indicated in Figure 4 [22].

0.40 _c G

d e

G 0.30

0.20

0.10

0.00

_

I __.,

0.2

....

I

0.4

J

I,

0.6

,

I.

0.8

,

_1 .

1.0

~

I

1.2

..z_

I

1.4

~

__1

1.6 -E/V

Figure 6. Differential capacity curves of Hg/1M NaC104 + 10-4M PEG 400 for different contents of pT: a) 0M, b) 0.0015M, c) 0.005M, d) 0.008M, e) 0.01M, f) 0.03M, g) 0.05M [23].

871 d

0.50

4 e

E ,

0.40

r

~2 0.30 0.20 0.10 0.00

I

0.2

I

I

0.4

J

I

0.6

i

I

0.8

,

I

1.0

,

I

1.2

~

I

1.4

I

I

1.6 -E/V

Figure 7. Differential capacity curves of Hg/1M NaC104 + 104M PEG 10000 for different contents of pT as indicated in Figure 6 [23].

The strong adsorption properties of PEG 400 and PEG 10000 are manifested in a significant decrease of the differential capacity over a wide range of potentials. For the concentration of 5"10-4M PEG 400, the range is from -0.35V to - 1.60V, while for 10-4M PEG 10000 the range of potential is still wider. It needs stressing t h a t the increase in the concentration of PEG 10000 affects only to a slight extent the differential capacity. Differences in the adsorption capacity of PEG 400 and PEG 10000 undoubtedly result from differences in size and shape of the molecules: PEG molecules with the mer n u m b e r of n<11 are zig-zag shaped, while those with the mer n u m b e r n>>11 are m e a n d e r shaped [24,25]. Figures 4 and 5 indicate t h a t the introduction of TU to a solution containing PEG results in increased differential capacity for the range of potentials from -0.3V to -1.2V. At potentials close to pzc, curves show the TU-typical h u m p which is larger and sharper in shape in the presence of PEG 10000, as compared to PEG 400. The presence of this capacity h u m p typical of strongly adsorbent anions or polar substances m a y be a result of changes in m u t u a l influences, mainly the electrostatic ones, between adsorbed molecules [26]. In the presented systems, the influences p e r t a i n both to TU as well as PEG molecules and m a y result in the e s t a b l i s h m e n t of a more orderly structure of the adsorption layer. The introduction of toluidine isomers to solutions containing PEG results in a dramatically different shape of the differential capacity curve, compared to the shape obtained in solutions without PEG: in the former case the curves show a strong adsorption peak and d e m o n s t r a t e no desorption peak (of course within the examined range of potentials), while in the latter case the adsorption peak is hardly visible, with the desorption peak being clear and its height increasing as

872 the concentration of toluidine goes up. Similarly to the T U - PEG systems, in solutions containing toluidine isomers, the adsorption peak is larger and sharper in shape in the presence of PEG 10000, compared to PEG 400. The appearance of a strong adsorption peak in the P E G - toluidine solution is connected more likely with competitive adsorption among PEG and toluidine molecules, t h a n with the previously postulated [27] change in the orientation of toluidine molecules. The comparison of mT and pT differential capacity curves in the presence of PEG does not d e m o n s t r a t e any significant differences in their shape, while in 1M NaC104 solution without PEG, desorption peaks for pT are markedly higher t h a n for mT. 2.2.

Determining

surface excess

The differential capacity of double layer values at sufficiently negative potentials presented in Section 1.1 significantly differ from the respective values obtained in the basic electrolyte 1M NaC104. Therefore an integration of the differential capacity curves was conducted starting from pzc. The values of pzc were m e a s u r e d for each solution using the streaming mercury electrode. Interfacial tension at pzc was m e a s u r e d by the m a x i m u m bubble pressure following Schiffrin's method. The values of pzc obtained facilitate a projection of the orientation of molecules adsorbed on the surface of mercury. The increase in the TU concentration results in shifting pzc towards negative potentials which is an indication of the adsorbtion of the polar molecule of TU oriented with its negative pole towards the surface of mercury. The addition of toluidine isomer to the solution of 1M NaC104 causes a clear shift of the pzc value towards negative values but only at their lowest concentrations. A further increase of the concentration of both isomers results in a slow shift of the pzc value towards positive potentials. Such changes of pzc resulting from toluidine concentration changes suggest t h a t at lower concentrations, toluidine molecules demonstrate a fiat orientation on mercury surface, thus being more responsive to the influence of u electrons of the aromatic ring from the surface of the electrode. At higher concentrations, the impact is reduced as a result of the changed orientation of toluidine molecules from fiat to oblique. Similar effects are observed in the following systems: toluidine i s o m e r s - PEG 400, yet at toluidine concentration values significantly higher t h a n in the 1M NaC104 solutions. In the presence of BU and PEG 10000, the increase in the concentration of toluidine results in shifting pzc only in the direction of negative potentials, similarly to solutions containing I- ions [28]. The introduction of BU, PEG 400 or PEG 10000 into 1M NaC104 results in shifting the pzc towards positive potentials and only at high concentrations of BU a small shift of the pzc in the opposite direction is observed. Up to the point where the full covering is achieved, BU molecules have flat orientation on the surface of the electrode [29]. As indicated by the electron density of BU molecule on carbon atoms, the first and third carbon atoms have the highest electron density. The atoms, which display some hydrophilic properties, together with the oxygen atom are oriented towards the solution [30]. On the other hand, the

873 orientation of PEG molecules on the mercury surface confirms the orientation postulated earlier by Jehring [31]: -CH2-CH2-O group is oriented towards the mercury surface with its carbon atoms which constitute the positive pole of the dipole, while oxygen atoms are oriented towards the solution. The values of the surface tension determined at pzc decrease with the increase of BU, TU and toluidine concentration in a comparable manner, despite significantly lower concentrations of toluidine in relation to the concentrations of BU and TU which suggests a higher surface activity of toluidine molecules with mT demonstrating slightly stronger adsorption properties t h a n pT. The values of pzc obtained in the mixtures studied are a resultant of the values obtained for the individual substances. In the B U - TU mixture, with concentration of both substances being equal, the value of pzc indicates a clear domination of TU in shaping the adsorption equilibrium. A similar domination is to be found in B U - mT and B U - pT systems, but only at maximum toluidine concentration values. The results obtained following integration of curves of differential capacity were used to calculate the Parsons' function ~ = 7 + ~'E, where ~, is the surface tension; a is the electrode charge; E is the electrode potential. The total surface excess at the constant charge was determined from equation:

F;

RT c3In c org

/

where Corg is the concentration of organic substance if constant ionic strength is supposed. The obtained values of total surface excess in the B U - TU, B U - pT and B U - mT mixtures are much higher t h a n for individual substances which indicates the synergistic character of adsorption [32, 33]. The fundamental information concerning the adsorption of the substances studied were obtained from the relative values of surface excess F', free adsorption energy AG ~ , interaction constants and electrostatic parameters of the inner layer. According to the Gibbs adsorption isotherm, the relative surface excess of TU, pT and mT is given by: F=-

1 (0(I) ) RT c~lncp c~,Ch

(2)

where the "p" index stands respectively for TU, pT or mT and the "h" index stands for BU, PEG 400 or PEG 10000. It was assumed in equation (2) that the average activity coefficients of the individual substances in the solution does not change because of the increasing concentration of TU, pT or mT. The surface pressure 9 required to determine Fp was calculated using the Parsons' function: t

874 O = A~ =~0 _~ [34, 35], where ~ s t a n d s for the average value obtained from the solution containing the inhibitor a n d the accelerating agent, while ~0 s t a n d s for the value obtained b a s e d on the solution with a c o n s t a n t concentration of the inhibitor. Figures 8, 9 a n d 10 p r e s e n t the dependence of surface p r e s s u r e vs. log Cp in solutions containing a c o n s t a n t q u a n t i t y of P E G for various electrode charges. On the other h a n d in solutions containing BU the dependence was p r e s e n t e d only for ~M = 0 (Figure 11).

PEG 400

"7

E 30

+2

+1

E e

PEG 10000

30

~1

o

-~

20

20

-

+1 ~01

-2

10

10

0

I

-2.5

I

-2.0 -1.5 -1.0 -0.5

I

~-3

0

0.0

I

I

I

I

-2.5 -2.0 -1.5 -1.0 -0.5

I

0.0 log cru

Figure 8. Surface pressure as a function of TU concentration in the bulk in 1M NaC104 + 10-4M PEG, the electrode charges ((~M/10-2C 9m -2) being indicated by each curve [22]. 40

"7

E

PEG 400

e

30

PEG 10000

30

+4

E

-2 -4

+4 +2

20 20

10_

10

0 -3.0

I

I

I

I

-2.5

-2.0

-1.5

-1.0

-3.0

I

I

I

-2.5

-2.0

-1.5

I

-1.0 log cpT

Figure 9. Surface pressure as a function of pT concentration in the bulk in 1M NaC104 + 10-4M PEG, the electrode charges (~M/10-2C 9m 2) being indicated by each curve [23].

875 4o

• e

PEG 400

PEG 10000

30_

30 +5 +4 +2 0 -2 -4

20_ 20

+2

lO

10

-4

0 -3.0

I

I

I

I

-2.5

-2.0

-1.5

-1.0

-3.0

I

I

I

-2.5

-2.0

-1.5

I

-1.0 log CmT

Figure 10. Surface pressure as a function of mT concentration in the bulk in 1M NaC104 + 104M PEG, the electrode charges (cyM/10-2C 9m -2) being indicated by each curve [36].

16 c

14 e

12

m

b

10 m

-

//

a//

m

-3

-2

-1 log cp

Figure 11. Surface pressure as a function of concentration in the bulk of TU(a), pT(b), mT(c) in 1M NaC104 + 0.55M BU at (YM= 0 [37].

876

S e l e c t e d d e p e n d e n c i e s of r e l a t i v e s u r f a c e e x c e s s e s Fp of a c c e l e r a t i n g a g e n t on its c o n c e n t r a t i o n in a s o l u t i o n a t pzc a r e p r e s e n t e d in F i g u r e s 12 a n d 13.

~E

a o

3

9

o

b

2

J

ca

[]

1

0 0.0

I 0.1

I 0.2

I 0.3

I 0.4

I 0.5

I

0.6 Cru / M

Figure 12. Relative surface excess of TU as a function of TU concentration in the bulk at pzc in the solutions" a) 1M NaC104 + 10 .4 M PEG 400, b) 1M NaC104 + 10 .4 M PEG 10000 [22].

~

3 E

c

0

E

x---

-3

I

I

I

-2

-1

0 log Cp

Figure 13. Relative surface excess of TU(a), pT(b), mT(c) as a function of bulk concentration of these substances in 1M NaC104 + 0.55M BU at ~M = 0 [37].

877 The values of relative surface excess of TU determined in the presence of PEG 400 are comparable to the values obtained in water solutions which do not contain the inhibitor [38], while when determined in the presence of PEG 10000 are much lower. The results obtained univocally indicate weaker adsorption of TU in the l a t t e r case. In the P E G - toluidine isomers system for selected concentrations of PEG, a synergistic n a t u r e of co-adsorption was observed which results most likely from the formation of a new structure of the adsorption layer, with fewer water molecules. The course of the dependence of F'p as a function of the concentration of TU, pT and mT in the presence of 0.55 M BU at pzc (Figure 13) indicates t h a t toluidine isomers are adsorbed more strongly t h a n TU despite significantly lower toluidine concentrations. These systems confirm also the fact of stronger adsorption of mT compared to pT. The effect was also observed in KC1 solutions [28]; it is no doubt connected with a different positioning of the hydrophilic group -NH.9 in relation to the hydrophobic r a d i c a l - C H a in the aromatic ring. On the other hand, stronger adsorption of toluidine compared to TU results most likely from a different affinity to mercury of the aromatic ring and the sulphur atom.

2.3. F r u m k i n a d s o r p t i o n i s o t h e r m s The adsorption of TU and isomers of toluidine was further analysed on the basis of the surface pressure data and by using the method previously developed by Parsons [34,38 and 39]. Individual pressure curves were superimposed graphically by shifting t h e m parallel to the log Cp axis. It was stated t h a t the interaction constant in the F r u m k i n isotherm i.e., the A p a r a m e t e r varied with electrode changes. Therefore, the constants F r u m k i n isotherm were determined from the equation:

~x= I 0 Jexp(-2A0)

(3)

where x and p are the molar fraction in the solution and the adsorption 1

_

\

coefficient is defined as exp/AG~ 0 is the coverage. The surface excess at s a t u r a t i o n Fs was e s t i m a t e d by extrapolating the 1/Fp vs. 1/% curve at different charges and different Ch to 1/% = 0. In the majority of the systems studied, the surface occupied by one molecule of TU, pT and mT (Sin - 1/Fs) was larger t h a n surface calculated from the molecular dimensions of TU, pT or mT molecule. The variances obtained between the theoretical and experimental values of Fs may be a result of residual presence of the molecules of the inhibitor and the remaining molecules of w a t e r in the adsorption layer with the m a x i m u m coverage of the surface of the electrode with adsorbants [29, 40]. Figure 14 presents the linear test of the F r u m k i n isotherm for P E G - TU systems.

878

-9.0

-9.0

m

m

PEG 400

PEG 10000

-8.0

-8.0

-7.0

-7.0

-6.0

-6.0 +3

-5.0

4.0

!

1

I

0.2

0.0

I

I

0.4

I

I

0.6

~21

-5"0

I

4.0

0.8

I 0.0

I 0.2

I

0.4

I

I

0.6

I

0.8

Figure 14. Linear test of the Frumkin isotherm for 10.4 M PEG - TU systems, the electrode charges (CYM/102C 9m -2) indicated by each line [22].

The slope of the lines p r e s e n t e d in Figure 14 indicates t h a t the values of the i n t e r a c t i o n c o n s t a n t A in the solution containing 10 .4 M P E G 400 c h a n g e with the decrease in ~i~ f r o m - 1 . 8 t o - 2 . 9 . Figure 15 p r e s e n t s v a r i a t i o n in the A p a r a m e t e r depending on the electrode charge for 10-4M and 10-4M P E G 400 and P E G 10000.

7

7 0

PEG 10000

_

6

5 4 ---o.

~ <>-

3 -O --O

--r

2

b I

I

I

I

I

-5 -4 -3 -2 -1

I

I

I

I

0

1

2

3

-5

I

I

I

I

I

I

I

I

-4

-3

-2

-1

0

1

2

3

-2

-2

O'M/ 10 C ' m

Figure 15. Variation of the interaction parameter A due to surface charge density for TU in the presence of: a) 10-4M, b) 5 9 10-4M PEG 400 or PEG 10000 [22].

879 Based on the values of the A p a r a m e t e r obtained, it needs concluding t h a t the repulsive interaction between the adsorbed TU molecules is lower in the presence of PEG 10000, as compared to PEG 400. The values of A are also lower t h a n those obtained in w a t e r solutions [38,40] (the comparison of the A p a r a m e t e r was conducted based on its absolute values). It needs stressing t h a t in the solution of 0.5M Na2SO4 [40], along with the increase of the positive charge of the electrode, the repulsive interaction among the adsorbed TU molecules decreases. A similar effect is observed in the presence of PEG 400 and PEG 10000. It is most likely connected with the screening action of C10-4 ions being adsorbed in the vicinity of the positive pole of the TU polar molecule whose adsorption increases with the increase of the positive charge of the electrode. The value of AG ~ was determined using the extrapolation of the straight lines in Figure 14 to the value of 0 = 0. Figure 16 presents the dependencies between AG ~ and the electrode charge for P E G - TU systems.

22

22 9

E

PEG 400

21 _

,~ ' ~ / ~

PEG 10000 / ~

21 bn

=D <~ 20

/o/

20 19

b

|

_

18

17 16 -

15 I

1

I

I

I

I

-6 -5 -4 -3 -2 -1 0

I

.I

1 2

I

3

I

-4

-3

-2

-1

0

1

2 -2

O'M / l O

3 -2

C'm

Figure 16. Variation of the free energy of adsorption AG Odue to surface charge density for TU in the presence of: a) 10-4M, b) 5 9 10-4M PEG 400 or PEG 10000 [22].

Linear dependencies of AG ~ on O'M indicate a preferential contribution of stable dipoles in the generation of free energy of adsorption, similarly to the process taking place in the water solution of TU [34]. It needs stressing t h a t in the presence of PEG 400, changes in the value of

AG ~ against the electrode

charge are much smaller in comparison to the respective changes of AG ~ in the presence of PEG 10000. The result indicates a greater impact of PEG 10000 on

880 the TU absorption compared with PEG 400, in particular in its higher concentration. Figure 17 shows linear tests of the F r u m k i n isotherm for P E G - pT systems.

-9.0 D

--9.0 EG 400

f

PEG 10000

-8.5 -8.0

,.....,

-8.0 --7.5 --7.0 --7.0

+4 D

+2 -6.5 0.1

I

I

0.2

I

I

0.3

I

I 0.4

-6.0_ 0.0

-

1

I

0.1

I

I

0.2

I

_

I

0.3

I

0.4

|

0.5

Figure 17. Linear test of the Frumkin isotherm for 10-4M P E G - pT systems, the electrode charges ((SM/10-2C 9m 2) indicated by each line [23].

As indicated by Figure 17, the value of p a r a m e t e r A for pT in the presence of 104M PEG 400 or 10 .4 M PEG 10000 does not depend on the charge of the electrode and ranges from-4.5 a n d - 3 . 5 respectively, while in the presence of 5 9 10-4M PEG 400 or 5 9 10 .4 M PEG 10000, the values are r e s p e c t i v e l y - 3 . 6 and -3.1. The results obtained indicate that in the presence of 10-4M PEG 400 the repulsive interaction among pT molecules is slightly stronger. In the remaining systems the interactions are similar to the respective interactions in 1 M NaC104 with no PEG added. The values of AG ~ for pT both in the presence of 10nM PEG 400 and 5 9 10 .4 M PEG 400 do not depend on the charge of the electrode and a r e - 2 4 . 6 k J " tool -1 and - 2 3. 7kJ " mol 1. In PEG 1 0 0 0 0 - pT systems, the values of AG ~ for pT depend linearly on the charge of the electrode, similarly to the PEG TU systems. The effect of the increased value of AG O for pT in the presence of PEG 400 and PEG 10000, yet only for ~M > 0, compared to the appropriate values obtained in the basic electrolyte 1M NaC104 comes as a surprise. The effect is most likely the result of fewer water molecules being present in the adsorption layer in the presence of adsorbed PEG molecules which to a more limited extent inhibit pT adsorption compared to water molecules. The linear tests of the F r u m k i n isotherm for PEG - mT systems are presented in Figure 18.

881

PEG 400

PEG 10000

-9

-8

+4

-2 -4 _ I O. 1

I 0.2

0.3

I+2 ' 0.4

'l 0.5

I

0.0

I

0.2

I

I

0.4

I

_ 0+2 I

0.6

Figure 18. Linear test of the Frumkin isotherm for 10.4 M PEG - mT systems, the electrode charges (CYM/10.2 C'm "2) indicated by each line [36].

As indicated by Figure 18, the adsorption properties of mT in the presence of PEG 400 and PEG 10000 are f u n d a m e n t a l l y different. In the presence of 10-4M PEG 400, the value of the A p a r a m e t e r changes along with the increase in the electrode charge f r o m - 4 . 7 t o - 5 . 9 . , while in the presence of 5 9 10-4M PEG 400, the repellent interaction between the adsorbed molecules of mT is more limited and the value of the constant A changes ranging from -2.9 to -3,.4 respectively. In the presence of PEG 10000, the value of the A p a r a m e t e r does not depend on the charge of the electrode and is similar for both c o n c e n t r a t i o n s , - 1 . 9 a n d - 2 . 2 for 10 .4 M PEG 10000 and 5 9 10 .4 M PEG 10000 respectively. The comparison of the value of the impact of the constant A demonstrates t h a t only in the presence of 10-4M PEG 400, the repulsive interaction among the adsorbed mT molecules is stronger compared to the same interaction in the solution of 1M NaC104. In the r e m a i n i n g P E G - mT systems studied, the interaction is weaker. The values of AG ~ determined for mT in the presence of PEG 400 do not depend on the charge of the electrode and indicate a slightly stronger adsorption of mT in the presence of PEG 400 compared to the adsorption in the solution of 1M NaC104. In the presence of PEG 10000, similarly to pT, there is also a linear relationship between the value of the AG ~ and the charge of the electrode. The value of AG O for pT and mT increases along with the increase of the positive charge of the electrode, with a stronger increase found in the presence of 5"10 .4 M PEG 10000 compared to the increase of AG ~ observed in the presence of 10 .4 M PEG 10000. A slightly higher adsorption of mT compared to pT from the 1M NaC104 solution, particularly visible at positive charges of the electrode, finds no

882 c o n f i r m a t i o n in s y s t e m s c o n t a i n i n g P E G in which the value of AG ~ for m T a n d pT are comparable. T h e r e f o r e s t r u c t u r a l differences of toluidine molecules h a v e no i m p a c t u p o n t h e i r a d s o r p t i o n in the p r e s e n c e of P E G polymers. F i g u r e 19 p r e s e n t s a l i n e a r test of the F r u m k i n i s o t h e r m for B U - T U a n d B U - m T systems.

~ (~-7.0

-10.0 \,Q\-. 9 A ",," ~ ~ 9149

= -6.0

-9.0

-5.0

-4.0

. .o

!i

-3.0

0.0

0.2

0.4

0.6

0.8

0.0

|

0.2

0.4

0.6

0.8 |

Figure 19. Linear test of the Frumkin isotherm for systems: a) 0.55 M BU + TU, b) 0.55 M BU + roT, the electrode charges (oM/10 -2 C'm ~ indicated by each line [37].

As i n d i c a t e d by the above Figure, the values of p a r a m e t e r A d e m o n s t r a t e a clear d e p e n d e n c e on the electrode charge only for the B U - TU systems. The effect is i l l u s t r a t e d in F i g u r e 20. 10 8

6 -o---

o

4 2 0 -2

I

I

I

I

0

2

4

6 -2

oM/lO C'm

-2

Figure 20. Variation of the interaction parameter A due to surface change density for TU in the presence of: a) 0.55 M BU, b) 0.88 M BU [37].

883

Figure 20 shows t h a t in general the repulsive interaction among TU molecules decreases with the decrease of the BU concentration and with the increase of the positive charge of the electrode. An exceptional situation is observed for ~ M ---- -t- 0.01C.m-2, at which the value of the p a r a m e t e r A effectively does not depend on the concentration of BU. The m a x i m u m change of the p a r a m e t e r for both concentrations of BU is very similar which can be an indication of similar changes occurring in the TU molecule orientation in both cases. It needs stressing t h a t the values of p a r a m e t e r A for TU in the presence of BU are on the whole either comparable with those obtained in the solution of 1M NaC104, or lower. Exceptionally higher values of the p a r a m e t e r have been obtained in the presence of 0.88 M BU for (~M < 0. The area of surface charges of (YM < 0 in the BU TU system is equivalent to the m a x i m u m adsorption of the substances used and it may be connected with the relatively higher values of p a r a m e t e r A at these charges. Not insignificant is also the possibility of changes in the butanol cluster structure and concentration in the bulk stage which can effect the surface stage as well. The value of the interaction constant for toluidine isomers in the presence of 0.55 M BU o f - 1 . 5 3 for mT a n d - 3 . 3 for pT does not depend on the charge of the electrode. In the presence of 0.44 M BU, the values of p a r a m e t e r A for these isomers indicate weaker repulsive interaction and to a slight degree depend on the charge of the electrode. It needs stressing t h a t in each solution containing BU the values of p a r a m e t e r A are lower in terms of the absolute values t h a n the respective values obtained in the solution of 1M NaC104, with such changes being more a p p a r e n t in the case of mT, as compared to pT. The value of free adsorption energy for 0 = 0, depending on the electrode charge in B U TU and B U toluidine isomer systems was presented in Figure 21. 30 7...., 0

C

26

_

a

b

3~ 22 d

18

_ a

14

I -2

0

2

4

J 6

O"M / 10 .2 C" m "2

Figure 21. Variation of the free energy of adsorption AG 0 due to surface charge density for: a) TU, b) mT, c) pT in the presence of 0.55 M BU and for d) TU in the presence of 0.88 M BU [37].

884 The analysis of the above Figure indicates significantly stronger adsorption of toluidine isomers compared to TU despite much lower concentration of toluidine used in the experiment. The linear nature of the relationship between

AG 0= f((~M) for toluidine isomers is a result of the chemical interaction of the aromatic ring ~ electrons of toluidine with mercury, resulting from partial transfer of the charge [41,42]. Absence of linearity from the relationship presented in Figure 21 for TU in the vicinity of the pzc may be connected with a more physical interaction between TU molecules and the surface of the electrode, as against the adsorption at positive charges. The above effect taking part within the area of the water capacity hump undoubtedly is also related to the relatively loosest structure of the surface water and therefore, the existence of a high dipole polarisability. It should also be noted that in all solutions examined containing BU, the increase in the concentration of BU results in an increase in the value of AG ~ both for TU and toluidine isomers. The effects contrast with the results obtained in systems containing PEG, and in particular PEG 10000. Easier adsorption of the substances under examination in the presence of higher concentrations of BU is most likely connected with increased degree of order in the adsorbed molecules on the surface of the electrode, resulting from selfassociation of BU molecules and also, as indicated earlier, lower number of water molecules displaced by the adsorbed organic molecules.

2.4. Virial a d s o r p t i o n i s o t h e r m s The values of Fs obtained for TU, pT and mT are in the majority of systems at variance with the theoretical values, and therefore to describe the adsorption of these substances, the virial isotherm was applied: J3c = F.exp2BF

(4)

where: B is the second virial coefficient. A linear test of the virial isotherm for PEG - TU systems is presented in Figure 22. Using a linear test of the virial isotherm, the value of the second virial coefficient B was determined based on the slope of the lines, while the value of AG 0 was determined by extrapolation of the lines to the value of F'- 0 in the standard state of 1 mol'dm -3 in the bulk of solution and 1 molecule 9 cm -2 on the surface of the electrode. The value of the second virial coefficient B at pzc in the presence of PEG are slightly lower compared to the value of B = 1.2 nm 2 9 molecule I, obtained in the solution without PEG [34], yet in the presence of PEG I0000, the values are slightly higher. The divergent changes in the value of the B parameter in systems containing PEG 400 or PEG I0000 are connected with the fact that the parameter constitutes the resultant of the inter-molecular repulsive interaction

885 a n d molecule size [43]. In the c h a n g e s of the AG ~ relative to the m o l a r m a s s of PEG, its c o n c e n t r a t i o n as well as the electrode charge, t h e r e is a n a n a l o g y to the c h a n g e s in the r e l e v a n t v a l u e s of AG O o b t a i n e d for the F r u m k i n isotherm. Similar effects were o b s e r v e d in o t h e r s y s t e m s studied.

18

18 -

o~

17

17

16

16 +3 +2 ' ~ ~1

15

14

I 1

0

I 2

,""i-3 3

PEG 10000

15

~

14

I -3 1

0

+2

I0 2 18

-2

F ' v u / 1 0 molec, m

Figure 22. Linear test of the virial isotherm for TU in the presence of 10 -4 M PEG, the electrode charges (CYM/10.2 C'm -2) indicated by each line [22].

Table 1 The values of p a r a m e t e r A b a s e d on the F r u m k i n i s o t h e r m a n d p a r a m e t e r B b a s e d on the virial i s o t h e r m as well as the calculated v a l u e s of p a r a m e t e r B according to P a y n e for the s y s t e m as follows: T U - 10 .4 M P E G B/nm2"molecule -1, ~M/IO-2 C.m-2

A

B (virial isotherm)

B - (2A+l)/2Fs

(JM

P E G 400

P E G 10000 P E G 400

P E G 10000 P E G 400

P E G 10000

+2

4.72

3.60

0.85

1.31

0.84

1.22

+1

5.08

3.80

0.88

1.43

0.90

1.28

0

5.28

4.04

0.91

1.44

0.93

1.34

-1

5.60

4.72

0.98

1.52

0.98

1.52

-2

6.04

4.84

1.00

1.58

1.04

1.55

886 Based on the relationship contributed by Payne [44]: B = (2A +l)/2F s

(5)

the value of B was calculated for selected systems. A satisfactory conformity of the values of B so calculated and the values of B determined based on the virial isotherm was found. Table 1 presents an example of the values of A and B p a r a m e t e r s calculated for the P E G - TU systems.

2.5. E l e c t r o s t a t i c

parameters

of the inner layer

An insight into the potential drop changes in the inner layer (I)M2 at a constant charge caused by adsorption offers information on the structure of the double layer. The changes are the resultant of the contribution of the free charges and oriented dipoles. Experimental separation of the effects is in principle impossible [45]. According to the electrostatic model of Parsons [34], the potential ~M-2 is a sum of two components depending on the surface density of the charge: (I)M -2 _ 4~x----L1cyM + 4~pp F'p

(6)

where: pp is the dipole m o m e n t of the isolated molecule of the accelerating agent: for TU, ~ = 16.31"10 .30 C'm, for pT, ~ = 4.43"10 .30 C'm and for mT ~ = 4.76"10 ~0 C'm. The values of the dipole moments in the inner layer undergo usually certain changes caused by the field of the electrode and additional interactions with the adjacent dipoles. The analysis of equation (6) ignores these effects. Other variables used in equation (6) or arising out of it include: Xl - the inner layer thickness; ~ i - electrical permittivity of the inner part of the double layer; K i - the integral capacity of the inner layer. The value of potential (I)M-2 was calculated based on the relationship: (I)M-2 = E - E z - (I)2-S where E stands for the potential equivalent to the given value of F'p and CyM and Ez stands for the potential of the zero charge for the solution which does not contain the accelerating agent, while the drop of the potential in the diffusion layer (I)2.s was calculated based on the theory of G o u y - C h a p m a n [46]. The dependence of the value of (I)M-2 on F'TU at a constant surface charge density and in the presence of 10 .4 M PEG is presented in Figure 23. The relationships presented in the above Figure are linear in nature, similarly to the results obtained in other systems containing TU [47, 48]. The linearity of the dependence of ~M-2 on F'TU is additionally confirmed by the congruency of the adsorption isotherms described relative to the charge. An analysis of these relationships was conducted using the method applied previously [48]. The p a r a m e t e r s obtained which describe the inner layer are presented in Table 2.

887 PEG 400 -0.4

PEG 10000 -3

-2

-1

-0.4 - 3 - 2 -1 0 +1 +2

o

~e

1 +2

-0.3

-0.3

+3

-0.2

-0.2

-0.1

-0.1

0.0

0.0

0.1

0.1 0

1

2

3

0-

1

2 -18

10

-2

F'wu/molec" m

Figure 23. Potential drop across the inner layer q)M-2 as a function of the quantity of TU adsorbed at constant electrode charges (CYM/10.2 C'm -2) in the presence of 10.4 M PEG [22].

Table 2 I n n e r l a y e r p r o p e r t i e s for TU a d s o r b e d at the m e r c u r y / w a t e r - N a C 1 0 4 -

PEG

m i x t u r e interface 102 ~M/C'm -2, xl/nm, 10-~SF'Tu/molecule-m-2 102K iF ' /F'm-2122] P EG 400 (~M CPEG=COnSt

Ei

10 -4 M -3 9.3 - 2 10.2 - 1 11.0 0 12.4 +1 13.3 CPEG----5" 10 -4 M -3 7.1 -2 7.5 -1 8.4 0 9.4 +1 10.0

F'TU = 0 Ki

P EG 10000 F'TU - 1

xl

Ki

Xl

26.3 0.31 30.3 0.30 32.3 0.30 3 - 8 . 5 0.29 47.6 0.28

17.9 20.2 22.2 26.3 31.3

0.46 0.45 0.44 0.42 0.38

35.7 38.5 45.5 50.0 58.8

18.9 21.3 23.8 27.8 32.3

0.33 0.31 0.31 0.30 0.27

Ei

F ' T U -- 0

F ' T U -- 1

Ki

x1

Ki

Xl

5.2 6.2 6.8 7.5 8.0

20.8 24.4 27.0 30.3 38.5

0.22 0.22 0.22 0.22 0.18

9.8 13.3 15.6 18.5 20.8

0.47 0.44 0.39 0.36 0.34

4.3 5.1 6.1 6.7 7.8

23.3 28.6 33.3 35.7 42.0

0.16 0.16 0.16 0.16 0.16

8.9 11.4 14.7 16.4 19.2

0.43 0.40 0.38 0.36 0.36

C P E G ----

0.18 0.17 0.18 0.17 0.15

The v a l u e s of Ei for TU o b t a i n e d in the p r e s e n c e of P E G are in g e n e r a l lower c o m p a r e d to the v a l u e of Ei = 11.4 o b t a i n e d for TU in w a t e r solution at aM = 0 [34]. The fact t h a t the electrical p e r m i t t i v i t y of the i n n e r l a y e r i n c r e a s e s along

888 with the increase of the positive charge of the electrode, may prove the existence of free PEG molecules in the inner layer, subject to partial rotation. The effects are in conformity with changes of the constant value of the interaction of A determined on the basis of the F r u m k i n isotherm. The drop in the integral i with the increase of TU concentration is connected with the capacity of K F, increase of the thickness of the inner layer Xl. Small values of Xl for F~,U = 0 indicate fiat orientation of PEG molecules adsorbed on the surface of the mercury electrode, as previously suggested [31]. The values of K iF, determined in the presence of PEG 10000 are lower t h a n the respective values determined in the presence of PEG 400. The effect is due to lower values of ~i in the presence of PEG 10000, as compared to PEG 400. The results obtained indicate a more condensed structure of the inner layer in the presence of PEG 10000. Figures 24 and 25 present changes in the value of (I)M-2 relative to F'p for toluidine isomers in the presence of PEG.

>

-0.4

-0.4

m

PEG 10000 -0.3 _

-4

-0.2

-0.2

-2

-0. l

+2

0

"-a. -2

-0.1 _

-4

-0.3

~

j_oj_----o-o-~-__~__>___q0 0.0

0.0

0.1 0.4

I

1

I

I

I

0.8

1.2

1.6

2.0

2.4

0.1

I 0.5

1 1.0

1

[

1.5

2.0 18

2.5 -2

F ' ~ j / 1 0 molec, m

Figure 24. Potential drop across the inner layer (I)M-2 as a function of the quantity of pT adsorbed at constant electrode charges (C~M/10.2 C'm -2) in the presence of 10 -4 M PEG [23].

The relationships presented in the above Figures are linear in n a t u r e effectively only in the presence of PEG 10000. Unfortunately, p a r a m e t e r s of the inner layer calculated in these systems vary significantly from rational values. The general absence of linearity in the relationship between (I)M-2 and F'p in the presence of PEG 400 in the case of mT may be due to the changed orientation of mT molecules on the surface of mercury, which is also indicated by the values of

889

>

e

-0.6

-0.20

P E G 400

-2

-0.16 -0.4 -0.12 -0.2

..- .::-

_

0

~~----__~ -0.08 "<--2 _ ~

0.0

-

- - ~ 2 , - - - - - - - - - ~ ~ _ . ~

k

,

--~=-

0

-0.04

0.2 0.4

018

1 !.2

".6

I 2.0

~ 2.4

000 0.5

1.0

1.5

2.0

F',,,v / 1018molec-

2.5 m -2

Figure 25. Potential drop across the inner layer (I)M-2 as a function of the quantity of mT adsorbed at constant electrode charges (cyv/10.2 C'm -2) in the presence of 10-4 M PEG [36].

the interaction c o n s t a n t A d e t e r m i n e d based on the F r u m k i n i s o t h e r m as well as the value of the second virial coefficient obtained b a s e d on the virial isotherm. Such a conclusion c a n n o t be d r a w n for pT. O t h e r r e a s o n s for the absence of linearity in the relationship b e t w e e n ~M-.9 and F'p m a y include: p a r t i a l t r a n s f e r of the charge from the pT molecule to the m e r c u r y electrode t a k i n g place during the adsorption process [49] or competitive adsorption p r e s e n t in the solution of molecules and ions [50]. The dependence of the value of (I)M-2 on F'p for TU or mT in the presence of BU h a s been p r e s e n t e d in Figure 26. The linear dependence of r o n F'p p r e s e n t e d in the above Figures served as a basis for the calculation of the p a r a m e t e r s of the inner layer. The values of ~i for TU at pzc are slightly lower t h a n the value of ~i for TU in a solution without BU, as well as slightly lower t h a n the value of ~i for TU in m e t h a n o l [51], while the values of ~i for pT a n d mT in the presence of BU are similar to the values of ~: d e t e r m i n e d for TU in methanol. The values of the integral capacity K i at pzc in all s y s t e m s with BU are close to 0.306 F-m -2 obtained for TU in a w a t e r solution [34]. The values of Xl r a n g i n g from 0.10 to 0.24 n m are m u c h lower c o m p a r e d to the respective value of Xl for TU in a w a t e r solution. The r e a s o n for the discrepancy is most likely the simplifications i n h e r e n t in the electrostatic model of the inner layer [34], referred to earlier. Therefore, the analysis of electrostatic p a r a m e t e r s in this case m a y be t r e a t e d as s o m e w h a t approximated.

890

-0.4

-17)(o/+/~~7

,-q

-0.3

-10

+3

e -0.3

-0.2

,,

+1 +2 +3

-0.2 -0.1 -0.1 a

0.0

0.0

0.1

0.0

I

I

I

I

0.5

1.0

1.5

2.0

0.1

0.0

I

I

I

I

0.5

1.0

1.5

2.0

18

-2

F' e/10 molec" m

Figure 26. Potential drop across the inner layer (I)a2 as a function of the quantity of: a) TU, b) mT adsorbed at constant electrode charges (<SM/10.2 C'm -2) in the presence of 0.55 M BU [37].

3. KINETICS OF Zn(II) ION REDUCTION IN SYSTEMS S T U D I E D The results of studies presented earlier provide us with information on mixed structures of adsorption layers in terms of the potential in which strong adsorption of the studied organic substances occurs. The expansion of the scope of research to include the study of the range of potentials in which a significantly weaker adsorption takes place proved feasible by examining the kinetics of Zn(II) ion reduction acting as "the adsorption probe". The concentrations used in this Section refer to the bulk concentrations, since adsorption m e a s u r e m e n t s do not allow determining surface concentrations of substances in systems used for this range of potentials so very distant from the pzc. The value of ksapp, the apparent standard rate constant for the Zn(II) ion reduction was determined in the majority of systems using the impedance spectrum method with the formal potential E~. The impedance spectrum of the electrode was determined for various frequencies from 100 to 25000Hz. The ohmic resistance of the basic electrolyte was determined as the actual component of impedance at a frequency of 10kHz and a potential distant from the F a r a d a y region. The activation resistance RA was calculated for the potential E~ using the relationship: Z' = f(mZ") or Z" = f(Z') [52, 53], where Z' is the real component and Z" an imaginary component of cell impedance. calculated using the following relationship:

The value of

k app was

891 kapp =

RT n2F2cRA

(7)

where c is the concentration of the depolariser. In certain systems, where Zn(II) ion reduction was clearly an irreversible process, the m e a s u r e m e n t of the impedance spectrum failed to produce satisfactory results.

Therefore in these systems the values of k app were

determined using cyclic v o l t a m m e t r y curves based on the irreversible process theory [54]. This same method was also used to determine the value of the reversible half wave potentials E~/2 . Studies of the electroreduction of Zn(II) ions in the presence of TU in 1M NaC104 showed t h a t the value of E1~/2 change along with the increase in its concentration f r o m - 0 . 9 8 5 V to -0.975V which indicates the absence of stable Zn(II) - TU complexes in the solution. In a solution containing PEG 400 only, the value of E1~/2 shifts maximally towards the value o f - l . 0 5 1 V , while in the presence of PEG 10000 the same value is-1.095V. The introduction of increasing quantities of TU to the solution containing PEG results in the shift of the value of El/2 in the direction of positive potentials and for the m a x i m u m TU concentration of 0.55M, the values of E~/2, regardless of the concentration and the molar mass of PEG, are very similar to those obtained in a solution which did not contain PEG. Figure 27 presents a logarithmic dependence of ks pp of Zn(II) ion reduction on TU and PEG 400 concentrations. It needs stressing t h a t the value of k app obtained in the solution without TU depends to a certain extent on the concentration of PEG 400 and the m i n i m u m value of k app - 2.63"10-Scm's -1. A similar dependence on the concentration of PEG 400 is also found in the presence of TU (lines b and c). In the presence of PEG 10000, the rate constant of Zn(II) reduction does not depend on its concentration and is 6.00-10-6era's-I; a similar situation is found in the presence of TU (line d). Based on the intersection of b, c and d lines with the dashed line, the relationship of TU: PEG concentrations was determined in which a compensation of the inhibiting and accelerating effect takes place. In the case of PEG 400, the relationship between concentrations is 4000 to 220, while for PEG 10000 from 27000 to 540. A decrease of the relationship with an increasing PEG concentration is arguably due to an easier access of TU molecules to the surface of the electrode in a higher PEG concentration which results from a better ordered and less h y d r a t e d structure of the adsorption layer. The TU:PEG concentration relationships presented referring to the compensation of the inhibiting and accelerating effect confirm in an obvious m a n n e r the stronger adsorption of P E G 10000 compared to PEG 400 also at potentials distant from

892

a 9

b 7X

C

-2

-3

-4

I

-3

-2

-1

0 log CT~

Figure 27. Plots of log k app for the (Hg)Zn(5"10-3M)/Zn(II) (5"10-3M) couple vs. log CTU in the presence PEG: a) cpE~ = 0M, b) 104M PEG 400, c) 5-10-4M PEG 400, d) 5-10-4M PEG 10000. The dashed line denotes k app = 3.31.10 -3cm.s-I for the Zn(II) in 1M NaC104 [22].

the pzc.

Based on the slope of the Alogk app/Alog cTu, the electrode acceleration

process coefficient m a y be d e t e r m i n e d which for TU in 1M NaC104 (line a) is 1.00, while in the presence of P E G 400 or PEG 10000 it increases to 1.67 and 1.47 respectively. It appears t h a t the increased efficiency of the acceleration of the electrode process in the presence of the inhibitor compared to the 1M NaC104 solution is also due to the reduced n u m b e r of w a t e r molecules on the surface of the electrode and most likely the increased lability of the adsorbed TU molecules compared to their behaviour in the pure basic electrolyte. In order to d e m o n s t r a t e the impact of the structure of PEG 400 and PEG 10000 polymers , the values of k app of the Zn (II) ion reduction in a solution containing TU + 2.5-10 .3 M PEG 400 were determined. The selected concentration of P E G 400 is equivalent to the n u m b e r of mers -CH2CH2-O- contained in the

893 10 .4 M PEG 10000 solution. The results indicate a limited impact of the PEG 400 concentration on the kinetics of the reduction of the pilot ion, however, the values of ksapp in the presence of PEG 10000 are in general ten times lower t h a n they are in the presence of PEG 400. The different inhibiting effect of both polymers studied is undoubtedly due to their different structures: PEG 400 molecules are zig-zag shaped, while the PEG 10000 molecules are m e a n d e r shaped [24,25] which without doubt has some impact upon their form as the adsorbate. Figure 28 presents a logarithmic dependency of k app of Zn(II) ion reduction on the pT concentration in P E G - pT systems.

9

-2

_

b

c -3 _

d

-4

-6 -3-

I

I

-2

-1 log %,,.

Figure 28. Plots of log k app for the (Hg) Zn(5 10-3M)/Zn(II) (5" 10-3M) couple vs. log CpT in the presence PEG: a) CpEG= 0M, b) 10-4M PEG 400, c) 5"10-4M PEG 400, d) 5"10-4M PEG 10000. The dashed line denotes ksapp = 3.31.10 3cm's -1 for the Zn(II) in 1M NaC104 [23].

Similar dependencies were determined for mT; the slightly stronger accelerating properties of pT compared to mT are reflected also in solutions containing PEG. The dependencies presented referring to solutions containing

894 PEG are not linear compared to the respective dependencies determined for TU (Figure 27). The similarity in the changes of the value of ksapp in the presence of TU and toluidine isomers lies in the fact t h a t the values depend on the concentration of PEG 400 and not on the concentration of PEG 10000. The effect of compensation of inhibition and acceleration of Zn(II) ion electrolytic reduction occurs in the presence of toluidine isomers at far lower concentration relationships then in the presence of TU which seems to be due to the stronger adsorption of such isomers on the surface of mercury. This s t a t e m e n t is of course based on the results of adsorption m e a s u r e m e n t s conducted in the vicinity of the pzc, however with a great degree of caution it could be applied to the Zn(II) ion reduction potentials, a p p r o x i m a t e l y - 1 . 0 V which lie in the toluidine isomer adsorption region, since the cathodic m a x i m u m on differential capacity curves obtained for toluidine isomers in 1M NaC104 solution occurs at the potential of approximately -1.3V. The slopes describing the efficiency of the acceleration of Zn(II) ion reduction in 1M NaC104 determined based on the straight-line sections of the curves presented in Figure 28 are 0.64 for pT, while in the presence of both PEG 400 and PEG 10000 they are 2.44. The values for mT are 0.48 and 1.48 respectively. The effect described which was obtained in P E G - toluidine isomer systems is more clear t h a n in P E G - TU systems. Figures 29 and 30 present a logarithmic dependence of the value of ksapp for reduction of Zn(II) ions on the concentration of BU, TU and pT in BU + TU and BU + pT systems. 0.0 e~

-1.5

9

*

a .

.

.

.

.

.

.

.

.

-3.0

-4.5 -2.4

.

.

.

.

.

.

.

.

*

-2.0

-1.6

-1.2

.

.

.

.

.

9

-0.8

-0.4

0.0

log C-ru(Bu~ 3

Figure 29. Plots of log k app for the (Hg) Zn(5.10 M)/Zn(II)(5.10-SM) couple vs. log cvu in the presence BU: a) cBu = 0M, b) 0.11M, c) 0.33M, d) 0.55M, e) 0.88M or vs. log cmj when cvu = 0(f). The dashed line denotes ks = 3.31.103cm.s l for the Zn(II) in 1M NaC104 [33].

895 0.0 0

-1.0

I

-2.0

-3.0 -2.8

-2.4

-2.0

-1.6

-1.2 log CpT(BU)

Figure 30. Plots of log k app for the (Hg) Zn(5" 10-3M)/Zn(II)(5 910-3M) couple vs. log CpT in the presence BU" a) CBU -- 0M, b) 0.11M, c) 0.44M, d) 0.55M. The dashed line denotes ks = 3.31" 103cm-s -z for the Zn(II) in 1M NaC104 [32].

The increase in the concentration of BU to 0.88M results in a reduced value of ksapp to 7.4.10 .5 cm-s -1. The effect of compensation of inhibition and acceleration of Zn(II) ion electroreduction in T U BU systems occurs in a constant relationship between concentrations of these substances at approximately 0.08. The efficiency of the acceleration of this process in the presence of BU is similar to t h a t observed in the presence of PEG. Stronger accelerating properties of Zn(II) ion electroreduction in 1M NaC104 by pT compared to mT occur also in all systems containing BU. The effect of compensation of the process of inhibition and acceleration occurs in the systems studied where the relationship between concentrations of pT:BU ~ 0.012 and of mT:BU ~ 0.029. Also in these systems, the efficiency of acceleration of the Zn(II) ion electrolytic reduction by pT and mT is higher in the presence of BU t h a n in the pure 1M NaC104 solution. The Zn(II) ion reduction takes place in the region of adsorption of toluidine isomers. As indicated by adsorption m e a s u r e m e n t s , at sufficiently negative values of ~M, the adsorption of these isomers both in the 1M NaC104 solution and in the presence of the inhibitors used should be comparable. Therefore the greater ability to accelerate Zn(II) ion reduction by pT as compared to mT should be more the result of differences in the structure of weaker Zn(II) - toluidine complexes having their source in spherical effects. The slight shifts in the value

896 r for the reduction of Zn(II) ions in the 1M NaC104 solution which together of El/2 with an increase in the pT concentration change f r o m - 0 . 9 8 5 V to -0.995V and for mT from -0.985V to -1.003V m a y evidence the existence of weak complexes. The results of the m e a s u r e m e n t s of the kinetics of Zn(II) ion reduction indicate much weaker inhibiting properties of BU compared to PEG 400 and PEG 10000. The potential to eliminate the inhibiting influence of the organic substances studied on the electroreduction of Zn(II) ions was also d e m o n s t r a t e d by using NaI as a component of the basic electrolyte (x M NaI + (l-x)M NaC104). The choice of I- ions was prompted by their -known capability to accelerate electrolyte reduction of Cd(II) and Zn(II) ions, the strongest from among halide ions [55,56]. The logarithmic dependence of the value of ksPP for the reduction of Zn(II) ions on the concentration of I- ions in the presence of BU or PEG is presented in Figure 31.

-1.0 a tz~o

-2.0 c

-3.0 b -4.0

d

-5.0 -2.0

I

I

I

I

I

-1.6

-1.2

-0.8

-0.4

0.0 log c,-

Figure 31. Plots of log k app for the (Hg) Zn(5.10-3M)/Zn(II)(5 910-3M) couple vs. log ci_ in the following solutions" a) x M NaI + (1-x)M NaC104, b) with the addition of 0.55 MBU, c) with the addition of 10.4 M PEG 400, d) with the addition of 10.4 M PEG 10000. The dashed line denotes ksapp = 3.31.10 -3cm.sl for the Zn(II) in 1M NaC104.

The course of the dependence presented in Figure 31 indicates a far weaker power to accelerate Zn(II) ion reduction by I- ions as compared to TU or toluidine isotherms. As a consequence of that, both in B U - I- systems as well as PEG 10000 - I- even at the m a x i m u m concentration of I- ions of 0.8M, the effect of compensating the inhibition of the Zn(II) ion reduction process by the accelerated

897 influence of I- ions is missing. Also in the systems studied the increase of the efficiency (lines a and b, on the one hand, and lines a and d are almost parallel) is absent. This effect appears however in the P E G - I- system which indicates a slightly different adsorption of I- ions in the presence of PEG 400, as compared to BU and PEG 10000.

4.

CONCLUSIONS

The account of mixed properties of adsorption layers presented in the above Section is to a certain extent the result of a certain convention customarily adopted in this type of research. It applies in particular to the thermodynamic analysis of the systems studied. The choice of organic substances based on their opposing impact on the Zn(II) ion reduction kinetics proved purposeful, for in effect it allowed a significant expansion of information on mixed adsorption layers. The information derived from adsorption m e a s u r e m e n t s d e m o n s t r a t e d as follows: 9 The values of relative surface excesses of substances accelerating the Zn(II) electroreduction increase together with the increase of the positive charge of the electrode which is in a g r e e m e n t with the postulated course of the TU molecule adsorption oriented with the sulphur atom towards mercury and toluidine molecules adsorbing with the aromatic ring. In the presence of all inhibitors studied, the relative values of surface excesses most often decrease in the following order: F'mT > 1-"pT > F'TU and all this in much lower concentrations of toluidine (conditioned by its solubility in the solutions studied), compared to the concentration of TU. 9 The increase in the positive charge of the electrode in all systems studied results in the increase of the value of free adsorption energy AG 0 which may be connected according to Payne [57] with squeezing w a t e r molecules out of the surface of the electrode at aM > 0.

9

9

The linear dependence of AG o on aM accompanying mainly the ion adsorption arrived at in the majority of systems indicates t h a t the effects connected with the polarity of molecules (stable dipole moment) constitute the dominating impact of the field of the electrode on the adsorbate molecule. Larger area occupied by a single molecule calculated on the basis of the Fs value compared to the area determined based on the molecular size indicate flat positioning of molecules on the surface of the electrode and their m u t u a l repulsion, confirmed by relatively larger values of the virial coefficients. It can also result from residual presence of other molecules. Negligible differences in the adsorption p a r a m e t e r s determined for toluidine isomers both in the 1M NaC104 solution as well as in the presence of inhibitors indicate t h a t the location of amino groups in the aromatic ring has a limited impact upon the adsorption process. The stronger adsorption of

898 toluidine isomers compared to TU confirms an exceptionally strong impact of the aromatic ring upon the adsorption process [58]. 9 The values of the A parameter determined on the basis of the Frumkin isotherm indicate repulsive interaction between the adsorbed molecules of the accelerating agent. In all systems studied containing TU, the increase of the positive charge of the electrode was accompanied by the reduction in the value of the A parameter resulting from the penetration of C104 anions between the positively charged TU amino groups at ~M > 0. In solutions containing toluidine isomers, the value of the A parameter in general does not depend on the charge of the electrode which suggests stability of the orientation of the adsorbed toluidine molecules. 9 The linear dependence of the potential drop in the inner layer ~M-2 to F'p found in the majority of systems confirms the congruence of the determined adsorption isotherms in relation to the charge. The values of Xl arrived at confirmed the fiat positioning of BU and PEG molecules on the surface of the electrode postulated earlier. Relatively large values of electric permittivity of Ei for F'p = 0 (~i > 10) obtained in some systems indicate a considerable freedom of inhibitor molecule rotation in the field of the double electric layer. The results of the Zn(II) ion electroreduction measurement as an ion piloting the adsorption equilibrium in mixed adsorption layers at a potential distant from the pzc indicate significantly higher efficiency of toluidine isomers in eliminating the inhibiting impact of PEG as compared to TU. In systems containing BU, the relationship between the concentration of the accelerating agent and the inhibitor with a view to the compensation effect is weak, i.e. it does not depend on the concentration of BU, while in the presence of PEG, such relationship does occur. The results arrived are indicated a higher BU adsorption lability compared to PEG at Zn(II) ion reduction potentials which additionally enables the adsorption of the accelerating agent. It is worth noting that at such potentials, the surface concentrations of the low-molecule substances studied are undoubtedly small and occur in the area of the initial, approximately linear, course of the isotherm which explains the stability of the relationship between the surface concentration and the bulk concentration. The above relationship does not apply to high-molecular compounds. The examination of the Zn(II) ion reduction mechanism as an adsorption probe using other inhibitors will undoubtedly allow expanding the information available on the properties and structure of mixed adsorption layers at potentials distant from the pzc. The results of research presented include information which can be used to understand better the adsorption equilibrium generated in complex corrosion systems comprising inhibitors with a specific structure and properties. Effectiveness of toluidine isomers as potential inhibitors of corrosion should be greater than that of TU owing to stronger adsorption of toluidine isomers. This adsorption is conditioned by the m e t a l - aromatic ring ~ electrons which are the effects of a partial charge transfer between the adsorbate and metal causing on

899 the whole, fiat orientation of the adsorbed toluidine molecules. An additional favourable effect of using toluidine as the inhibitor of corrosion is the fact that this substance causes a local increase of pH value which does not occur while using TU. Comparison of h i g h - and low molecular inhibitors activity allows for making the statement that the adsorption layers formed by high molecular inhibitors, as a rule, are not tight which enables their penetration by various factors which can affect the protective function change. Formation of mixed adsorption layers created by the adsorbate molecules with chemical interactions between them (e.g. hydrogen bonds), undoubtedly causes the increase of the adsorption layer tightness. In metal corrosion taking place in natural conditions, besides basic processes occurring on the metal surface and leading to oxidation of metal atoms, and their possible transition into the aqueous phase an essential role is also played by the redox processes caused by active components of the solution and leading to formation of the mixed potential. After its establishment there is formed on the metal a surface layer of a complex structure which can be stimulated by adsorption processes which can be conditioned by both electrostatic and donor-acceptor interactions. These considerations lead to a general conclusion that reliable results of the investigations of organic inhibitor effects on metal corrosion can be expected in the model systems very close to natural conditions and the fundamental investigations enable better understanding of the nature of these complex dependences.

REFERENCES 1. 2. 3.

4. 5. 6. 7. 8.

9. I0.

D.Jonhua, S.Guangling, L.Haicao, C.Chunan, Zhongguo Fushi Yu Fanghu Xuebao, 16 (1996) 53. M.Metikos-Hukovic, R.Babic, Z.Grubac, S.Brinic, J.Appl.Electrochem., 26 (1996) 443. R.M.Souto, M.M.Laz, S.Gonzalez, J.Chem.Soc.Faraday Trans., 92 (1996) 2725. N.Rajendron, K.Ravichandran, S.Rajeswari, Anti-Corros.Methods Mater., 42 (1959) 9. Q.J.M.Slaiman, H.M.AI-Saaty, Ann.Univ.Ferrara, Sez.5 Suppl. 1995, I0 (8 th European Symposium on Corrosion Inhibitors, 1 (1995) 333. R.H.Rahim Abdel, B.Hassan Hanaa, M.W.Khalil, Materialwiss.Werkstofftech., 28 (1997) 198. M.S.Abdel-Aal, M.S.Morad, Z.A.Ahmed, Ann.Univ.Ferrara, Sez. 5 Suppl. 1995, 10 (8 th European Symposium on Corrosion Inhibitors 1 (1995) 343. A.A.El-Shafei, M.N.H.Moussa, A.A.EI-Far, J.Appl.Electrochem., 27 (1997) 1075. A.M.Ai-Mayouf, Corros.Prev.Control, 43 (1996) 68. A.A.Aksut, A.N.Onal, Bull.Electrochem., II (1995) 513.

900 11. R.T.Vashi, V.A.Champanerii, Trans.SEAST, 32 (1975) 5. 12. S.Muralidharan, B.Babu Ramesh, S.Iyer Venkatakrishna, S.Rengamani, J.Appl.Electrochem., 26 (1996) 291. 13. M.Elachouri, M.S.Haji, M.Salem, S.Kertit, J.Aride, R.Coudert, E.Essassi, Corrosion (Houston), 52 (1996) 103. 14. E.Khamis, H.A.Al-Lohedan, A.Al-Mayouf, Z.A.Issa, Materialwiss.Werkstofftech., 28 (1997) 46. 15. Wang Jianming, Cao Chunan, Lin Haichao, Wuli Huaxue Xuebao, 12 (1996) 1036. 16. E.Kalman, J.Lukovits, G.Palinkas, ACH-Models Chem., 132 (1995) 527. 17. T.Suzuki, H.Nishihara, K.Aramaoki, Corros.Sci., 38 (1996) 1223. 18. M.Th.Makhlouf, M.H.Wahdan, Polish J.Chem., 69 (1995) 1072. 19. K.Sykut, J.Saba, B.Marczewska, G.Dalmata, J.Electroanal.Chem., 178 (1984) 295. 20. G.Dalmata, J.Nieszporek, Polish J.Chem., 68 (1994) 2009. 21. A.Baars, J.W.J.Knapen, M.Sluyters-Rehbach, J.H.Sluyters, J.Electroanal.Chem., 368 (1994) 293. 22. J.Saba, Electrochim.Acta, 41 (1996) 297. 23. J.Saba, Monats.Chem., 128 (1997) I. 24. M.R6sch, Kolloid Z., 147 (1956) 80. 25. M.R6sch, Kolloid Z., 150 (1957) 153. 26. R.Parsons, Can.J.Chem., 59 (1981) 1898. 27. B.B.Damaskin, S.L.Diatkina, V.K.Venkatesan, Elektrochimija, 5 (1969) 524. 28. K.M.Joshi, S.J.Mahajan, M.R.Bapat, J.Electroanal.Chem., 54 (1974) 371. 29. M.R.Moncelli, M.L.Foresti, R.Guidelli, J.Electroanal.Chem., 295 (1990) 225. 30. A.F.Nesterenko, W.S.Burykina, A.J.Aprasjuchin, M.A.Loszkariew, Elektrochimija, 23 (1987) 425. 31. H.Jehring, Elektrosorptionsanalyse mit der Wechselstrompolarographie, Akademie Verlag, Berlin 1974. 32. J.Saba, Collect. Czech. Chem. Commun. , 60 (1995) 1457. 33. J.Saba, Electrochim.Acta, 39 (1994) 71 I. 34. R.Parsons, Proc.Roy.Soc., A 261 (1961) 79. 35. R.Parsons, Trans. Faraday Soc., 51 (1955) 158 I. 36. J.Saba, Gazz.Chim.ltal., 127 (1997) 53. 37. J.Saba, Collect. Czech. Chem. Commun. , 61 (1996)999. 38. R.Parsons, P.G.Symons, Trans.Faraday Soc., 64 (1968) 1077. 39. J.D.Garnish, R.Parsons, Trans. Faraday Soc., 63 (1967) 1754. 40. G.Pezzatini, M.R.Moncelli, R.Guidelli, J.Electroanal.Chem., 301 (1991) 227. 41. D.Rolle, J.W.Schultze, J.Electroanal.Chem., 229 (1987) 141. 42. S.Trasatti, Electrochim.Acta, 37 (1992) 2137. 43. S.Trasatti, J.Electroanal.Chem., 53 (1974) 335. 44. R.Payne, J.Chem.Phys., 42 (1965) 3371. 45. S.Trasatti, J.Electroanal.Chem., 64 (1975) 128. 46. D.C.Grahame, Chem.Rev., 41 (1947)441.

901 47. S.Minc, J.Jastrz@ska, M.Jurkiewicz-Herbich, J.Electroanal.Chem., 152 (1983) 223. 48. M.Jurkiewicz-Herbich, J.JastrzQbska, Pol.J.Chem., 58 (1984) 1125. 49. O.Ikeda, H.Jimbo, H.Tamura, J.Electroanal.Chem., 137 (1982) 127. 50. M.Jurkiewicz-Herbich, Pol.J.Chem., 66 (1992) 1695. 51. W.R.Fawcett, M.D.Mackey, J.Chem:Soc., Faraday Trans. I, 69 (1973) 634. 52. J.H.Sluyters, J.J.C.Oomen, Rec.Trav.Chim., 79 (1960) 1101. 53. J.H.Sluyters, Rec.Trav.Chim., 79 (1960) 1092. 54. R.S.Nicholson, J.Shain, Anal.Chem., 36 (1964) 706. 55. D.Zytner, Elektrochimija, 11 (1975) 1368. 56. R.Andreu, M.Sluyters-Rehbach, J.H.Sluyters, J.Electroanal.Chem., 171 (1984) 139. 57. R.Payne, J.Electroanal.Chem., 41 (1973) 145. 58. E.Dutkiewicz, J.Stuczyflska, Electrochim.Acta, 34 (1989) 1513.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

903

T h e b i o l o g i c a l s i g n i f i c a n c e of H a l o b a c t e r i a o n n u c l e a t i o n a n d s o d i u m chloride crystal growth A. LSpez-Cort@s and J. L. Ochoa The Center for Biological Research, P.O. Box 128, La Paz 23000, BCS, Mexico

ABSTRACT

The participation of halobacteria in halite formation has been previously considered by a number of authors. In fact, the release of bacterial nucleation factors has been suspected but never properly documented. Some studies show that certain chemicals are capable of modifying sodium chloride crystal habit, and this possibility is of the outmost importance for the correct management of salterns. For example, from over 500 compounds tested, only 12 have been found able to modify the crystal arrangement of halite. The halobacteria can also get entrapped in fluid inclussions within the crystal and retain viability for several years. Thus, they may play some role in contamination of salt preserved foods. Here we review the subject of halobacteria participation in sodium chloride crystal formation and present some results with regard the influence of cell surface layer (S-layer) components on crystal habit. As appears, the halobacteria may influence crystal growth rate and crystal habit, allowing the optimization of saltern management. 1. I N T R O D U C T I O N Cell morphology and cell surface components have been implicated in a wide variety of biological and adhesion phenomena, including mineral and organic crystal formation. Crystallization is used in industry to recover and purify many inorganic and organic materials, but very little work has been done to investigate the feasibility of bulk crystallization for the recovery and purification of proteins and mineral crystals [1]. Crystallization of large molecules, such as proteins, is less known than crystallization of mineral salts, and their mutual effects are poorly understood. For example, it is known that for each protein there appears to be a set of mineral substrates that promote nucleation of protein crystals at lower critical levels of supersaturation than required for spontaneous growth [2,3]. Judged et al. [1] studied the crystallization of ovalbumin in ammonium sulfate and observed that crystal growth could occur without nucleation at a relative supersaturation value of 20. The crystal size distribution was measured and the

904 crystal growth rate was found to be of second-order depending on ovalbumin supersaturation. There was only a slight effect of the ammonium sulfate concentration at pH above 4.9, while the effect of pH itself allowed a 10th-fold increase in the crystal growth rate constant in the range of 4.6-5.4. Under these conditions, the presence of other protein components did not effect the crystal growth rate of ovalbumin. In general, protein supersaturation can eliminate impurity effects [4] and temperature seems to favor crystallization. As a rule, slow crystallization of proteins is due primarily to impedance of the elementary act of entering the growth site, indicating a decisive role of entropy, not of energy barriers, in the crystallization of biological macromolecules. Pressure may actually inhibit crystal yields since, at equilibrium, two conformations of the protein (in the case of lysozyme) have been postulated: one capable of being incorporated into the nuclei of the crystal, and the other which is not [5]. The addition of polyethylene glycol (PEG) into the crystallization medium favors crystal growth by disordering protein aggregates either, in the medium, or at the surface of the crystal, therefore allowing the formation of much larger crystals [6]. In humans, the organic matrix of renal calculi influence the crystal growth that occurs in such pathological mineral deposits and the use of antibodies against these molecules has allowed to visualize their distribution in a variety of normal and pathological mineralized tissues [7]. In the case of cholesterol, nucleation time referred as the time of the first appearance of cholesterol crystals from isotropic crystal-free biles on light microscopy, is being used to assess the potency of nucleating agents such as immunoglobulins [8]. As demonstrated by this test, biliary IgM (M i n m u n o g l o b u l i n s ) , are more potent t h a n biliary IgG (G i m m u n o g l o b u l i n s ) as a potential nucleating agent. A 42 kilo D a l t o n s (KDa) biliary glycoprotein has been also shown to be related with cholesterol crystallization promotion on the pathogenesis of gallstone disease [9]. The protein is an extensively glycosylated (37%) monomer with an i s o e l e c t r i c p o i n t , pI, of 4.1, probably due to its sialic content. Enzymatic N-deglycosylation removes the carbohydrate moiety and inactivates the promoting activity. Enzymatic proteolysis results in both a complete structure degradation and functional inactivation. Biomineralization is the term used to refer to biological metal precipitation, however it should be borne in mind that not necessarily true crystal minerals are formed through this process. Thus, the concept may be missleading. The metal precipitates can exist either intra- or extracellularly, attached or unattached to the cell surface. The adsorption of metals to living or dead cells, which does not involve metabolic energy or transport processess, has been termed "biosportion" [10]. Biosorption has been the metal-microbe interaction most widely studied and covers both the terrestrial and the marine environment. It should be emphasized, however, that biosorption by marine bacteria may have special implications since they occur in a high ionic medium with high concentrations of some metals. Aside of its use for detoxification of heavy metals in aqueous systems [11], very little

905 information is available about the influence of bacteria in metal crystal formation. One important exception may be the cycling of manganese in the marine environment. In general, metal-precipitating bacteria are abundant in hydrothermal vents and the mechanism of metal precipitation may differ greatly among the different types of microorganisms and metals. Metal-microbe interaction can be classified into two classess [12]: a passive process which does not require the direct participation of living organisms and can occur whether the microbes are alive or dead; and an active process, in which some metabolic or enzyme activities are involved. In the first type interaction results as a consequence of the negative charge of microbial cell surfaces and their exopolymers, or through reactions with extracellular complexing agents that can be attached to the cell surface, or released into the medium. This passive sorption of metals is mediated by various functional groups which promote adsorption, ion-exchange, chelation, and/or covalent binding through carboxyl, hydroxyl, sulfhydryl, amino, imino, imidazole, sulfate and/or sulfonate groups, present in the cell surface, in the cell wall, or within the cytoplasm, as polyssachhcarides or glycoproteins. A classical application of this mechanism of sorption is wastewater treatment. Microorganism that can produce extracellular complexing agents such as "siderophores" do so to trap the metals by a passive mechanism that later are internalized by an active process. The active removal of metals occurs via extracellular precipitation, redox reactions, intracellular accumulation, or volatilization [12]. The biomineralization of gypsum, for example, seems to be a two-step process initiated by the binding of calcium to the cell surface following of the binding of sulfate to the calcium. In photosinthesizing cells, the sulfate is eventually replaced by carbonate to form calcite at the cell surface as the pH increases. Thus, the cells provide essential nucleation sites and the chemical conditions necessary for mineral formation [13].

2. THE P H Y S I O L O G I C A L ROLE OF CRYSTAL F O R M A T I O N In polar and sub-polar fishes, for example, some glycopeptides protect their body fluids from freezing. Such peptides prevent the growing of ice by a noncolligative process thus inducing the development of unusual and strikingly similar crystal habits suggesting that the peptides show some affinity for similar crystal faces of ice [14,15]. Ideally, the protein has an exact octapeptide repetition and is assumed to have an helical conformation to control crystal formation [3]. Certain bacteria promote the formation of ice in super-cooled water by means of ice nucleators; the opposite effect, inhibition of ice formation, is common for a group of glycoproteins found in different fish and insect species. These substances termed anti-freeze glycoproteins promote the supercooling of water with no appreciable effect on the equilibrium freezing point, or melting temperature, by binding to a growing ice crystal and slowing crystal growth [16]. In a similar fashion, glycosaminoglycans (GAGs) and some sulphated polysaccharides are

906 involved in preventing urinary stone formation by inhibiting crystal growth and agglomeration, and possibly also nucleation. They can prevent crystal adherence, correct an abnormal oxalate flux, and avoid renal tubular cell damage [17]. Other synthetic polyanions, including peptide analogs of naturally occurring proteins, inhibit the nucleation and growth of calcium salt crystals [18] under physiological conditions.

3. CRYSTAL PROMOTION Bacteria interact with metals, not only because they are needed as nutrients, but also as important agents in their geochemical cyling. Such metal cycles are driven by diverse chemical and biological processes and may be biotechnologically important. For example, in mining industry the bacteria may participate in the oxidation and solubilization of metal sulfide ores and to recover valuable metals from low grade ores. Metal precipitation by bacteria refers to the transformartion of a soluble metal to an insoluble form. Usually, in the first stage of the process the formation of amorphus, highly hydrated, precipitates are obtained, but with aging they can be transformed into crystals [19]. In the crystallization of biomolecules two critical steps, the nucleation of the initial seed and the enlargement of this seed, determine the quality of the final crystal. However, the degree of supersaturation required to nucleate crystals is often higher than the optimal concentration necessary for enlargement. Thus, even under conditions suitable for crystal growth, kinetic factors may prevent the onset of nucleation and crystal growth. Also, spontaneous nucleation may occur at such frequencies that the resulting microcrystalline precipitates are indistinguishable from their amorphous counterparts. It many situations, it is advisable to decouple crystal growth from nucleation in order to grow large, regular crystals. One must not only control the number of seeds, but also reduce the supersaturation level and, therefore, decrease the incorporation of defects detrimental to crystal quality. Seeding techniques provide a preformed, regular crystal surface onto which further molecules may be aggregated in an orderly form, generally at a lower degree of supersaturation than is required for nucleation. Such techniques are ideally suited to bypass the nucleation step, and hence accomplish the decoupling between nucleation and crystal growth. Three aspects for seeding should be considered: 1) preequilibration of the solution to be seeded and determination of the proper supersaturation level for seeding; 2) the environment and necessary precaution for seeding; and 3) the streak seeding technique and how it can be used in conjunction with microseeding and macroseeding [20]. In the case of some proteins, crystallization may be affected by the aminoterminal segment that sticks out to interact with a symmetry related molecule through an intermolecular salt-bridge. For example, removal of the r e s i d u e lysine in position 38 (Lys 38), in the case of an endonuclease from Clostridium thermocellum, or the substitution of its bridge-forming residues by site-directed

907 mutagenesis, promotes crystal packing arrangements different from the wild type enzyme [21]. Flexible amino-, and/or carboxy-terminal extensions, influence crystal nucleation but not crystal growth [21]. Following this idea, an attempt to facilitate crystallization has been the use of engineering cysteins to promote formation of a "back-to-back" dimmer that occurs in different crystal forms of wild-type and m u t a n t lysozymes [22]. The designed double m u t a n t in which the surface residues of a s p a r a g i n e Ash68, and a l a n i n e Ala93, were replaced by cysteines, formed dimers in solution and crystallized isomorphously to wilde-type but a much faster rate. Overall, the m u t a n t structure remained very similar to wild-type despite the formation of two intermolecular disulfide bridges. The results suggest that the formation of the lysozyme dimmer is a critical intermediate in the formation of more than one crystal form and that covalent cross-linking of the intermediate accelerates nucleation and facilitates crystal growth. Another role seems to be played by polysaccharides in the case of the calcifying algae Pleurochrysis carterae, which produces structures known as coccoliths in homogeneous cell cultures. The polysaccharides PS-1 and PS-2 have been localized in the crystal coats of mature coccoliths, and in electron dense Golgi particles. These polyanions are synthesized in medial Golgi cisternal and co-aggregate with calcium ions into discrete 25 nm particles. The polysaccharides remain with the mineral phase after the coccoliths are extruded from the cells [23]. The effect of many different compounds on calcium stone formation has been evaluated [24]. A few appear to inhibit the nucleation rate, growth and suspension density (crystal mass produced) in proportion to its concentration. On the other hand, glucose, glycerol and certain amino acids which are recognized as osmoregulators, and are produced by halophylic microorganisms, have been also evaluated as halite crystal habit modifiers [25]. In the same way, silica gel as an inert particle, and the ferrocyanide ability to form dendrite crystals of halite, have been already reported [26,27]. Altogether these data point to a dramatic effect of different compounds on mineral and protein crystallization and have induced us to study the influence of halobacteria, and of its components, in the nucleation, shape and growth of NaC1 crystals under natural conditions. 4.

I N F L U E N C E OF HALOBACTERIA ON HALITE FORMATION

It seems that since halobacteria have not being connected to any dreadful disease, and because they are easy to grow, albeit very slowly for practical purposes, interest on their biotechnological use is very scant. In fact, their ecological role is still poorly understood [28]. The term halobacteria refers to the halophylic Archaea, not to halotolerant bacteria, and this distinction is of outmost importance. The order Halobacteriales contain six genera and the number of newly found species is increasing. Phylogenetic data indicate that they are among the most modern Archaea, as their strong preference for aerobic life

908 would suggest. Halobacteria require a minimum of 1.5 M (9%) sodium chloride for growth, and in most cases the optimum lies between 3.5-4.5 M (21-27%) NaC1. Such salt concentration exceeds by far the total saltiness of sea water (which is about 0.6 M or 3.5% of dissolved salts). As a particular feature, halobacteria exhibit active growth and motility in saturated salt solutions, only reduced when entrapped into the salt crystals they bump into. In the absence of salt, all except the coccal forms of halobacteria disrupt promptly and dramatically. In any location where the basic requirement of salt is met, halobacteria will be found. They may become the dominant microflora in what appears to be a classical ecological succession [28]. The halobacteria can be grouped into three major types reflecting their natural source: in the first category are habitats in which the salt mixture derives from evaporated seawater. The commercial salterns consist of shallow evaporating pools containing brines of steadily increasing salinity. They are an ideal ecosystem to observe and study microbial succession. As the brine concentrates under the hot sun and wind, its density and, therefore, thermal storage capacity increase, while its ability to hold dissolved oxygen declines. Certain ions reach saturation earlier than others and precipitate out. For example, calcium and sulfate crystallize out as gympsum. In the less-saline stages (<2M salts), halotolerant Prokarya predominate. Between 2-3 M, halobacteria progressively displace Prokarya. At 3 M, halobacteria alone are found as a thriving component of a characteristic hypersaline ecosystem that includes green algae (Dunaliella spp), brine flies, and brine shrimp. Above 3 M, a more subtle succession of progressively more halophylic Archaea appears as the brines evaporate to the saturation point and salt crystal begin to form. On the other hand, underground salt mines reflect their marine origin as evaporitic sediments from past geological era and are composed mainly of sodium chloride and mixed with lesser amounts of potassium, magnesium and calcium salts and layers of fine clay particles. The halobacteria can be recovered from samples of brines that have come to the surface, and also from intact rock salt. Finally, the salt lake habitats are by no means generic. Each is a terminal lake and collects ion-rich runoff from its own drainage area. Chemically, they resemble the composition of surrounding rocks and mineral deposits. Warm, slightly acid to neutral-pH lakes, such as the Great Salt Lake and the Dead Sea, tend to have high sodium and magnesium concentrations. Warm alkaline lakes, also called soda lakes, have high sodium and carbonate concentrations, with very low amounts of magnesium [28]. As one could expect, hypersaline habitats are intensely stressful: first, the halobacteria have to withstand the elevated ionic concentration that no other life form seems to tolerate. All halobacteria require sodium ions for structural integrity. Requirements and tolerances for other ions depends on the strain's original habitat. To deal with the osmotic challenge that salt concentration imposes, the halobacteria accumulate intracellular substances to balance external stress, for example potassium ions up to 5 M. In spite of being shallow habitats affected by wind, the concentration of dissolved oxygen in brine waters

909 is low, and becomes even more critical as temperature rises. Midday temperatures in salterns can average 450C rising to more than 60~ occasionally. To survive under such conditions the halobacteria have adopted several strategies: one appears to be the flexibility of their outer layers which allow the cells to take a variety of flattened shapes including squares, rectangles, flattened discs, and triangles. Such morphologies help the bacteria to exchange nutrients and gases more efficiently. In addition, active flagellar motility is common in the majority of fresh halobacteria isolates, suggesting that such cells use positive taxis to reach oxygen. Other halobacteria have gas vesicles which help them to position in the water column [28]. The ability of halobacteria to survive long dry seasons indicate that brine inclusions are perhaps the most critical strategy of the bacteria for survival. It has been proposed that halobacteria may act as nucleation factors for halite and other minerals [28], in a similar fashion as the nucleation protein of Pseudomonas syringae in ice. The halobacteria however, lack peptidoglycan and the non-coccoidal genera exhibit a high-molecular weight complex glycoprotein bearing sulfated oligosaccharides. This protein, as a rule, has an excess of acidic amino acids over basic amino acids, and enables the cell to bind and organize large amounts of cation. The cytoplasm membrane is composed mainly of isoprenyl glycerol dieter lipids with chain lengths of 20-25 carbons. Both, sulfated and glycosylated lipids are used as taxonomic criterium. When sodium chloride crystallizes in natural salterns, the microflora normally present in the ponds influence the overall process in different ways. Some living halobacteria may even get entrapped within the fluid inclusions as the crystal develops, and thus affect the physical characteristics of the product. The microorganism might be able to survive under such conditions for extended periods (about 4 years) and this fact imposes some recommendations for the proper handling and application of the contaminated product. The problem becomes evident when the salt is used for preserving fish, meats and hides: it has been shown that red discoloration of food preserved with salt is due to the halobacteria that remain viable in crude solar salt after harvesting [29]. The survival of the entrapped bacteria is determined by storage conditions, and sometimes they may even reproduce [30]. The water content of fresh solar salt is usually in the range between 2-6% (w/w), and in some cases up to 15% (w/w) [31]. The fluid inclusions are readily observed under low power magnification and may be described as negative cubes, or oblongs, with slightly rounded corners. They contain aqueous solutions and, in some cases, small gas bubbles. They range in size from less then 1 pm in the longest dimension to several millimeters. Most however are in the micrometer size range. If the salt is not exposed to extreme heat, pressure, or recrystallization, inclusion fluids may be considered similar in composition to the evaporitic fluids from which the salt crystals originated. The extent of bacterial entrapment in fluid inclusions has been studied using pure cultures of halobacteria added to saturated salt solutions [32].

910 With regard the influence of halobacteria in NaC1 crystal habit formation, much less information has been collected in spite of the many investigations on materials affecting the halite crystal habit. About 500 different organic and inorganic compounds have been tested to modify the crystal habit of NaC1 and only 12 have shown to be effective [33], among these urea, cysteine, creatinine, papain, monosodium glutamate, some cadmium salts, the sodium hexametaphosphate, aluminon, and the chlorides of Zn 2§ and Mn 2§ cause the formation of octahedra, while ferrocyanides were capable of modifying the habit formation (US Patent 2 642 335 (16 June 1953) and 3 090 756 (21 May, 1963)). A combination of cubes and octahedrons have resulted from the action of NaOH, boric, phosphoric and hydrochloric acids [34]. The addition of mercuric chloride causes combinations of the cube and dodecahedron crystals, and antimony chloride causes a combination of cubic, octahedron and dodecahedron ones. The greatest effect seems to be played by the ferrocyanide salts which in concentration as low as 10 ppm can cause the formation of vicinal faces that prevents salt bridges between contiguous crystals. By controlling the concentration of ferrocyanide it is possible to favor a particular crystal formation which constitute the basis for the anti-caking and low-bulk density sodium chloride patents. The dendrite crystals formed at high concentrations of ferrocyanide salts are in fact cryptocrystalline extensions of cube corners which grow to lengths of 2.5 cm or more. On the other hand, when a combination of sodium hexametaphosphate and a soluble aluminum salt are used, the formation of octahedra resulting from the action of hexametaphosphate is suppressed and cubic crystals with a basket shape are formed. Such structures have been described as aggregates of a geometrical arrangements of tetrakaidecahedra, a combination of 14-faced of cube-octahedron with eight hexagonal faces and six square faces, all with edges with equal length [26]. Guerrero Negro appears to be the largest solar saltern of the world and is located on the west coast at the middle of the Baja California Peninsula, in M6xico. Its products, resulting from the evaporitic action of solar and wind energy make a business worth several hundreds of millions of dollars every year and are exported worldwide. Thus, it is just reasonable to become interested in studying the influence of halobacteria in halite production. Our aim was not only supported by a scientific curiosity, but also in view of increasing competition of other countries in the international salt market. Another goal was to consider the possibility of increase production capabilities by taking advantage of the role of halobacteria in halite formation. To determine if halobacteria indeed influence the production of halite, an experiment was designed to study the effect of an endemic isolate (Haloarcula SP8807) and of its S-layer on halite crystal habit, and crystal growth rate.

911

5. MATERIALS AND METHODS 5.1. B a c t e r i a s t r a i n s and g r o w t h c o n d i t i o n s The bacteria strains tested in this study were Halobacterium halobium NRC 817, Haloarcula vallismortis ATCC 29252, Haloferax mediterranei ATCC 33500, Haloarcula SP8807, Vibrio parahaemolyticus MMF6, Planococcus spp. M6P2, and Azospirillum brasilense Cd DSM 1843. The halobacteria Haloarcula SP8807 was isolated from a commercial saltern in Pichilingue, La Paz, B.C.S., M~xico, while the eubacteria V. parahaemolyticus, and Planococcus spp. were obtained from an hypersaline ecosystem at Guerrero Negro, B.C.S., M6xico. All the other halobacteria, as well as A. brasilense, which is an organism not found in hypersaline environments, were acquired from an international collection. The halophylic bacteria were grown in HEC medium formulated as follows in 25% seawater (g/l): NaC1, 195; MgC12"H20, 16.25; MgSOn-7H20, 25; CaC12-2H20, 0.6; KC1, 5; NaHCO3, 0.2; NaBr, 0.6; NH4C1, 2.5; FeC13"6H20, 0.0062; KH2PO4, 0.62. The carbon and nitrogen source came from yeast extract (5 g/l) and casein hydrolysate (1.0 g/l) [35]. Incubation was done at 38~ under stirring at 150 rpm. A. brasilense on the other hand, was cultivated in Nutrient Broth (Merck, Germany), and incubated at 30~ under agitation as above. 5.2. D e t e r m i n a t i o n of the n u m b e r and size of halite cubic crystals with scanning electron microscopy One ml of bacteria culture at the exponential-phase was centrifuged for 5 min., 16,000 x g at 20~ The pellet was washed three times with sterile 30% NaC1 solution, and resuspended in 1 ml of the same solution. From this suspension, 5 ~1 drops were placed on a gold-coated glass coverslip as described by [25]. The drops were incubated at 24~ without disturbance. After 10, 20, and 30 min., excess of brine was eliminated with a filter paper. The crystals were air-dried and gold-coated at 20 mA during 8 min. The blank consisted of 5 ~1 drops of sterile 30% NaC1 solution treated under the same conditions. The crystals were viewed using a Philips 515 scanning electron microscope at an accelerating voltage of 25 kV. 5.3. E v a l u a t i o n of effect of strains, s-layer and c h e m i c a l m a t e r i a l s on the crystal form of NaC1 One ml of bacteria culture at the exponential-phase was centrifuged for 5 min., 16,000 x g at 20~ The pellet was washed three times with sterile 30% NaC1 solution, and resuspended in 1.0 ml of the same solution. From this suspension, twenty drops of 5 ~1 each were placed on a clean microscope slide and kept in a chamber for 24 h at 35~ and 40% relative humidity. The slides were observed with a phase-contrast microscope (Nikon Labophot, Japan) at low magnifications of 2.5X, 10X and 20X. The following samples were mixed with sterile 30% NaC1 solution: S-layer from Haloarcula SP8807 (20 mg/ml); potassium ferrocyanide, glucose, glycerol, casein hydrolysate, and silica gel (63-200 mm) from Sigma Chemicals Co. St.

912 Louis, Mo, and amino acids from Merck, Germany (8003-8004), at concentrations between 200 to 2000 mg/ml. Twenty 5-~1 drops of each mixture were placed on clean microscope slides, dried, and observed as described above.

5.4. Preparation of s-layer and spheroplast of Haloarcula SP8807 The S-layer was isolated as follows: one ml of bacteria culture in its exponential phase was centrifuged for 5 min., 16,000 x g at 20~ The pellet was resuspended in 0.1 M MES [2(N-morpholino)ethane sulfonic acid] buffer, pH 7.0, plus 0.5 M sucrose, 0.25 M NaC1, and 0.01 M MgC12 [36]. After 30 min., the cells were converted into spherical bodies by a decrease in external sodium chloride and magnesium salt concentration. Under these conditions, the S-layer can be removed and dissociated from the cell surface [37,38]. The new cell suspension was centrifuged for 5 min., 16,000 x g at 20~ yielding a pellet of spheroplasts and a supernatant containing the dissociated S-layer subunits. The S-layer subunits were recovered and dialyzed according to [39]. After dialysis, the S-layer subunits were reassembled in a salt solution composed of 4 M NaC1, 25 mM KC1, and 80 mM MgSO4, pH 3.2 [37]. The self assembled S-layer was examined by SDS-PAGE, transmission electron microscopy, and tested in crystal formation studies. For negative staining preparations we used a solution of CaC12 10 mM pH 6 to promote the reassembly of the S-layer subunits [40,41]. 5.5. S o d i u m dodecyl sulfate-polyacrylamide gel electroforesis (SDSPAGE) Electrophoresis was done using gradient PhastGel 4-15% with sodium dodecyl sulfate (SDS) buffered strips (Pharmacia, Uppsala, Sweden). The molecular weight markers were: bovine albumin (66 kDa),egg albumin (45 kDa), glyceraldehyde-3-phosphate dehydrogenase (36 kDa), bovine carbonic anhydrase (29 kDa), bovine trypsinogen (24 kDa), soybean trypsin inhibitor (20.1 kDa) and a-lactalbumin (14.2 kDa) from bovine milk (Sigma, SDS-7 Dalton Mark VII-L, USA). The gels were stained for protein by the silver impregnation method with PhastGel electrophoresis media (Pharmacia, Sweden), and periodic acid-Schiff (PAS) staining procedure for the detection of glycoproteins using the PhastSystem [38].

5.6. Negative staining for electron microscopy Negative staining preparations for electron microscopy, EM, were done with fixed and unfixed self-assembled S-layer. Reassembled S-layer was prefixed in 3% (v/v) glutaraldehyde in buffer solution of 4 M NaC1, 25 mM KC1, 80 mM MgSO4, pH 3.2, during 1 h at 4~ in the dark. After this time the sample was centrifuged for 5 min., 10,000 x g, at 20~ and the supernatant replaced with clean fixing solution under the same conditions. The fixed S-layer was washed and preserved at 4~ with the same buffer solution. A few drops (10 ~1) of fixed samples containing S-layer fragments were placed on a small sheet of Parafilm and diluted 1:40 with a buffer solution of 2 M NaC1,

913 12.5 mM KC1, 40 mM MgSO4, pH 3.2. Immediately 10 pl of sample were transferred to a copper grid previously exposed to glow discharge. After 3 min., each grid with sample was stained with 2-3 drops of 2% of aqueous uranyl acetate for 1 min. Excess stain was removed by blotting with filter paper [39]. 10 ILl of unfixed samples containing reassembled S-layer in solution of CaC12 10 mM, pH 6, were placed on copper grids previously exposed to glow discharge. After 3 min., each grid with sample was stained with 2-3 drops of 2% of aqueous uranyl acetate for 1 min. Excess of stain was removed by blotting with filter paper [39]. Some samples were micrographed without staining to avoid the dilution of microcrystals. The samples were viewed in a Philips 410 transmission electron microscope at an accelerating voltage of 80 kV.

6.

RESULTS AND DISCUSSION

6.1. C r y s t a l l i z a t i o n rates Following the methodology described above, we found no significant differences in the weight of sodium chloride crystals harvested with, or without, halobacteria. That is, in either case the amount of salt crystals produced was about the same so we may conclude that the halobacteria does not improve productivity of salterns. Interestingly, the halobacteria did affect the number and size of the sodium chloride crystals (Fig. 1; for convenience, crystals smaller t h a n 1 mm were disregarded).

350

A4

300 <J I-09 >-n,' o u_ 0 n," w rn

250

~2 200 150 ,4

100

/m___--m/m

1

,m ~

::3 Z

50 . . . . .

10

3'o

4'0

I

T I M E (h)

Figure 1. Influence of halobacteria on the number and size of halite crystals. For convenience, only crystals larger than 1 mm were considered. The data were used to performed a one-way analysis of variance (ANOVA) test with a confidence of p < 0.05. 1 - NaC1 (1-8 mm), 2 - Halobacterium halobium NRC 817 (1-25 mm), 3 - Haloarcula vallismortis ATCC 29252 (1-15 m m ) , 4 - Haloarcula SP 8807 (1 - 15 mm).

914 When our data were statistically analyzed by a one-way analysis of variance (ANOVA) to evaluate the effect of halobacteria on crystallization rates, we found that the number of crystals formed in the presence of the bacteria tested, differed significantly from the control (Fig. 1). The blank produced few crystals of similar size, while in the presence of halobacteria many cubic crystals of different size were generated (Fig. 1). The analysis also indicated that the shape, and/or morphology, of the halobacteria provoked different effects on crystal habit [43]. In particular, bacteria with triangular, or square-shape (e.g. Haloarcula SP8807) seem to provide a template and serve as a means of mechanical nucleation to promote crystal formation in saturated solutions [32].

6.2. Crystal f o r m a t i o n of NaC1 Halite has a crystalline form based on a cubic symmetry [26,27]. The influence of Haloarcula strain SP8807 on NaC1 crystal habit growth was observed by scanning electron microscopy (Figure 2). The micrograph illustrates the close association that occurs between the bacteria and halite crystals. In the caption the crystal is shown as if emerging from the cell. This caption is a good physical evidence on the role of halobacteria on crystal growth. By transmission electron microscopy, on the other hand, and in the presence of unfixed and unstained samples of S-layer, symmetric particles of 50 nm of diameter, which could be microcrystals of sodium chloride in their early stage of mineralization, were observed (Fig. 3). These two approaches suggest that either whole cells, and/or their S-layer, can be involved in the nucleation and crystal formation of NaC1. Haloarcula strain SP8807 also caused the formation of dendrite crystals of halite. The SEM micrograph of Figure 4A shows crystal formations in which dendrite crystals lie outside, and in between, cubic crystals. A high magnification of this caption shows that the dendrite crystals are close to the halobacteria surface, and confirms its ability to modify the crystal habit into irregular shapes (Fig. 4B). The other strains of bacteria tested in our study were also capable of inducing formation of dendrite crystals, except in the examples of V. parahaemolyticus and Planococcus spp. which suggest that not all the microbiological members present in the crystallizer ponds are involved in the modification of crystal habit. This assumption was further supported by the inability of A. brasilense, a non-hypersaline eubacteria, to form dendrite crystals. SDS-PAGE analysis of purified S-layer from Haloarcula SP8807 showed a single protein with a molecular weight of 66 kDa (Fig. 5). Noteworthy, the staining procedure for the detection of glycoproteins failed to reveal any band in the electrophoresis gel. It is well known that the S-layer of most halophylic archaebacteria appear to be composed of glycoproteins [44], a report on the detailed chemical structure of a glycopeptide of H. salinarum [45], and the primary structure of the cell surface glycoprotein of H. halobium [46], confirmed such observations [47,48]. The cell surface glycoprotein of H halobium has a molecular mass of about 120 kDa (core protein = 87 kDa). Recently, Sumper et al. [49] reported a partial chemical characterization of the S-layer glycoprotein of

915

Haloferax volcanii in which the m a t u r e polypeptide contains 794 amino acids with a calculated molecular mass of 81 kDa. Although glycosylation is not an obligatory step in S-layer biosynthesis, it represents an i m p o r t a n t proteinmodification reaction which can add a great potential to the diversification of bacterial cell surface properties [50]. In this context, it is interesting to note t h a t upon continuous cultivation under optimal conditions, some bacterial strains loose their ability to glycosylate S-layer proteins [44].

Figure 2. (A) Micrograph of halobacterial cell showing a cubic structure assumed to be a crystal of sodium chloride. The bar represents 1.0 gin. (B) High magnification showing some detail of a sodium chloride crystal on the surface of an halobacterial cell. The bar represents 0.5 gm.

916

Figure 3. An unstained mount of a mineralized S-layer fragment in which the contrast comes from the sodium chloride crystals. The arrows show symmetric bodies of 50 nm of diameter, probably sodium chloride crystals, on an early stage of formation. The bar represents 0.5 gm.

Negatively stained preparations of S-layer from Haloarcula SP8807 strain examined by transmission electron microscopy show structures with a honeycomb appearance of the reassembled S-layer (Figure 6). These are the units suspected to act as templates for crystal nucleation and growth. As in the case of the S-layer protein isolated from Bacillus coagulans, which was recrystallized on an air/water interface and on phospholipid films [51], the hydrophobic face of the protein m a y actually associate with the air/water interface, leaving the negatively charged inner face oriented itself towards the switterionic head groups of the lipids. Hence, individual monocrystalline areas m a y grow isotropically in all directions, until the front edge of a neighboring crystals is met, forming the honeycomb structure observed in Fig 6. Recently it was proposed t h a t the S-layer m a y serve also as a template for fine-grained gypsum and calcite formation [52].

917

Figure 4. Details of dendrite crystals. (A): Dendrite crystals built by flat cubic microcrystals (arrows). Bar represents 5 gin. (B): A high magnification showing the presence of halobacterial cells (arrows) combined with amorphous halite crystals (empty arrow). Bar represents 2 gm.

918

kDa

-,,1-,-,66 ~~45

U

Q

.91,,,,- 36

,~-~1--" 29

/,~1-,,--24 . '~20

14 Figure 5. SDS-PAGE 4-15% gradient gel, silver stain. Lane A: profile of proteins from spheroplast of Haloarcula SP8807; Lane B: profile of whole-cell protein from Haloarcula SP8807; Lane C: purified S-layer subunits from Haloarcula SP8807; Lane D: profile of molecular weight markers.

As shown in Figure 7, the S-layer of Haloarcula SP8807 (20 pg/ml) modifies the crystal habit of halite yielding dendrite crystals, as well as cubic crystals (Fig. 7). The dendrite crystals and their branches at micrometer level showed flat cubic microcrystals combined with amorphous crystals and halobacterial cells (Fig. 4). Naked cells, or spheroplasts of Haloarcula SP8807, on the other hand, showed to be extremely fragile to influence the crystal production. In fact, partial lysis was observed which translated into many protein bands in SDS-PAGE. A 66 kDa light band from spheroplast samples (Fig. 5) confirmed the observations of Jarrel and Sprott [36] who reported that spherical bodies produced from logarithmic-phase cells contained a cell wall that was thinner than the wall of the original rod shaped halobacterial cells.

919

Figure 6. Negatively stained (2% uranyl acetate) preparations of S-layer self-assembly products of Haloarcula SP8807. (A): Unfixed S-layer reassembled with CaC12 10 mM, pH 6. The arrow shows an area with trace of periodicity in one dimension). The bar represents 100 nm. (B): Selfassembled flat sheet-like S-layer showing a honeycomb appearance of an hexagonal arrangement (circle), stained in presence of 2 M NaC1, 12.5 KC1, 40 mM MgSO4, pH 3.2. Bar represents 200 nm.

In addition to the above, we studied the effect of different compounds on the induction of dendrite crystals of halite, as shown in Table 1. Again, ferrocyanide was found the most active agent inducing the formation of dendrite crystals and at high concentration (250 ~g/ml) caused dendrite growth of sodium chloride crystals e n t r a p p e d within cubic crystals. Casein hydrolysate (1200 to 2000 pg/ml), and some aminoacids tested, did not induce the production of dendrite crystals (Table 1). Similarly, the absence of dendrite crystal formation in the presence of silica gel demonstrates t h a t this phenomenon is not the result of a inert particle effect.

920

Figure 7. Dendrite halite crystals caused by S-layer (20 [ag/ml) from Haloarcula strain SP8807. Bar represents 50 [am.

Table 1 Effect of dissolved and suspended materials on the formation of halite crystals Shape Relative Frequency Agent Concentration abundance a (%)b (og/ml) Dendritic c (+++) 100 Ferrocyanide 250 Cubic (+++) 100 Glucose 250 Cubic (+++) 100 Glycerol 250 Cubic (+++) 100 Amino acids e 200-2000 Cubic (+++) 100 Casaminoacids 200-600 Dendritic d (++) 15 1200 Dendritic (+) 55 Cubic (++) 30 Dendritic d (+++) 10 2000 Dendritic (++) 40 Dendritic (+) 40 Cubic (++) 10 Cubic (+++) 100 Silica gel 100-2000 Dendritic d (+++) 50 Halobacteria cells 200 f Cubic (++) 50 Dendritic d (+++) 50 S-layers 20g Cubic (++) 50 a= (+++) Abundant; (++)= common; (+)= present

921 b= Frequency found in 60 experiments c= Inside the cubic crystals (50 gm average width of the branches) d= Between the cubic crystals (10 gm average width of the branches) e= D-L alalnine, L(-) proline, glycine, D-L serine, L cysteine, D-L threonine, L(-) tyrosine, L histidine, L(+) lysine, and L glutamic acid f= On basis of dry weigth g= On basis of Coomassie protein determination.

In conclusion, the observation of cubic structures associated with halobacterial cells and S-layer samples suggests that they may serve as templates in the nucleation and halite formation. The modification of halite crystal habit, resulting in dendrite shape, was attributed to the .proteinaceous component of the S-layer of the halophytic archaebacteria. Monitoring the types and concentration of dissolved organic carbon compounds, and of halobacteria, in natural solar salterns may be an important biotechnological tool in the operation of salterns.

ACKNOWLEDGEMENTS

This research was part of the doctoral thesis of ALC at the CCH-UNAM Biotechnology Program. ALC wishes to t h a n k Dr. Sergio Sfinchez-Esquivel, Dra. Amelia Farres for their encouragement and support during such studies, Dra. Susanne Schultze-Lam for valuable suggestions and comments, Dra. Eugenia Klein from the Weizmann Institute of Science for the electron microscopy facilities. The Figs. 1, 4B, 5, 6, 7, and Table 1 are reproduced from [24] with the kind permission of Taylor and Francis Ltd., Publishers.

REFERENCES

1. R.A. Judge, M.R. Johns and E.T. Whit, Biotechnol Bioeng., 48 (1995) 316. 2. McPherson and P. Shlichta, Science-Wash., 239 (1988) 385. 3. H. Mizuno, PROTEINS: Struct. Funct. Genet., 5 (1989) 47. 4. P.G. Vekilov, M. Ataka and T. Katsura, Acta Crystallographica Section D Biological Crystallography, 51 (1995) 207. 5. M. Gross and R. Jaenicke, Biophys. Chem., 45 (1993) 245. 6. W.J. Ray Jr., PROTEINS: Struct. Function Genet., 14 (1992) 300. 7. M.D. McKee, A. Nanci and S.R. Khan, J. Bone and Mineral Research,

I0 (1995) 1913. 8. P.R.C. Harvey and G.A. Upadhya, J. Lipid Res., 36 (1995) 2054. 9. Abei, P. Kawczak, H. Nuutinen, A. Langnas, J. Svanvik and R.T. Holzbach, Gastroenterol., 104 (1993) 539. I0. M. Tsezos, in: Microbial Minerak Recovery, H.L. Ehrlich and C.L. Brierley, (eds.) McGraw-Hill, NY (1990) 325.

922 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

R.W. Olafson, Env. Health Persp., 65 (1986) 71. M.T. Bradley, in: Genetic Engineering, J.K. Setlow (ed.), 17 (1987) 231. J.B. Thompson and FG. Ferris, Geology, 18 (1990) 995. J.A. Raymond and A.L. DeVries, ACS-Symposium-series-American-ChemicalSociety, USA, 444 (1991) 249. A. Chakrabartty and C.L. Hew, Eur. J. Biochem., 202 (1991) 1057. A. Parody-Morreale, K.P. Murphy, E. Di-Cera, R. Fall, A.L. DeVries and S.J. Gill, Nature, 333 (1988) 782. E.R. Boeve, L.C. Cao, C.F. Verkoelen, J.C. Romijn, W.C. De-Bruijn and F.H. Schroder, World Journal of Urology, 12 (1994) 43. E. Mueller and C.S. Sikes, Calcified Tissue Int., 52 (1993) 34. T.J. Beveridge, Annu. Rev. Microbiol., 43 (1989) 147. E.A. Stura and I.A. Wilson, Methods (DULUTH), 1 (1990) 38. V. Chitarra, H. Souchon, S. Spinelli, M. Juy, P. Beguin and P.M. Alzari, J. Mol. Biol., 248 (1995) 225. D.W. Heinz and B.W. Matthews, Protein Engineering, 7 (1994) 301. M.E. Marsh, Protoplasma, 177 (1994) 108. K. Kohri, T. Umekawa, M. Kodama, Y. Katayama, Y. Ishikawa, M. Takada, Y. Katoh, K. Kataoka, M. Iguchi and T. Kurita, Eur. Urol., 17 (1990) 173. A. LSpez-Cort~s, J.L. Ochoa and R. V~zquez-Duhalt, Geomicrobiol. J., 12 (1994) 69. R.S. Ploss, Science, 144 (1964) 143. A.C. Shuman, 2nd Symp. Salt North Ohio Geol. Soc., Ohio, 12 (1965) 246. C.F. Norton, ASM News, 58 (1992) 363. F.C. Harrison and M.E. Kennedy, Trans. Royal Soc. Canada, 16 (1922) I01. H.P. Dussault, Proc. 2nd. Int. Symp. of Food Microbiology, 13 BP Eddy (ed.), London, H. M. Stationery Office, (1958). S.J. Lefond, Handbook of World Salt Resources, NY, Plenum Press, (1969). C.F. Norton and W.D. Grant, J. Gen. Microbiol., 134 (1988) 1365. H.E. Buckley, Crystal Growth, Wiley NY 2nd. Ed., (1952) 247, 556. D.W. Kauffman, Sodium Chloride, Reinhold. NY, (1960) 15. M.F. Torreblanca, F. Rodriguez-Valera, G. Juez, A. Ventosa, M. Kamekura and M. Kates, System. Appl. Microbiol., 8 (1986) 89. K.F. Jarrell and G.D. Sprott, Curr. Microbiol., 10 (1984) 147. S. Cohen, A. Oren and M. Shilo, Arch. Microbiol., 136 (1983) 184. K. Hecht, F. Wieland and R. Jaenicke, J. Biol. Chem., 367 (1986) 33. U.B. Sleytr, P. Messner and D. Pum, in: F. Mayer (ed.), Methods in Microbiology, Academic Press, New York, 20 (1988) 29. M. Kessel, I. Wildhaber, S. Cohen and W. Baumeister, EMBO J., 7 (1988) 1549. S.F. Koval and R.G.E. Murray, Microbiol. Sci., 3 (1986) 357. I. Van-Seuningen and M. Davril, Electrophoresis, 13 (1992) 97. W.D. Grant and H.N.M. Ross, FEMS Microbiol. Rev., 39 (1986) 9. P. Messner and U.B. Sleytr, Glycobiology, 1 (1991) 545.

923 45. M. Sumper, Biochim. Biophys. Acta, 906 (1987) 69. 46. J. Lechner and M. Sumper, J. Biol. Chem., 262 (1987) 9724. 47. M.F. Mescher, J.L. Strominger and S.W. Watson, J. Bacteriol., 120 (1974) 945. 48. M.F. Mescher and J.L. Strominger, J. Biol. Chem., 251 (1976) 2005. 49. M. Sumper, E. Berg, R. Mengele and I. Strobel, J. Bacteriol., 172 (1990) 7111. 50. P. Messner and U.B. Sleytr, in: Advances in Microbial Physiology, A.H. Rose, (ed.), Academic Press, New York, 33 (1992) 213. 51. D. Pum, M. Weinhandll, C. Hoedl and U.B. Sleytr, J. Bacteriol., 175 (1993) 2762. 52. S. Schultze-Lam, G. Harauz and T. J. Beveridge, J. Bacteriol., 174 (1992) 7971.

Adsorption and its Applications in Industry and Environmental Protection Studies in Surface Science and Catalysis, Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

925

A d s o r p t i o n - b a s e d optical t r a n s d u c t i o n in optical fibre c h e m i c a l sensors for e n v i r o n m e n t a l a n d i n d u s t r i a l a p p l i c a t i o n s F. Baldini a and S.Bracci b aIstituto di Ricerca sulle Onde Elettromagnetiche "Nello Carrara", CNR, Via Panciatichi 64, 50127 Firenze, Italy bCentro di Studio sulle Cause di Deperimento e Metodi di Conservazione Opere d'Arte, CNR, Via G. Capponi 9, 50121 Firenze, Italy 1.

INTRODUCTION

Optical fibre sensors for chemical compounds have undergone remarkable development in recent years. This is perfectly justified, both because the detection of chemical parameters is extremely important in many industrial and chemical processes, in environmental control and in the biomedical field, and also because optical fibre chemical sensors offer considerable advantages compared to traditional sensors [1,2]. The high degree of miniaturization, considerable geometrical versatility, easy handling, absence of electromagnetic interferences, absence of electric contacts, and the possibility of remote monitoring are only some of the advantages which optical fibres can offer in comparison with other types of sensors. Moreover, the ease in combining the sensing process with an optical network provides an additional advantage for optical fibres, namely the capacity to interrogate many sensors for different parameters simultaneously and with the same optoelectronic unit. An optical fibre sensor consists of three main parts: an optoelectronic system, an optical link, and a probe. The optoelectronic system includes the source, the photodetector and the driving electronic board. The optical source can be a lamp, a laser, a light emitting diode or a laser diode: the last two are preferable, because they are more easily coupled to optical fibres by means of commercially available components and also because they can be modulated electronically, thus avoiding the use of any mechanical coding for the optical signal (e.g. a chopper). The detecting system consists of a photodetector (photomultiplier, photodiode, etc.) coupled to appropriate electronics for the signal processing. Plastic or glass fibres, either single or in bundles, are generally used as the optical link. The most crucial part is unquestionably the probe, where modulation of the light carried by the fibres takes place. Besides assuring the sensitivity and accuracy

926 necessary for the particular type of application, the probe m u s t be characterized by long lifetime, ease of manufacturing, and easy connection with the optical link. The use of a chemical transducer inside the probe, the optical properties of which depend on the concentration of the investigated species, introduces other critical aspects: i) a transfer of mass often exists between the external environment and the sensor since the chemical transducer and the investigated species can be in different phases; thus a lengthening of the response time occurs; ii) since the chemical transducer is often an organic compound, particular attention m u s t be paid to its long-term stability and to its photodecomposition. M a n y sensors have been described for the m e a s u r e m e n t of different chemical p a r a m e t e r s [1,3]. The first optical fibre sensor was a pH sensor, and dates back to 1980 [4]; since then, m a n y chemical p a r a m e t e r s have been investigated with optical fibres, for both aqueous and gaseous species. After a description of the principles at the basis of an optical fibre sensor, a general review on the optical fibre sensors for environmental and industrial applications based on adsorption is presented.

2.

THE O P T I C A L F I B R E

Before entering into details of an optical fibre sensor, a brief description of an optical fibre will be given. An optical fibre is a cylindrical dielectric waveguide made of low-loss materials, such as silica glass. Two main parts can be identified: the core and the cladding, which constitute the central and the outer part of the cylinder, respectively (Figure 1). In general, another layer, the jacket, covers the cladding, for protective reasons. The refractive index of the cladding, n2, is lower than the core refractive index nl: this is a necessary condition for the t r a n s p o r t of electromagnetic radiation.

core

..~---

......

n

.

fl 2 ,:

cladding

,.,

I {

n

1 /

/,"

/

/

1

/

\

n

2

i

i

R

Figure 1. Cross-section of an optical fibre.

r

Figure 2. Refractive index profile in a graded-index fibre.

927

A step-index fibre is an optical fibre characterized by c o n s t a n t values of the refractive indices of the core and the cladding. A graded-index fibre is an optical fibre in which the refractive index of the core n(r) changes gradually along the radius r of the cylinder, from a m a x i m u m nl on the axis to the value n2 on the core/cladding interface (Figure 2). According to the ray optics theory, the description of the light propagation in an optical fibre is based on meridional and skew rays: a meridional ray is an optical ray passing through the fibre axis, and a skew ray is a ray which never crosses the fibre axis. I

~

!

I

Figure 3. Optical path of a meridional ray in a step-index fibre.

A meridional ray is easily identified by the incident angle 3 on the core/cladding interface, or by its complementary angle ~ which is formed by the ray with the fibre axis (Figure 3). The angle a in the figure is the angle formed by the optical ray with the fibre axis in the outer medium. The following relationship exists between a, and ~: n o sin a = nl sin ~ = nl cos

(1)

with no equal to the refractive index of the outer medium. A skew ray is characterized by an helical path. At each reflection, a skew ray touches the core/cladding interface and the surface of an inner cylinder of radius p. In a step index fibre, a meridional optical ray which enters the fibre from the air is guided, if it is incident with an angle ~ t h a t is greater t h a n the critical angle 3 c on the interface core/cladding:

ac = arc sin( n2 / \nl) Total reflection takes place, and the ray is guided along the fibre if:

(2)

928 7~ -- > ~ > ac

(3)

2

In a graded-index fibre each ray is characterized by the i n v a r i a n t ~" = n(r)sin 9(r)= n(r)cos ~(r)

(4)

where ~ has already been defined as the angle formed by the optical ray with the axis of the fibre. The guiding condition is: m

n2 < [ 3 < n l

(5)

A p a r a m e t e r is introduced to define the acceptance cone inside of which the radiation entering the fibre is guided: the numerical aperture NA. In a step-index fibre this p a r a m e t e r is defined as: NA = sina a

(6)

Angle (la is the angle in the outer medium formed with the axis of the fibre by the optical ray t h a t is characterized by a reflecting angle equal to 9 c on the interface core/cladding. In t e r m s of refractive indices, the numerical aperture of a fibre can be written:

:(nl This definition is considered valid also for the graded-index fibre, with nl equal to the m a x i m u m value of n(r), t h a t is, the refractive index on the fibre axis. In terms of electromagnetic waves, every optical fibre can support a certain n u m b e r of modes according to the Maxwell's equations. A monomode fibre is a fibre in which only one mode of electromagnetic radiation is supported; on the contrary, in a multimode fibre, more t h a n one mode is guided. Each mode is characterized by a different value of the constant propagation ~ along the axis direction. The values of the constant propagation can be obtained by solving the eigenvalue equation which derives from the Maxwell's equations applied to the waveguide; as a m a t t e r of fact an exact solution is obtained for only a few cases, among which there is t h a t of a step-index fibre. In all the other cases, only an approximate solution can be found. In any case, propagation constants of guided modes m u s t always satisfy the following disequation: n2ko < [3 < n l k o

271;

ko = ~o

(8)

929 w h e r e k0 is the p r o p a g a t i o n c o n s t a n t of a p l a n e wave in the v a c u u m , a n d ~.o is the free-space w a v e l e n g t h of light. O t h e r limitations on the accepted ~ values depend on the refractive index profile of the fibre. An i m p o r t a n t p a r a m e t e r is the V number, k n o w n also as fibre parameter. It is defined as: V = 2~R In 2 _ n 2 ~/2 _ 2~___R_RNA ~-0 ~0

(9)

w h e r e nl is the core refractive index in a step-index fibre, or its m a x i m u m value in a g r a d e - i n d e x fibre, n2 is the cladding refractive index a n d R the core radius. The value of V n u m b e r is related to the n u m b e r of guided modes supported by the fibre. In a step-index fibre, if V is less t h a n the value of 2.405, the fibre is monomode, o t h e r w i s e it is multimode. In a n optical fibre sensor, t r a n s m i s s i o n properties of the fibres are very i m p o r t a n t since they m u s t be t r a n s p a r e n t at the working wavelengths. Table 1 s u m m a r i z e s the w o r k i n g r a n g e a n d the typical a t t e n u a t i o n for the m a i n types of fibres.

Table 1 The w o r k i n g r a n g e a n d a t t e n u a t i o n of the most u s e d optical fibres Fibre type

W o r k i n g r a n g e (~m)

Silica fibres

0 . 2 - 1.9

A t t e n u a t i o n (dB/m)

0.77 - 1.35 0.003 - 0.005 m a i n optical windows 1 . 0 5 - 1.35 0 . 0 0 0 5 - 0.002 1.45 - 1.75 0.0002 - 0.003 Plastic fibres 0.4 - 0.8 0.3 - 3 Fluoride fibres 1.5 - 4.5 0.002 - 0.02 Chalcogenide fibres 3 . 0 - 11 0.5 - 5 Polycristalline silver-halide fibres 4.0 - 20 0.5-5

Silica fibres g u a r a n t e e low a t t e n u a t i o n s for w a v e l e n g t h values of b e t w e e n 500 nm a n d 1.9 ~m. For ~.<500 nm, as the w a v e l e n g t h decreases, the a t t e n u a t i o n increases up to values on the order of 3 dB/m for L~200 nm. This m a k e s it n e c e s s a r y to use very short fibre lengths. In the visible band, plastic fibres can also be used, but their a t t e n u a t i o n permits utilization for only s h o r t distances. W h a t m u s t be considered is t h e i r lower cost in c o m p a r i s o n with t h a t of o t h e r fibres; therefore, in some cases, t h e i r utilization can be t a k e n into account, n o t w i t h s t a n d i n g the h i g h e r a t t e n u a t i o n . For w a v e l e n g t h s longer t h a n 2.0 ~m, different fibres m u s t be used, a l t h o u g h the l a t t e r are c h a r a c t e r i z e d by h i g h e r a t t e n u a t i o n s . This p r e v e n t s t h e i r use if long optical links are necessary.

930 Moreover, since all the optoelectronic components (sources, detectors, lenses, etc.) are available at a low price only in the band ranging from 400 nm to 1.9 pm, it is apparent that this is the optical window in which most of the optical sensors proposed works. 3.

W O R K I N G P R I N C I P L E S OF AN OPTICAL F I B R E S E N S O R

An optical fibre sensor is a device capable of detecting a physical or chemical p a r a m e t e r on the basis of the modulation of any property of the light carried by the fibre: intensity, phase, polarization state, frequency/wavelength, etc. Optical fibre chemical sensors are mainly amplitude-modulation sensors: the intensity of the light transported by the fibres is directly modulated by the analyte under investigation or by an appropriate sensitive reagent whose optical properties vary with the variation in the concentration of the p a r a m e t e r being studied. The most important working principles on which most optical fibre chemical sensors are based are absorption and fluorescence. 3.1. A b s o r p t i o n In addition to the substances having their own absorption bands, substances which, by interacting with an appropriate reagent, vary their absorption (e.g. acidbase indicators vary their own absorption, depending on the concentration of the hydrogen ions) can also be detected. If the light crosses a liquid sample, the Lambert-Beer law is the equation which relates light intensity I to concentration c of the absorbing species: A = log :_v_-= tic I

(10)

where A is known as absorbance; Io and I are the light intensities transmitted in the absence and in the presence of the absorbing sample, respectively; ~ is the absorption coefficient, and 1 is the optical path. Clearly, this is true if the substance under investigation is the only one absorbing at the considered wavelength; otherwise, the absorption of other substances present in the solution must be considered. If, instead, an appropriate reagent is immobilised on a substrate, the m e a s u r e m e n t is based on the reflection by the solid substrate. A special function (function of Kubelka-Munk) must be introduced which is proportional to concentration c of the substance under examination, according to the Kubelka-Munk theory [5]: F(R)= (1 - R) 2 = kc 2R

(11)

where R is the reflectance of an infinitely thick sample and k is a constant depending on both the absorption and scattering coefficients. If the thickness of the

931 sample cannot be considered infinite, the relationship between reflectance and concentration of the analyte is much more complex: the intensity of the light is partly transmitted, reflected, absorbed and scattered, so t h a t equations (10) and (11) are not followed exactly, and a proper algorithm has to be introduced. It is i m p o r t a n t to point out t h a t the two equations are for a definite wavelength. This means t h a t they can be correctly used only if a monochromatic source is used, i.e. lasers or sources coupled with very narrow interference filters. If light-emitting diodes (LEDs) are used, as often occurs in optical fibre sensors, a multiwavelength optical b e a m is used and equations (10) and (11) are no longer valid. For example, equation (10) would become:

A'= log II~

(12)

where the integral is evaluated on all the wavelengths emitted by the LEDs. Clearly, A' can not be denoted as absorbance and is no longer related in linear m a n n e r to the concentration of the chemical p a r a m e t e r .

3.2. Fluorescence Fluorescence can be used in optical sensors for detecting a chemical substance by using different approaches. Three main cases can be distinguished: 9 the substance under investigation is fluorescent; 9 the substance is not fluorescent, but can be labelled with a fluorophore; 9 the substance interacts with a fluorophore, causing a variation in the emission of fluorescence. Of particular interest, in the latter case, is the phenomenon known as fluorescence "quenching", in which the intensity of the fluorescence decreases as a consequence of the interaction with the substance (quencher) being tested, which can thus be detected [6]. This is one of the most used approaches in optical-fibre chemical sensing. Fluorophore (F) can interact with quencher (Q) at the ground state (static quenching), with the consequent formation of a nonfluorescent complex (FQ): F+Q ~

F+Q

(13)

or can interact with it at the excited state (dynamic quenching): F* +Q---> F +

O~

(14)

Then, due to the interaction with the quencher, the fluorophore comes back to the ground state, without the emission of fluorescence. In both cases, the relationship between the fluorescence intensity I and the concentration of quencher [Q] is:

932 I

1

I0

1+ K[Q]

(15)

where Io is the fluorescence in the absence of the quencher; K is a constant equal to the dissociation constant of equation (13) in the case of static quenching, and is the Stern-Volmer constant in the case of dynamic quenching (the Stern-Volmer constant is equal to the product kq.to, between the diffusion-controlled rate constant kq and the fluorescent lifetime to of the excited state F* in the absence of the quencher). A decrease in the intensity of the fluorescence may also be due to an energy transfer from fluorophore F* in the excited state to another molecule, acceptor A, whose absorption spectrum, modulated by the chemical species under investigation, overlaps the emission spectrum of the fluorophore. Therefore, fluorescence and absorption can be combined to detect a chemical parameter. In this case, the fluorescence intensity in presence of acceptor, I is given by: I

--=I-~] Io

(16)

where Io is the intensity of the fluorescence in the absence of the acceptor, and ~l is a term depending on the distance between fluorophore F and acceptor A. If other chromophores are present in the solution being tested, a decrease can be observed in the fluorescence, caused by the absorption of the excitation light (primary inner filter effect) or of the emission light (secondary inner filter effect) by these chromophores. It is obvious t h a t in this case, the previous equations are no longer valid, but t h a t corrective terms are necessary. In the case of dynamic quenching, it is more convenient to consider the timedependent decay [7]. In the presence of an interaction with the excited state, the lifetime of the fluorophore is decreased: the higher the concentration of the quencher, the greater the decrease in the lifetime. This is not the case for static quenching, in which the lifetime of the fluorophore is not affected by a change in the concentration of the quencher. Typical fluorescence decay times are in the range between 2 and 20 nsec, while phosphorescence decay times are in the 1 psec + 10 sec range. According to Stern and Volmer, the relationship between the decay time and the concentration is: T to

:

1

(17)

1 + Ksv[Q ]

where t and to are the decay times of the excited state of the fluorophore in the presence and in the absence of the quencher, respectively. Lifetime can be measured either in the time domain or in the frequency domain. In the former case, the fluorophore is excited with a narrow pulse and the

933 fluorescence decay is monitored. In the latter case, a modulated excitation is used: the fluorescence emission is still modulated at the same frequency, but is diminished in amplitude and phase shifted. The extent of the amplitude decrease and of the phase shift depends on the lifetime of the fluorophore. When ground-state reactions are involved, lifetime measurements can be performed if reagents and products Of the reactions are fluorescent species characterized by different decay times, as occurs in some acid-base reactions [8,9]. The advantage of this approach lies mainly in the fact that there is no more dependence on loading or photobleaching of the chemical transducer fixed at the end of the optical fibres, which is one of the greatest drawbacks of intensity-modulated chemical sensors. Moreover, no problems arise from possible fluctuations or drift in the source intensity or photodetector sensitivity which, on the contrary, heavily affect intensity of modulated sensors. 3.3.

Evanescent

wave field

Interaction of the chemical species with the evanescent wave field can be exploited for sensing purposes [10]. The electromagnetic field which propagates along the fibre inside the core extends also in the cladding region. The solution to Maxwell's equations shows that, in the presence of total internal reflection, a standing wave (called evanescent wave) exists in the cladding, propagates in the direction of the fibre axis and decays exponentially in the direction perpendicular to the core/cladding interface. The penetration depth of the evanescent wave is a key parameter for sensing purposes; it is the distance, from the cladding, at which the amplitude of the electromagnetic is decreased by a factor equal to 1/e and is expressed by the following formula (valid for a step-index fibre): dp=

~-0 1 2~n1 [sin2~_sin2~c~/2

(18)

Typical values of penetration depth are in the order of the utilized wavelength. For example, if n1=1.5 and n2=1.33 (aqueous medium) the minimum value of the penetration depth ( 3 = 90 ~ is about U5 and increases upwards by about 1 wavelength for angles 1~ greater than the critical angle. The penetration depth goes to infinity in correspondence of the critical angle. However this fact can be disregarded since, for angles close to the critical angle, the fibre is characterized by losses due to the scattering coming from the surface roughness. In practice, the evanescent wave field is limited to within few microns or less from the core surface. The efficiency of the approach depends on the fraction r of the optical power carried by the fibre which propagates in the cladding. The optical power carried by the core represents a background since it is not modulated by the analyte: higher this background is, lower are the performances of the evanescent wave sensors. This fraction is clearly high for monomode fibres (r>0.5) and, in multimode fibres, for the modes close to the cut-off condition (the so-called higher modes). In

934 multimode fibres the average power in the cladding, coming from the contribution of all modes, has to be considered. It has been shown [11] that for weakly guiding multimode fibres (nl -~ n2) the value of r is given by the following relationship: r = Pcla____dd= 4x/-2 Ptot 3V

(19)

For multimode fibres typical values of r are in the range of 0.01. Any changes in the microenvironment close to the fibre core which is ascribed to chemical species and modifies the evanescent field distribution can be used in the development of optical sensors. Modification of the refractive index of the cladding due to the penetration of the chemical species in the region close to the core is the sensing mechanism most followed. This approach is not characterized by selectivity, which can be reached by combining the evanescent wave analysis with the absorption or fluorescence coming either directly from the analyte or from the proper chemical transducer, which is located in the proximity of the fibre core. 4.

THE OPTODE

The term optode (or the more used word optrode: the first is the correct term, however, since its etymology comes from the Greek " 6 ~ ! 65o~") denotes the probe of the optical fibre chemical sensor: that is, the chromophore with its mechanical envelope, if existing, and the portion of the fibre in contact with this chemical transducer. The simplest case is when the investigated species has optical properties: in this case, development of the optode is reduced to the manufacturing of an optimized optical cell to be connected to the fibre. As explained above, if a chemical transducer is utilized, use is made of a special reagent, whose optical properties vary in accordance with the variation in the concentration of the p a r a m e t e r under examination. The reaction may be direct as in pH sensors in which the hydrogen ions react with an acid-base indicator or a fluorophore, causing a variation in the absorption or fluorescence, respectively. In other cases, the reaction which takes place in the optode gives an opticallydetectable compound as final product. An example is the detection of carbon dioxide, which is based on the detection of the pH of a carbonated solution, the acidity of which changes according to the quantity of CO2 dissolved therein. Enzyme-based sensors are included in this latter category. Detection is based on a selective conversion, catalysed by a special enzyme, of the p a r a m e t e r under investigation. Among the reagents or the products of the enzymatic reaction, there is a species which is optically detectable. In glucose sensors, the consumption of oxygen is detected when glucose reacts in the presence of glucose oxydase; in a penicillin sensor, instead, the production of hydrogen ions in the presence of the enzyme penicillase is exploited.

935 Two types of optode can be distinguished: - extrinsic optodes: the chemical transducer is immobilized on an external support, such as glass or polymeric matrix; in this case, a mechanical envelope is necessary for attaching the support to the tip-end of the fibre; - intrinsic optodes: the chemical transducer is directly immobilized on the fibre. This can be done at the fibre tip or along the fibre on the core: in this case, a compact and highly-miniaturized structure is attained, since the probe is practically the fibre itself. In an optode, a fast and reliable interaction with the investigated analyte m u s t be guaranteed. Moreover, if a chemical transducer is used, an appropriate immobilization process should be developed. Adsorption can play an i m p o r t a n t role in both cases, either by assuring a good exchange of the analyte between the probe and the external environment, or by offering a simple immobilization process for the sensitive reagent. 5.

A D S O R P T I O N AS AN E X C H A N G E M E C H A N I S M F O R T H E DETECTED ANALYTES

Adsorption of the chemical compounds under investigation on appropriate layers deposited along the fibre core can offer some advantages on traditional bulk procedures.

cladding

core

'

............................ " ';

I

modified cladding Figure 4. Sketch of the modified fibre for sensing purposes by means of evanescent wave technique.

Figure 4 provides a sketch of the fibre modified for sensing purposes. The deposited layer works as new cladding for the fibre, provided t h a t it has a refractive index smaller t h a n the refractive index of the fibre core. Since the interaction between the light carried in the fibres and the external environment occurs by means of the evanescent field, the sensing area is limited to a few microns from the core/cladding interface. If the new cladding has a thickness greater t h a n the penetration depth, the investigated volume is inside the cladding; therefore, the effect of interfering elements which may be found in the investigated sample can be

936 eliminated through the choice of an appropriate layer. A clear a p p a r e n t example is in the case of the analysis of species in aqueous samples, such as dissolved gases or nonpolar and organic compounds. If the m e a s u r e m e n t is performed with bare fibres, the evanescent field extends in the aqueous phase, and the presence of dissolved species cannot be detected, because of the strong absorption of the water. The choice of hydrophobic and organophilic layer can prevent the diffusion, close to the fibre core, of w a t e r molecules. These cannot interfere with the m e a s u r e m e n t , which makes it possible to detect the dissolved species, which is adsorbed inside the layer (Figure 5). 0 0.,,,.

0

~

0

oo o o,O,o , o Oo&, o, o.O, o d_.,.o o o o . o o

o

water

9~ a n a l y t e

modified cladding uJ

fibre c o r e

Figure 5. Adsorption of the analyte in the modified cladding: on the left side the exponential decay of the electric field outside the fibre core along with the penetration depth is shown.

For sensing purposes, the ideal layer should have the following characteristics: - a refractive index smaller t h a n the refractive index of the fibre core; - selectivity to the investigated compounds; - efficient adsorption and desorption processes, which would lead to a good reversibility of the sensor; - fast diffusion of the investigated compounds, which would m e a n a fast response time of the sensor. Modulation of the light carried by the fibre can be induced in two ways: 9 a change in the refractive index of the cladding after the adsorption, which gives rise to a change in the penetration depth of the evanescent field. The selectivity of the sensor is determined only by the selectivity of the cladding material, since the sensor responds to all the adsorbed species. Its a d v a n t a g e lies in the simplicity of the optoelectronic system, since no requirements are given regarding the choice of the wavelength; 9 light absorption in correspondence with the absorption bands of the analysed species. Higher selectivity can be attained, but a multiwavelength system should be used either in the NIR region, if the overtones of chemical compounds are utilized, or in the infrared region, if the vibrational bands are exploited.

937 Clearly, t e m p e r a t u r e changes can affect this type of m e a s u r e m e n t [12]. A variation in temperature has two effects: i) a change in the cladding density, due to the volume change, and ii) a change in the refractive indices of the core and of the cladding. The first effect gives rise to a change in concentration of the polymer cladding, while the second gives rise to a change in the penetration depth of the evanescent field. In any case, a change in the detected signal is observed, and a t e m p e r a t u r e compensation is necessary. 5.1.

Sensors based on refractive index changes

Since, in this case, the sensor is practically a refractometer, a requisite characteristic for the deposited layer is its high capability to change its refractive index with the adsorption of the investigated chemical compound. Refractive index changes of the order of 10 .5 can be detected with optical fibres [13]. As already pointed out, the selectivity of the sensor in this case depends only on the selectivity of the cladding; therefore its choice is crucial. The first system based on this principle was proposed for the detection of hydrocarbons in water [14]. An unclad fibre (core diameter 140 ~m) was silylated for a length of 80 cm. Different chemical reagents (RnSiX4-n) were used to realize the organophilic cladding: in particular, two silylating agents having the same R group (R = -ClsH37) but different leaving group (X=OCH2CH3, X=C1), namely, octadecyltriethoxysilane and octadecyltrichlorosilane. Other reagents used have the same leaving group (X=C1) but different R groups (R--Ph or R=n-C10H21), diphenyldichlorosilane and n-decyltrichlorosilane. An He-Ne laser (632.8 nm) was used as source, and the detector consisted of appropriate electronic circuitry with a simple pin photodiode. Laboratory tests on different contaminants indicated that the type of coating and the method of applying it affected the capacity for absorbing hydrocarbons. The detection limit was variable (400 rag/1 for p-xylene, 3 mg/1 for crude oil). Reversibility of the sensor was obtained by washing the coated fibre with acetone and methanol. Heteropolysiloxane polymers were also used for the detection of chemical vapours [15]: by incorporating different functions (amino, vinyl, glycidoxypropyl) inside the polymers, different sensitivities and selectivities were obtained for different chemical vapours. In the experimental tests, light from a laser diode 0~=670 nm) was coupled into two fibres (core diameter 600 pm). Angular excitation of the fibre was performed in order to excite the higher modes of the fibre and to increase the fraction of the optical power carried by the cladding. One of the fibres was stripped for a length of about 2.5 cm and coated with the vapour-sensitive coating; the other one was used as a reference, as it was insensitive to the vapour in contact with this fibre. Both fibres were inserted into a flow-cell. The signals coming out from the two fibres were detected by two photodiodes; their ratio resulted less sensitive to temperature changes. A detection limit of 100 ppm was obtained for toluene. The same system was tested for methane using a polyoxyethylene lauryl ether polymer as cladding material. A detection limit of 2% (in vol) in the air was obtained [13].

938 Selectivity still remains a problem when this approach is followed, but in principle the analysis of compound mixtures could be performed with several sensors, each coated with a different organophilic compound and properly calibrated. An appropriate processing system should be capable of discriminating the different compounds. On the other hand, selectivity is not a fundamental requirement for some applications. An example of this is the detection of hydrocarbon leakages from storage tanks, or remediation efforts. Klainer et al. developed an optical fibre sensor for the analysis of petroleum hydrocarbons, both in water and in vapour [16]. The new version of the system, called Petrosense CMS 5000, is at present distributed by Whessoe Varec [17]. Figure 6 shows a photo of the optical probe: the length of the probe is 25.5 cm, and the sensing area is limited to 2.5 cm of a glass fibre covered with a proprietary coating which attracts C6 and higher petroleum hydrocarbons. Table 2 describes the technical specifications as declared by the distributor.

Figure 6. Photo of the optical fibre probe for direct monitoring of hydrocarbons.

Table 2 Technical specifications of the hydrocarbon sensor Vapor Operating range

Water

0-20,000 ppm as TPH

Lower detection limit

<10 ppm as xylene

<0.1 ppm as xylene

Accuracy/Precision

+ 15% 12 seconds

_+ 10%

< 1 minutes 0 o _ 55oC

< 5 minutes 0 o . 55oC

Response time (initial) Response time (to 95%) Probe operating temperature range

12 seconds

939 5.2. S e n s o r s b a s e d on s p e c t r o s c o p i c a n a l y s i s In this case, the new cladding should be carefully chosen so as to avoid the overlapping between absorption peaks of its components and the absorption peaks of the detected compounds, since this would decrease the sensitivity of the measurement. The spectrophotometric investigations can be performed either in the IR region (2-20 pm) by means of fluoride or silver halide fibres, or in the NIR spectral region (1-2 pm), where silica fibres may be used. As previously outlined, in cases of analysis of aqueous samples, the choice of a highly hydrophobic layer makes it possible to avoid the influence of water which might thwart the detection of a small quantity of dissolved species: this is because the evanescent wave field which extends within the cladding, does not ,,see" the H20 molecules, the diffusion of which inside the cladding is not possible. This is particularly important in this spectroscopic analysis, since water is characterized in the NIR region by the broad H20 1.440-~m OH-stretch first overtone band and in the IR by broad vibrational bands around 2.6 ~m and 6.3~m. IR analysis make it possible to obtain high selectivity coupled with high sensitivity, since the 2-20 ~m spectral region is the most informative for detection of the various molecules. It is known as the "fingerprint" region, because it covers most of the absorption bands of the fundamental molecular vibrations. Proposed systems have been developed for the detection of hydrocarbons [18] and pesticides [19]. Different polymers were used for the cladding material: aliphatic polymers, such as polyethylene and polyisobutylene or polyvinylchloride. Due to the different IR absorption bands that characterizes the different hydrocarbons, the capability of the method of simultaneous detection of more than one species was demonstrated [20]. Silver halide fibres are preferable to chalcogenide fibres, since they are nonbrittle, and are quite flexible. Optical fibres are coupled to a Fourier transform infrared spectrophotometer [21] or to tunable lasers coupled with highly-efficient detection systems [22]. Detection limits of the order of few ppm or less have been obtained for different hydrocarbons. A detection limit of 300 ppb for tetrachloroethylene has been recently achieved [21]. The disadvantages are represented by the impossibility of utilizing long optical links, due to the high attenuation of chalcogenide fibres and to the very expensive optoelectronic system, which is neither compact nor transportable. Also, the near-infrared spectral region 1-2 pm provides some useful information, since at these shorter wavelengths, overtones or combination bands occur, presenting a much weaker intensity than the fundamental bands in the infrared region. This disadvantage is counter-balanced by the advantages coming from the fact that, at these wavelengths, quartz fibres are characterized by low attenuation. This means:

940 - very long optical link can be used (up to many kilometres); - optical fibre networks can be easily developed, and multiple detection in different sensing points is feasible. Commercially-available silica fibres with a siloxane cladding were used and tested for the first time by DeGrandpre and Burgess [23]. The same approach was then followed by Ache and coworkers for the detection of hydrocarbons [24,25]. The fibres (core diameter ranging from 200 ~m to 500 ~m) were coupled with a spectrophotometer and an halogen lamp was used as optical source. Good detection limits were reached (0.4 ppm for p-xylene and chlorobenzene, 0.9 ppm for toluene), proving the effectiveness of the approach. Moreover miniaturized optical fibre spectrophotometers in the NIR region (1-2 ~m) will be available on the market in the near future, facilitating in-situ measurements as required by new environmental regulations. The possibility of utilizing NIR spectroscopy and multivariate calibration techniques should make possible a selectivity in the analysis of samples containing more than one contaminant [26]. Absorption in the UV region caused by electronic transitions can be also utilized for the detection of hydrocarbons in air or water. Recently, a system has been proposed using silica fibres as optical link, a diode array spectrophotometer as detecting system, and a deuterium lamp as optical source [27]. Fibre attenuation and sensitivity of the array spectrophotometer limit the UV spectral analysis to 225 nm. Experimental tests on toluene, naphthalene and fluorene demonstrate the feasibility of the system with a detection limit of a few ppm for toluene and 30 ppb and 3 ppb for naphthalene and fluorene, respectively. 6.

A D S O R P T I O N AS AN I M M O B I L I Z A T I O N P R O C E S S FOR CHEMICAL TRANSDUCERS

When a chemical transducer is necessary, it may be immobilized by using different techniques on either the fibre itself or on a solid support. E n t r a p m e n t of the chemical transducer in polymeric membranes, in glassy networks (by means of a sol-gel procedure), in Langmuir-Blodgett films, or immobilization by means of covalent bonds and adsorption, are the main procedures followed in manufacturing an optode. Embedding in polymers ensures good compactness: the choice of the proper polymer, selective to the analyte and capable of avoiding leakage of the chemical transducer, is the critical point. Many evanescent wave-based sensors based on the approach of using the sol-gel process for depositing a sensitive layer on the core of the fibre have been described [28]. A porous glassy network can be obtained from the hydrolysis of an organometallic precursor, followed by condensative polymerization. The precursor can easily be doped with chromophores, which are thus entrapped in the glassyoxide network. A critical issue is the porosity of the final structure, which

941 determines both the capacity of maintaining entrapped the chemical transducer and the response time of the optode. The realization of Langmuir-Blodgett films with an incorporated chromophore is a very promising technique for the realization of optical fibre chemical sensors. The deposition of these films on the fibre core allows the manufacture of extremely compact and miniaturized evanescent-wave sensors. On the other hand, the low stability of these films is a very critical aspect, which has so far prevented the development of reliable sensors. The method of coupling the optically-sensitive reagent to a support by means of a covalent bond seems to be very promising, since it avoids losses of the bound chromophore. A much-followed method for covalently bonding the chromophore exploits a silylation process [29]. In the case of adsorption, the chromophore is immobilized on special supports, such as polymeric resins. Binding results from weak electrostatic or van der Waals interactions. This technique is very simple, but sometime presents the drawback of progressive leakage of the reagent from the support. All these techniques present advantages and disadvantages at the same time. The exact criteria for choosing a particular technique will depend on the application in question and on the materials chosen for realizing the optode. For instance, covalent immobilization is possible only if the reagent is characterized by a chemical functionality which is able to react with the chosen support, or is possible for imparting a particular functionality without blocking the reactivity to the analyte. On the other hand, the adsorption technique is a very generally-used method which, in principle, permits the immobilization of a great number of chemicallydifferent reagents. Clearly, the success of the immobilization strongly depends also on the nature of the chosen support. The simplicity which characterizes this immobilization technique also gives rise to the main problem associated with the method: due to the non-specific interactions involved, the problem of leakage of the reagent is often of vital importance. This, in turn, makes many of the probes realized in this manner unsuitable for many applications, in particular for medical ones. In addition, probes realized by adsorption are affected by the problems common to the probes realized using others techniques. In general, the ideal sensitive support should have the following characteristics: - stability of the system support and chromophore; - selectivity to the investigated analyte and, consequently, no interferences; - reversibility; - high sensitivity. An aspect to be stressed is that, in general, reagents coupled to surfaces do not react in identical manner to the free solution forms. A thorough investigation of the spectroscopic properties of the chemical transducers after the adsorption should be undertaken in order to better understand their behaviour [30-33]. Typical results are shown in Figures 7 and 8. Here the optical properties of an acid-base indicator, phenol red, before and after the immobilization on beads of

942 0.4 ~58

7. t. 8

0.3

0.2

0.1 6.~,8

0

i

400

i

i

500

600

700

)~ (nm) Figure 7. Absorption spectra of phenol red in aqueous solution (1.5-10 -5 M) for different pH values.

1.2] [ 0.9

0.6

0.3

O

i

400

500

600

i

i

700

~. (nm) Figure 8. Reflectance spectra of phenol red immobilized on XAD-2. A'=log(1/R), where R is the reflectance of the sample.

943 Amberlite XAD-2, a non-ionic styrene/divinylbenzene copolymer, are reported. Broadening of the pH working range and a shift of the pK towards higher values were observed for all the described indicators. At the same time, the disappearance of, or a heavy reduction in, the pH dependance of the absorption band of the acidic form is observed [33]. If the broadening of the working range and the pK shift can be easily explained by a distribution of the dye molecules over slightly different sites (in solution all the molecules are equivalent), the reason for the disappearance of the acidic band is not completely clarified. A plausible explanation for this phenomenon is that the XAD-2 beads are coated with two or more layers of the dye: the inner layers are essentially constituted by undissociated molecules (HIn), while ionic species (In-) predominate in the outer layer exposed to the solvent (Figure 9). Accordingly, the inner layers act as a reservoir for undissociated HIn species, so that a nearly constant concentration of HIn can be "seen" on the surface of the beads. Only when that reservoir disappears can a concentration decrease of the undissociated species be perceived. From the above example, it is obvious that noticeable changes in the optical properties of the chemical transducer can take place: their effects should be considered within the design of an adsorption-based optode.

/

Figure 9. Sketch of the multilayered distribution of the undissociated (HIn) and dissociated (In-) forms of the acid-base indicator.

944 An additional problem is the realization of the optode that is the coupling of the sensing support and the fibre. This problem has been solved by using different approaches: the simplest way is that the support is contained in a flow-cell and the analyte, in solution or in gas phase flows in the system. The optical fibres coupled to the flow cell bring the light from the source to the cell and from the cell to the detector [34-36]. These sensors can operate in two different ways: i) the concentration approach and ii) the kinetic approach. The former consists of measuring the modulated signal before and after the passage of a fixed volume of sample solution. In this case, the information is related to the signal change after equilibration. This volume is chosen to encompass the desired range of sample concentration. In the second approach the sensor is exposed to an unlimited volume of sample, and the readout of the m e a s u r e m e n t is made at a fixed time. In this case, the time chosen depends on the response time and the range of concentration of interest [35]. The reversibility of the sensor is very important. Whenever possible, regeneration can be achieved by passing a solution of an appropriate composition in the system. A configuration which overcomes the problem of reversibility has been proposed by Egorov et al. [37]: the renewable chemical sensor uses minute amounts of beads, on which the reagent is adsorbed and which are renewed before performing a new analysis. The system has been tested for the detection of chromium (VI), but could in principle be used for other analytes. However, if characterized by good sensitivity, an irreversible sensor can be utilized as a "warning" sensor [38]. A somewhat more complicated scheme foresees the use of a m e m b r a n e for coupling the sensing support to the fibre [43-46]. The membrane acts as a physical barrier for maintaining the sensing support in place without impeding the chemical sensing capacity of the reagent. In a different configuration, the chemical transducer is adsorbed along the core surface of a silylated fibre. A polymeric layer deposited on the modified fibre acts as a selective membrane for the analyte under investigation [47]. The choice of the membrane has a pronounced effect on the performances of the sensor; better selectivity can be achieved with an accurate choice of the polymer. On the other h a n d the response time is generally subject to considerable lengthening, since it will be governed mainly by the diffusion coefficients of the analyte in the polymer. Bacci et al. have proposed an optode for pH detection which does not make use of a m e m b r a n e (Figure 10) [48]. XAD-2 is used as solid support. After the adsorption of the acid-base indicator, the polymer beads were inserted in a stainless steel capillary (external diameter 1.8 mm), terminated by a threaded mirrored cap. Very small slits on the lateral surface of the capillary made possible a fast exchange with the external environment. Table 3 shows the main chemical species investigated with adsorption-based optical fibre sensors: the chemical transducers, the utilized supports and references are reported.

945 Table 3 M a i n chemical species i n v e s t i g a t e d with a d s o r p t i o n - b a s e d s e n s o r s a n d r e l a t e d references. LOD is t h e limit of detection a n d R is the reversibility Analyte

Support

Chemical Transducer

CI2

Nylon 66

F

XAD-2

o-tolidine, o-dianisidine, N,N-diethyl-p-phenylenediamine Alizarin Complexone/Ce(III)

F-

XAD-4

Calcein blue/ZrC12

HCN

XAD-7

02 02

Optical fibre Kieselgel

Chloramine-T, 4-Methylpyridine, Pyrimidine trio ne Ru(DIP)3C12

pH pH

XAD-2 XAD-2

pH

pH Pb(II)

Cr(VI) nl(III) Cd(II) Fe(III)

Ru(bipy)aC12 Bromothymol blue Phenol red

Working Range and/or LOD

R

0+5 ppm No 0.043 ppm 1.6+9.5x10 -4 M Al(III) 1.1xl0 4 M 2.6+42x10 S M ZrC12 2.6x10 -5 M HC1 lpl/1 No 0+800 Torr 2 Torr 0+200 Torr 1 Torr 6.5+8.0 pH 0.02 pH 0.02 pH

[36] [34] [44] [38]

Yes

[47]

N2

[43]

Yes Yes

[39] [48]

Yes Bromophenol blue, Chlorophenol red, Thymol blue, 3,4,5,6-Tetrabromophenolsulphonphtalein, Bromothymol blue Bromocresol green 6.00+9.00 pH Yes XAD-4 0.05 pH 3+100x10 -7 M HC1, Diphenyldithiocarbazone XAD-4 1.0x10 -s M Citrate H2NOH 1,5-Diphenylcarbohydrazide 0.01+1 ppm Polysorb 35 ppb MP1 1.3+4.0x10 -5 M F XAD-2 Eriochromo cyanine R 1.0x10 -~ M 8- Hydroxy- 7-io do quinoline-5- sulfonic 2+700x10 -5 M KI DEAE8.0x10 -s M acid Sephadex 8+ 100x 10 .6 M 2,4-Dinitroresorcinol HC1 XAD-7 8.0x10 -6 M XAD-4

Ref.

.

[40]

[41]

[35] [37] [45] [421 [46]

946

XAD-2 optical fibres

threaded mirrored cap

Y 8 mm stainless steel

slits

capillary Figure 10. Adsorption-based optode for pH detection. 7. C O N C L U S I O N S Adsorption is a chemical process which finds application in the development of optical fibre sensor for chemical parameters. The critical issues are different, depending on whether the process is related to the interaction between the investigated analyte and the optode, or is utilized as immobilization process for the chemical transducer. One is a lack of selectivity (in the former case) and the other is possible leakage (in the latter). These problems are very important. The fact that only one of the proposed adsorption-based sensors is available on the market, the hydrocarbon sensor from Whessoe Varec, should not be completely ascribed to the poor reliability of the adsorption process for sensing purposes, but also to the typical difficulties associated with the manufacture of an optode. Most of the probes described in the literature are hand-made, no matter which chemical compound is investigated. Manufacture of the probe in large numbers, which is a fundamental point in the consideration of an industrial production of a sensor, is never considered in the design of the probe and in the choice of the materials used. This aspect, which is not at all trivial, is probably the main reason why the appearance of m a n y optical fibre chemical sensors on the m a r k e t has so far been prevented.

REFERENCES 1. 2.

O.S. Wolfbeis (ed.), Fiber Optic Chemical Sensors and Biosensors, Voll. I-II, CRC Press, Boca Raton - Florida - USA, 1991. A.G.Mignani and F.Baldini, Rep. Prog. Phys., 59 (1996) 1.

947 3. A.M. Scheggi and F. Baldini, Intern. J. Optoelectr., 8 (1993) 133. 4. J.I. Peterson, S.R. Goldstein, S.R. Fitzgerald and D.K. Buckhold, Anal. Chem., 52 (1980) 864. 5. A.J. Guthrie, R. Narayanaswamy and D.A. Russell, Analyst, II (1988) 457. 6. P. Yuan and D.R. Walt, Anal. Chem., 59 (1987) 2391. 7. H. Szmacinski and J.R. Lakowicz, Sensors Actuat. B, 29 (1995) 16. 8. M.E. Lippitsch, S. Draxler and M.J.P. Leiner, Proc. SPIE, 1796 (1992) 202. 9. S. Draxler and M.E. Lippitsch, Sensors Actuat. B, 29 (1995) 199. I0. B.D. Mac Craith, Sensors Actuat. B, 11 (1993) 29. II. D. Gloge, Appl. Optics, I0 (1971) 2252. 12. G.L. Klunder, J. Burck, H.J. Ache, R.J. Silva and R.E. Russo, Appl. Spectr., 48 (1994) 387. 13. M. Archenault, H. Gagnaire, J.P. Goure and N. Jaffrezic-Renault, Sensors Actuat. B, 8 (1992) 161. 14. F.K. Kawahara, R.A. Fiutem, H.S. Silvus, F.M. Newman and J.H. Frazar, Anal. Chim. Acta, 151 (1983) 315. 15. C. Ronot, M. Archenault, H. Gagnaire, J.P. Goure, N. Jaffrezic-Renault and T. Pichery, Sensors Actuat. B, II (1993) 375. 16. S.M. Klainer, J.R. Thomas and J.C. Francis, Sensors Actuat. B, II (1993) 81. 17. Whessoe Varec, Inc., 10800 Valley View Street, Cypress, California 90630. 18. R. Krska, E. Rosenberg, K. Taga, R. Kellner, A. Messica and A. Katzir, Appl. Phys. Lett., 61 (1992) 1778. 19. J.E. Walsh, B.D. MacCraith, M. Meaney, J.G. Vos, F. Regan, A. Lancia and S. Artjuschenko, Analyst, 121 (1996) 789. 20. R. Krska, K. Taga and R. Kellner, Appl. Spectr., 47 (1993) 1484. 21. M. Jakusch, B. Mizaikoff, R. Kellner and A. Katzir, Sensors Actuat. B, 38-39 (1997) 83. 22. J.F.Kastner, M. Tacke, A. Katzir, B. Mizaikoff, R.Gobel and R. Kellner, Sensors Actuat. B, 38-39 (1997) 163. 23. M.D. DeGrandpre and L.W. Burgess, Appl. Spectr., 44 (1990) 273. 24. J. Burck, J.-P. Conzen and H.J. Ache, Fresen. J. Anal. Chem., 342 (1992) 394. 25. J.-P. Conzen, J. Burck and H.J. Ache, Appl. Spectr., 47 (1993) 753. 26. J.-P. Conzen, J. Burck and H.J. Ache, Fresen. J. Anal. Chem, 348 (1994) 501. 27. G. Schwotzer, I. Latka, H.Lehmann and R. Willsch, Sensors Actuat. B, 38-39 (1997) 150. 28. B.D. Mac Craith, C.M. McDonagh, A.K. McEvoy, T. Butler and F.R. Sheridan, Sensors Actuat. B, 29 (1995) 51. 29. F. Baldini and S. Bracci, Sensors Actuat. B, II (1993) 353. 30. M. Bacci, F. Baldini and A.M. Scheggi, Anal. Chim. Acta, 207 (1988) 343. 31. R.T. Andres and F. Sevilla III, Anal. Chim. Acta, 251 (1991) 165. 32. S. Motellier and P. Toulhoat, Anal. Chim. Acta, 271 (1993) 323. 33. M.E.J. Baker and R. Narayanaswamy, Sensors Actuat. B, 29 (1995) 368. 34. R. Narayanaswamy, D.A. Russell and F. Sevilla III, Talanta, 35 (1988) 83. 35. W.A. de Oliveira and R. Narayanaswamy, Talanta, 39 (1992) 1499.

948 36. 37. 38. 39. 40.

41. 42. 43. 44. 45. 46. 47. 48.

S.A.Momin and R. Narayanaswamy, Anal. Chim. Acta, 244 (1991) 71. O. Egorov and J. Ru~iSka, Analyst, 120 (1995) 1959. S.M. Jawad and J.F. Alder, Anal. Chim. Acta, 246 (1991) 259. G.F Kirkbright, R. Narayanaswamy and N.A. Welti, Analyst, 109 (1984) 15. G. Boisd~, F. Blanc and X. Machuron-Mandard, Intern. J. Optoelectr., 6 (1991) 407. S. Motellier, M.H. Michels, B. Dureault and P. Toulhoat, Sensors Actuat. B, 11 (I 993) 467. J. Lu and Z. Zhang, Analyst, 120 (1995) 453. O.S. Wolfbeis, L.J. Weis, M.J.P. Leiner and W.E. Ziegler, Anal. Chem., 60 (1988) 2028. D.A. Russell and R. Narayanaswamy, Anal. Chim. Acta, 220 (1989) 75. M. Ahmad and R. Narayanaswamy, Anal. Chim. Acta, 291 (1994) 255. N. Mal~ik and P. ~aglar, Sensors Actuat. B, 38-39 (1997) 386. E. Singer, G.L. Duveneck, M. Ehrat and H.M. Widmer, Sensors Actuat. A, 41-42 (I 994) 542. M. Bacci, F. Baldini, F. Cosi, G. Conforti and A.M. Scheggi, in: Optical Fiber Sensors, H.J. Arditty, J.P. Dakin and R.Th. Kernsten (eds.), Springer-Verlag Berlin, Heidelberg, 1989, 425.

Adsorption and its Applications in Industry and EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

949

A d s o r p t i o n h e a t p u m p s : a n e w w a y for e n e r g y s a v i n g a n d C F C s replacement G. Cacciola and G. Restuccia National Council of Research, Institute for Research on Chemical Methods and Processes for Energy Storage and Transformation, S. Lucia sopra Contesse, 98126 Messina, Italy

1.

INTRODUCTION

In the field of h e a t i n g and cooling production, research activity is mainly addressed towards finding alternative solutions to vapour-compression heat pumps, since these machines use valuable energy (electricity) as p r i m a r y energy and polluting refrigerants, namely chlorofluorocarbons (CFCs) which are dangerous both for ozone depletion and the green house effect [1]. Among the new systems recently proposed, adsorption machines present m a n y characteristics which make t h e m a good techno-economic alternative to vapourcompression machines. In fact, adsorption machines can use environmental friendly refrigerants and efficiently medium-low t e m p e r a t u r e h e a t (100-200~ as p r i m a r y energy. In addition they have no moving parts and a simple regulation of energy production in response to load request can be realised [2,3]. As is well known, gas adsorption phenomena are strictly correlated to energy transfer and t r a n s f o r m a t i o n and it is likewise well known t h a t adsorption p h e n o m e n a are regulated by t e m p e r a t u r e and pressure. Taking into account these properties and combining the endothermic desorption with the exothermic adsorption processes in closed cycles it is possible to realise an adsorption heat pump whose external effect is equal to t h a t obtained with a vapour-compression machine using an inverse Carnot cycle. Adsorption h e a t pumps, even though having the above mentioned advantages, still have some critical aspects which can be s u m m a r i s e d in three m a i n points: - non-continuous (intermittent) operation; - low h e a t transfer efficiency between the t h e r m a l fluid and the solid adsorbent bed; - operating pressure far from atmospheric level. These problems have made the first prototypes uncompetitive with respect to traditional machines. Thus research activity in this sector has focused on solving such difficulties and very positive results have been obtained in the last years [4-7].

950 2.

OPERATING PRINCIPLES

Adsorption machines are mainly composed of three principal components, as shown in Figure 1: the adsorber reactor (R) which contains the solid adsorbent embedded in a suitable h e a t exchanger, the evaporator (E) and the condenser (C). In the adsorber reactor both the desorption and the adsorption processes occur at different times and at different t e m p e r a t u r e and pressure conditions. An adsorption machine works at four t e m p e r a t u r e levels, one for each component but in practice, as further described, it is designed so t h a t there are three t e m p e r a t u r e working conditions, high, medium, low (Th, Tin, T1).

@ Qra+Qa

Qri + Qd

9

/~.

R

"

user

. .....

,

",..

Qc ~ .......

,

....:..,.

.

,:.

~

E

.

i

Figure 1. Scheme of an adsorption machine: C - condenser; E - evaporator; R - adsorbent bed.

The t h e r m o d y n a m i c cycle for h e a t p u m p i n g can be divided in two phases: the charge and the discharge. During the first phase, the solid adsorbent is dried and the refrigerant fluid here adsorbed evaporates t h a n k s to the h e a t furnished to the system at high t e m p e r a t u r e Th. The desorbed vapour flows into the condenser (C) where it condenses at medium t e m p e r a t u r e Tm and, consequently, releases useful heat. At the beginning of this first phase the incoming h e a t into the adsorber increases both t e m p e r a t u r e and vapour pressure till the condenser pressure is overcome, then the vapour, previously in equilibrium on the solid, flows towards the condenser t h a n k s to the small difference in pressure now created between adsorber and condenser. The vapour m o v e m e n t makes the desorption process possible under isobaric conditions. Then, during this charge phase external h e a t (Ori + Qd) is furnished to the system at high t e m p e r a t u r e Th,

951 by means of a thermal vector fluid flowing into the heat exchanger embedded inside the solid adsorber (R), and, at the same time, heat (Qc) is released from the system at medium temperature Tm, from the condenser. The discharge phase, which consists of the vapour adsorption process on the solid adsorbent, allows the heat coming from the low temperature (T1) to be transferred to the medium temperature source (Tm). This phase starts with the cooling of the solid adsorbent and the consequent decreasing of the pressure, till it reaches the evaporator vapour pressure. The evaporator (E), in fact, is maintained at the low temperature T1 thanks to an external fluid which transfers heat coming from the low temperature source. As during the charge phase, the small difference in pressure between evaporator and adsorber creates the non-equilibrium conditions which allow the refrigerant vapour to flow from evaporator to the solid adsorbent bed where it is adsorbed. During the cooling of the solid and afterwards the adsorption process, thermal energy (Ora + Qa) is produced and immediately externally transferred through the heat exchanger embedded into the solid adsorbent. In this way the evaporator continuously produces vapour which is isobarically adsorbed in the solid bed. This phase represents the useful effect of the system, in fact, heat (Qe) is taken away from a low temperature and released to a medium temperature source, realising in this way a heat pump or cooling effect, similar to that obtainable with an inverse Carnot cycle.

3.

R E F R I G E R A N T S AND A D S O R B E N T S

One of the most important elements of any heat pump and refrigeration system is the refrigerant, since the working conditions and the system compatibility with the environment principally depend on it. Generally speaking, the refrigerant requirements are: high latent heat per unit volume and good thermal stability; furthermore, it must be non-polluting and non-flammable. From a technical and safety point of view its vapour pressure must be near atmospheric level, in the temperature range between 263 and 353 K. Table I shows some characteristics of several refrigerants that have been proposed [8].

Table 1 Properties of some refrigerants Normal Boiling Point K

Heat of Vaporisation J/g

Name

Formula

Ammonia

NH 3

239

1368

Water

H20

373

2258

Methanol

CH3OH

338

1102

952

Water, ammonia and methanol present high latent heat, but the working p r e s s u r e s are not in a f a v o u r a b l e r a n g e (Figure 2). The c h a r a c t e r i s t i c s of t h e s e r e f r i g e r a n t s are r e p o r t e d in Table 2. In conclusion, since from the t h e r m o d y n a m i c point of view the choice of a r e f r i g e r a n t is not unequivocal, it could d e p e n d on technological criteria; n e v e r t h e l e s s the best r e f r i g e r a n t m u s t be still sought. The s u i t a b l e a d s o r b e n t s are porous m a t e r i a l s t h a t should adsorb large a m o u n t s of r e f r i g e r a n t fluids in v a p o u r p h a s e a n d h a v e t h e following c h a r a c t e r i s t i c s : wide c o n c e n t r a t i o n c h a n g e in a small t e m p e r a t u r e range, r e v e r s i b i l i t y of a d s o r p t i o n process for m a n y cycles, low cost, good t h e r m a l conductivity. Table 2 C h a r a c t e r i s t i c s of some r e f r i g e r a n t s Ammonia Methanol - toxic - f l a m m a b l e in some concentrations - not compatible with copper - high o p e r a t i n g p r e s s u r e - good l a t e n t h e a t - t h e r m a l l y stable - non polluting

Water

- toxic - flammable - not compatible with copper at high temperature - u n s t a b l e beyond 393 K low p r e s s u r e good l a t e n t h e a t

perfect, except for very low o p e r a t i n g p r e s s u r e at low p r e s s u r e does not oxidise copper a n d only p a r t i a l l y s t a i n l e s s steel - not s u i t a b l e for cold climate zone

10000

1000

100

10 9 methanol 9 water

0

20

40

60

80

100

T (~

Figure 2. Comparison of the vapour pressure of some refrigerants.

953 The most interesting classes of solid adsorbents are: zeolites, activated carbons and silica gels. For each class, different kinds of materials are available with different characteristics; generally, within a given class, the smaller the pore diameter the higher the adsorption energy and the regeneration temperature. Specific research has been carried out to study the influence of solid adsorbent porosity modification on the adsorption properties [9]. The most studied gas-solid pairs come from the combination of the cited refrigerants and adsorbents; these are shown in Table 3, where, for each application, the most suitable pairs are indicated.

Table 3 Suggested applications for adsorption pairs Freezing Refrigeration Air conditioning (T< 253 K) Zeolite- NH 3

(T ~ 273 K) A.Carbon- CH3OH

(T= 278-288 K)

Space heating (T ~ 333 K)

Zeolite- H20 A.Carbon-NH 3

A.Carbon-NH 3

Zeolite- H20

A.Carbon-CH3OH

Silica g e l - H20

Silica gel- HeO

From a technical point of view the general opinion among people who work in this sector is that zeolite-water, silica gel-water, activated carbon-methanol and activated carbon-ammonia appear to be the most suitable pairs for cooling and heating [10]. Nevertheless, other considerations must be taken into account in order to evaluate the most suitable pair. In fact, the use of ammonia is forbidden or limited in several countries such as J a p a n and Germany, while the same refrigerant is widely diffused in USA. Furthermore the expanding of southern Europe air conditioning market offers a big opportunity for the zeolite-water pair whose performance reaches the highest values in mild climate conditions [11].

4. T H E R M O D Y N A M I C CYCLES The basic thermodynamic cycle is shown in Figure 3 where the gas-solid equilibrium curves (isosteres) are also represented. The isosteric line ~ represents an equilibrium curve where the weight ratio of adsorbate (ma) and dry adsorbent (mo) is constant, then ~ = ma / mo. Following the same two steps mentioned in the "operating principles" paragraph, it is possible to describe a complete thermodynamic cycle. This starts from the equilibrium point F which describes the process when the solid

954 p (mbar) 1000 4

~=20

16

1

1 Pc lO0-

i-

/

I

Pe 10-?

Oa/

I

1l

//,4/

0 Te

100

Ta=Tc

' 200 Td

T (~

Figure 3. Adsorption heat pump thermodynamic cycle in a Clapeyron diagram for a zeolitewater pair: ~ - adsorbate uptake (wt. %).

adsorbent is first heated up isosterically (F-G), and the amount of energy furnished to the system is: Q ri = ~; C(T)dT Where C(T) is the heat capacity of the global mass of dry adsorbent (moCpo) and refrigerant fluid adsorbed in it (maCpa). Therefore: C(T) = moCp o + maCPa that can be easily transformed in: C(T) = m oCpo (1 + ~ Cpa Cpo ) When temperature and pressure overcome TG and pc the desorption starts and develops isobarically while the desorbed water condenses at Tc in the condenser. During this part of the cycle, external heat is furnished to the system to make the desorption phase possible (G-H). This amount of energy depends both on the heat of adsorption of the refrigerant vapour as well as on the heat capacity of the global mass of dry adsorbent and refrigerant fluid adsorbed in it. This last

955 quantity varies during the desorption phase, the following equation must then be considered: Qd = mo ~12[AH(~)+ Cpo (1 + ~ CPa

Cpo)~

where AH(~)is the adsorption/desorption enthalpy, which depends on ~. Contemporaneously, at the condenser the following energy quantity is available: Qc = mo(~l-~2)r(pc) where r(p) is the latent heat of condensation of the refrigerant fluid at pressure po The second step of the cycle starts with an isosteric cooling phase during which the solid adsorbent is cooled down from TH to TI (H-I line), and the energy released from the system is: Qra = ~~ C(T)dT When temperature and pressure reach T~ and pe the isobaric adsorption occurs up to TF; at the end the system is again at the starting point of the cycle (F in Figure 3). At the same time the refrigerant evaporates in the evaporator. The heat that can be released from the system is described with a formula similar to that of the desorption phase:

Qa -- mo ~2L[AH(~)+ Co(l + ~ CPa Cpo)~ In the evaporator the heat that must be furnished to the system at temperature Te is:

Oe = mo(~l- ~2)r(Pe) Adsorption machine components operate at four temperature levels (Td, Ta, To, Te); nevertheless to release heat from the system at the same temperature, the thermodynamic cycle is designed in order that Ta and Tc are equal. Figure 4 and Figure 5 respectively show the energy-exchange of the adsorption heat pump at the three temperature levels and the cyclic evolution of the energy exchanges. From the combination of these figures the intrinsic intermittent nature of the machine is evident; in fact, the useful effect is not continuously available.

956 It is necessary to specify that the useful effect in the heat pump mode is the sum of Qra, Qa and Qc when they are obtained simultaneously. In the cooling mode (refrigeration machine), the useful effect is the heat extracted at low temperature Qe.

Qri+Qd

-

Za=Zc

F

~III,IIIIQI~IIII~ ~ ) Q r a + Q a

HIIIIIIIIIIII~I~

l?o It Fiure 4. Input and outputenergies for an adsorption heat pump.

Figure 5. Sequence of the cycle phases.

The above described adsorption system, composed of a single adsorbent bed alternatively connected to the condenser or evaporator, was initially proposed for energy storage systems with heat pump effect in discharge phase. But the discontinuity of the useful effect, especially for cold production, makes this system commercially unsuitable. Later, two-adsorbent-bed machines were proposed [12,13] in order to operate in a continuous mode. In the two-beds configuration, two cycles have been proposed in order to improve the overall efficiency of the system. With the first an internal heat recovery is realised between the two reactors which allows the heat coming from one reactor during isosteric cooling to be used for heating the other reactor. With the second a regenerative system is realised using a heat wave swing inside both reactors. The practical scheme of the two-reactors system in case of internal heat recovery is shown in figure 6; it is composed of two adsorbent beds (reactor R1 and R2), one condenser (C), one evaporator (E), one expansion valve (Ev). The components which allow energy exchange with the system are: - a boiler (B), which gives thermal energy at high temperature; generally 10-20~ higher than Ta. The heat is transferred from the boiler to the heat exchanger embedded in the solid adsorbent by means of a fluid vector. The characteristic of this fluid depends on the working temperature, for example in case of a zeolite adsorber which needs temperatures higher than 150~ special oils are used;

957 - a h e a t exchanger high t e m p e r a t u r e fluid vector/water, by m e a n s of which t h e r m a l energy is furnished to the user, when the two beds are alternatively cooled down; - one h e a t exchanger inside the condenser to transfer h e a t to the user (heat pump mode) or to waste it to the a m b i e n t (chiller mode); - one h e a t exchanger inside the evaporator to transfer h e a t coming from the low t e m p e r a t u r e source.

I B [IQ,. f '

' ~C ,

user

'~"

I

f

"" "

c(ii

R1

" :-:'::3,'

R2

E~I~Q',b_":":

" Y

e

Figure 6. Scheme of a two-reactor adsorption heat pump: B - boiler; C - condenser; E - evaporator; Ev - expansion valve; Rl, R2 - adsorbent bed.

In addition, with a couple of three-way valves the connection between different components is realised. In detail, the valves allow to connect the condenser and evaporator alternately with the two beds during the desorption and adsorption phases. The t h e r m o d y n a m i c cycle of a double bed adsorption machine is shown in Figure 7. The adsorbent beds operate the same cycle but out of phase, so t h a t the h e a t recovery is obtained by transferring the h e a t from bed R1 (HI-II-L1 in Figure 7a) to bed R2 (F2-G2-M2 in Figure 7b) until t e m p e r a t u r e equilibrium is obtained. Afterwards, additional heat, supplied by the boiler, is needed to complete the desorption from adsorbent R2 (M2-H2 in Figure 7b); simultaneously the adsorption process develops in bed R1 (LI-F1 in Figure 7a), and it produces heat t h a t is supplied to the user. During the next two steps, which complete the cycle,

958 the roles of the beds are exchanged: bed R2 is cooled, supplying heat to bed R1 and afterwards to the external user. The energies involved with each step of the cycle are schematically represented in Figure 7. Qin is the heat that must be supplied to the adsorbent bed during isobaric desorption; it is used for heating the bed and for desorption. Qc is the heat released in the condenser. Qr is the heat exchanged between the two beds during recovery steps. It consists mainly of isosteric heating/cooling of the solid adsorbent and its adsorbate content (sensible heat); a small amount of adsorption/desorption heat is also exchanged (I-L and G-M in Figure 7). Qout is the heat produced during the adsorption process, which is extracted from the bed. Oe is the heat that must be supplied to the evaporator for adsorbate evaporation.

In (p)

bed 1 G! Mi

bed 2

In (p) HI

G~ ,M2IQi" "

Q, F

FQ/out,~ L~ I1

-

~

,

H2

~ L2 I2

b v

a

In (p)

G

T

M l 1 Qin

b

In (p)

H~

G~, ,M2

T

/ H2

F2Qout l~ L2 I2

L l I1 L v

d

T

c

T

Figure 7. Thermodynamic cycle of a two-reactor adsorption heat pump operating with internal heat recovery.

4.1. R e g e n e r a t i v e

cycles

In order to describe the heat wave propagation system, let consider the scheme shown in Figure 8. The system is composed of two reactors, always in connection each other by a thermal vector flowing through them. The boiler and the external heat exchanger, connected to the circuit between the reactors, allow the extra input and output energy needed to heat and cool the reactors [14-16]. To explain the concept of such methodology, each reactor can be ideally represented as formed by n-isothermal beds, which can be consecutively heated

959 up by m e a n s of the e n e r g y r e l e a s e d from the c o r r e s p o n d e n t bed of the o t h e r r e a c t o r w h i c h is cooling down. As can be seen in F i g u r e 9, the t h e r m a l vector fluid acts also as a n a c c u m u l a t o r which allows to t r a n s f e r all the e n e r g y from one r e a c t o r to the other. D u e to the h e a t losses a n d the irreversibilities, this is only a n ideal b e h a v i o u r , in fact, a n e x t e r n a l c o n t r i b u t i o n from the boiler (to h e a t one r e a c t o r to t h e m a x i m u m t e m p e r a t u r e ) a n d from the h e a t e x c h a n g e r (to cool the o t h e r reactor) is needed.

lXl C ( ,

i I i _Qu Ev Qa- l I

, : - .... ,-...~ ..... -.-.,.-.....-.---.,. | ,"...L~

i)E

,

I

: . . . . . " ' . " ".

,.: .... -:...,: . . . . . .

"" "

" "'" ~ "':

I

- ~ : - :

E

: :" "1--""

.......::.--......,

R1

l

;;

~:. :.. . .,:: .:.- .. , .-.- .:-.: -..: - ::-:-.c .. : -.,

Qout

R2

9 reversible pump

|Qin

Figure 8. Scheme of the regenerative adsorption heat pump: B - boiler, C - condenser, E - evaporator, E• - heat exchanger, R], R2 - adsorbent beds.

In o r d e r to o b t a i n a b e h a v i o u r as close as possible to the ideal, the s y s t e m should h a v e t h e following characteristics: - low t h e r m a l i n e r t i a of the bed including the h e a t exchanger; - low t h e r m a l r e s i s t a n c e of the a d s o r b e n t bed in the direction p e r p e n d i c u l a r to the t h e r m a l vector flow a n d h i g h r e s i s t a n c e in the parallel direction; - low flow r a t e of t h e t h e r m a l vector fluid. W i t h t h e s e h y p o t h e s e s , the t h e r m a l effect of the h e a t or cold t h e r m a l vector fluid i n c o m i n g into a bed h a s a step b e h a v i o u r . As a c o n s e q u e n c e also the a d s o r b e n t bed h a s a similar t e m p e r a t u r e profile with r e s p e c t to time, as s h o w n in F i g u r e 9.

960

.4,-a 0

Tmax

_l:

> ~

0 o m

Train

-L.. -L

~X

rma x

Train wX

bed R l

JI Jl J Jl JI J I . 1,i 11 11 11 ]1 1n1 ~

I L i~t

2

3

4

4

3

boiler

"

Exchanger

I II

I

Ir

2

1

rL

bed R 2

"L ~L t4~L t3~L t2~L

tn

t.

Tmax~orO >

Zmin

x

o

t~ ~

t. ~ -

t4~

t3~

t2~

Tmax

~"~ ~

Train

o -'~ m

x~

Figure 9. Bed and oil temperatures in the reactors of a regenerative heat pump. In these conditions the bed should behave like a succession of small isothermal beds. The temperature profile as a function of time is thus moving along the whole bed obtaining a heat wave moving from one to the other edge of the bed. The whole process is described in the following with reference to the scheme of the system of Figure 8 and the thermodynamic cycle of Figure 10 [17].

961

1000

p (mbar)

r

16 8

Pc

100

4

Pe

l0

'Tv TG 100

',

'.200

T (~

Figure 10. Thermodynamic cycle of the regenerative adsorption heat pump.

Let us consider that initially bed R1 is at temperature TG and condensation pressure pc while bed R2 is at temperature TI and evaporation pressure pe. During this phase the thermal vector fluid is circulating clockwise, it is heated up in the boiler at TH and enters into the bed R1 at the same temperature; due to the heat transferred its temperature at the bed exit is Tc (see Figure l la). At the same time the thermal vector fluid leaves the heat hxchanger Ex and enter into the bed R2 at temperature TF where it cools a small portion of bed. The thermal vector fluid leaves the bed R2 at temperature TI and goes towards the boiler where it is heated up to TH. Continuing this process, in the beds the temperature fronts will move, as a function of time, from one side to the other of the beds. In the bed R1 the heat wave is between the temperature TH and Tc while in the bed R2 it is between TF and TI. When the temperature at the end of bed R2 is lower than TI and at the end of bed Ra is higher than TG, the phase is considered completed and they are in the conditions represented by point H and F of Figure 10 respectively. When the second phase starts the fluid flow changes direction to counterclockwise by reversing the pump rotation or by a reversing valve. The heat transfer fluid heated in the boiler at TH will thus flow into the bed R2 heating it, from right to left in Figure l l-b. Due to the initial conditions of this phase (TF and pe) and to the large temperature difference between the fluid and the bed, the first part of it will be heated at TH while the oil leaves the bed at

962 YH

a)

t2i t.\ \. \. k t~ flOW direction

TO

TF

/ / tl

~t~ TG TF

0

/

f

F

/

/t~

/ t2

/

J

t. flow direction 0

L

T.

YH

b)

f

t,\

"~~to.\_

Tl

t \ ~___

YG TF

flow direction 0

TH

0

L

TH

0

TH flowdirection -~

d)

5~

TH 1 f Zi /tfn /t21 /tItf / o TG j "/ "/ TF J ~ flow direction

T! ~ -~l -~l ~ C) tn\ t~', tl', TG 2~ ' TF C _.~_ flow \ ___._~ direction

Tj

flow direction

TF L

0

~

TGT,

T, F~to/f q 1 / ' ~/t

TG

TG

Tr

TF 0

L bed R~

n

flOWdirection 0 bed R2

Figure 11. Behaviour of the heat vector temperatures in the two beds during the four phases of the heat pump cycle: to - initial time of the phase; tn - final time of the phase; L - bed length.

temperature TG. At the same time the heat vector enters bed R1 at TF and leaves at TI. During this short phase the pressure in the bed R2 increases from pe to pc due to the desorption of the initial part of the bed. This produces an adsorption

963 on most of the bed with its heating from TF to TG. Similarly, the cooling of the initial part of the bed R1 creates a re-adsorption of refrigerant and thus a decrease in the pressure from pc to pe. This allows a further desorption of most of the bed with a decreasing of the t e m p e r a t u r e from TH to TI as shown in Figure 10. At the end of this phase bed R1 is at pressure Pe, its initial part - right side in Figure l lb - is at t e m p e r a t u r e TF while the most part is at temperature TI. On the contrary, the reactor R2 is at pressure pc and t e m p e r a t u r e TG except the initial part - right side in Figure 1 lb -that is at TH. These conditions are equal to the initial conditions but with opposite heat vector flow direction and reactor out of phase. Continuing with the same flow direction, the cycle continues similar to the first phase but with adsorption in R1 and desorption in R2 (Figure 1 l-c). The last phase is similar to the second one. As described, the thermodynamic cycle reported in Figure 10 differs from that described in Figure 3 because the heating and cooling of the bed (F-G and H-I) are not isosteric processes but are closer to isothermal processes. This cycle allows a much higher heat recovery t h a n t h a t operating with "isothermal beds" previously described and is called regenerative cycle, as a result higher COPs are also obtained.

5. THERMODYNAMIC PERFORMANCES On the basis of the expressions of the various amounts of energy involved in the cycle as reported in the previous paragraph, here are presented the performances of a two-reactor adsorption machine operating in heat pump or in chilling mode. The Coefficient Of Performance of a heat pump operating without internal heat recovery is the ratio of the useful heat produced to t h a t supplied for its operation. In detail, it is the energy obtained from the isosteric cooling of the bed, the adsorption heat and the condensation heat over the heat for desorption and isosteric heating of the bed:

coP h =

Qra + Q a +Qc Qri +Qd

When the adsorption system is working as chiller, the useful effect is the energy entering into the evaporator, which comes from the ambient being cooled, while the input energy is the same as the heat pump mode operation:

COPc

Qe

Qri + Qd

In those cycles with internal heat recovery between the two beds, this amount of energy is mainly for isosteric heating and cooling but the recovery phase could

964 continue even during part of adsorption/desorption step. For example, in figure 7 the heat recovery also includes the G-M and I-L parts of the cycle in order to reach the equilibrium in temperature of both beds. The extension of the recovery phase is dependent upon the operating conditions of the machine that determine the shape of the thermodynamic cycle. Since the energy recovered can be calculated after definition of the thermodynamic cycle, there is therefore no general formula for its calculation. On this basis the coefficient of performance for heat pump and chilling operation are respectively: COP h =

COPc =

Qra + Q a + Q c - Q r Qri + Q a - Q r

Qe Ori +Od - O r

where Qr is the energy recovered. The increase in performance obtainable by heat recovery cycles with respect to non recovery ones is immediately evident by comparing the cooling COP formulae. For the heating COP the increase is due to the fact that the same amount is subtracted from both terms of the ratio, so that the COP increases as much as the amount of the heat recovered. Many research groups have calculated the performance rate in terms of COP both for heating and cooling applications by means of different mathematical models for basic cycles as for heat recovery and regenerative cycles [18-21]. These evaluations have been made for different adsorber-adsorbate pairs. In the case of machines working with uniform distribution of temperature reactors, the thermodynamic analysis has evidenced that the COP depends on the maximum temperature of the desorption phase, on the evaporation and condensation temperature and on the minimum temperature of the adsorption phase, taking into account that this last parameter is generally equal to the condensation temperature. Figures 12 and 13 show the COP behaviour as a function of the maximum temperature cycle and of evaporation temperature respectively. The curves reported in the figures are related to zeolite-water pair in heat pump mode; Figure 13 also shows the influence of the condensation temperature on COP. In both figures the performances obtainable with a basic cycle and with a two-bed internal heat recovery cycle are reported. Analogously, Figures 14 and 15 show the performances of the system for cooling applications. One important characteristic that can be deduced from these results, especially for Figures 13 and 15, is that the performance slightly decreases when the so called temperature lift increases. This means that when the temperature difference between condenser and evaporator increases, the COP of the system

965

does not drastically decrease as happens in commercialised u s i n g f o r e x a m p l e a m m o n i a - w a t e r p a i r [22] ( F i g u r e 16).

1.7

absorption

machines

-

1.6

with recovery

1.5 0 o

9 Te=15~ 9 Te=10~ 9 Te= 5~

1.4 1.3

without recovery

1.2

-

100

l

150

9 Te=

0~

l

I

200

250

Td(~

Figure 12. C O P vs. desorption temperature in heat pump operation for zeolite-water pair: Te - evaporation temperature; Tc = Ta = 55~

1,8

-

1.7

-

with recovery

1.6 9 1.5 cD 1.4

9 9 9 9

without recovery

1.3 1.2

Tc=45~ Tc=50~ Tc=55~ Tc=60~

f

i

i

5

10

15

Te(~ Figure 13. COP vs. evaporation temperature in heat pump operation for zeolite-water pair: Tc - condensation temperature; Td = 200 ~

966

0.9

-

with recover,

0.8 0.7 0.6 9

0.5 0.4

~ o

9 Tc=25~ 9 Tc=30~

ut recovery

0.3

9 Tc=35~

0.2

9 Tc=40~ I

0.1

100

50

I

I

I

150

200

250

Td(~

Figure 14. COP vs. desorption temperature in cooling operation for zeolite-water pair: Tc - condensation temperature; Tc = 5 ~

1 0.9 with r

0.8

e

c

o

v

~

0.7

0.6

9 o 0.5

without recovery Te= 4~

0.4

9

0.3

9 Te=

6~

Te= 8~ 9 Te=10~

9

0.2 0.1

I

20

25

I

I

I

30

35

40

Tc(~

Figure 15. COP vs. condensation temperature in cooling operation for zeolite-water pair: Te - evaporation temperature; Ta = 200 ~

The above m e n t i o n e d property, more evident for zeolite-water pair, a s s u m e s an economic relevance for a chilling machine; in fact in this case the condensation a n d adsorption h e a t can be r e l e a s e d to the a m b i e n t by a simple h e a t e x c h a n g e r i n s t e a d of a cooling tower.

967

0,9 0,8

0 0,7 0,6

(z-w) without recovery

~I 0,4 25

I

I

I

I

l

30

35

40

45

50

AT

Figure 16. Cooling COP vs. temperature lift for zeolite-water (z-w) and for ammonia-water pair: AT = Tc- Te.

The performance related to other gas-solid pairs has been calculated and results are published in specialised literature [23]. In Figures 17, 18 and 19 several results have been grouped together and the COP relative to activated carbon-methanol and activated carbon-ammonia are reported. Also in these cases

0.6

-

Ta=35~

0.5 0.4 O 0.3 0.2

cry

0.1 [ I I

70

90

V I

110

!

9 w i t h o u t recovery I

I

130

150

Td(~

Figure 17. Cooling COP vs. desorption temperature for the activated carbon-methanol pair: Ta-adsorption and condensation temperature; Te = -10~

968

t h e t h e r m o d y n a m i c performances both for the basic and for internal h e a t recovery cycles are shown. Performances reported in Figure 19 refer to an evaporation t e m p e r a t u r e not o b t a i n a b l e w i t h water (Te - - 1 0 ~

0.7

-

0.6 0.5 0.4 0.3 0.2

I

0.1

I

.w!t recovery

9 wit outrecover i

70

90

110

150

130

Td(~

Figure 18. Cooling COP vs. desorption temperature for the activated carbon-methanol pair: Te - evaporation temperature; Ta = 35 ~ - adsorption and condensation temperature.

0.6

0.5

0.4 9 0.3 9 methanol 0.2

9 ammonia

0.1 70

I

i

I

I

90

110

130

150

Td(~

Figure 19. Cooling COP vs. desorption temperature for activated carbon-methanol and activated carbon-ammonia pairs: Ta = Tc = 25 ~ Te = -10 ~

969

All the above mentioned results refer to ideal systems where the heat capacity of the heat exchanger is neglected. If this heat capacity is suitably considered the calculated COP decreases even if an internal heat recovery is realised. This is due to the inert h e a t capacity t h a t subtracts heat from the useful process of the system (the adsorption process). In order to calculate the influence of the heat capacity of the heat exchanger a new p a r a m e t e r has been defined [24]:

Z

._.

m m 9C P m m a "Cp a

where m = metal and a = adsorbent. The K p a r a m e t e r indicates the increment of the inert h e a t capacity due to the presence of the heat exchanger. In figure 20 the COP related to zeolite-water pair is represented for different values of K. As can be seen when the K coefficient reaches values near 2 the COP decrease is near 10%. A further performance improvement with respect to the internal heat recovery system can be obtained with a regenerative cycle [25]. A comparison of COPs calculated for the two cases in the same operative conditions is shown in Figures 21 and 22 for heating and cooling applications respectively.

2

-

1.5

9 r~

0.5 9 heat pump 9 refrigeration I

0.5

I

I

I

i

1.5

2

K

Figure 20. COP vs. metal/adsorbent heat capacity ratio for zeolite-water pair.

970 1.8 Te =

1.7

5~

T c = 38~ 1.6 1.5 -" 9

1.4

1.3 9 zeolite-water, basic cycle 1.2 9 zeolite-water, h e a t recovery cycle

1.1

9 z e o l i t e - a m m o n i a , r e g e n e r a t i v e cycle 0.9 50

I

I

I

I

100

150

200

250

Td(~

Figure 21. Heating COPs of basic, heat recovery and regenerative cycles.

0.8

-

0.7

_

Te =

5~

Tc = 38~ 0.6 0.5 9

0.4 0.3 9 z e o l i t e - w a t e r , b a s i c cycle 0.2 9 z e o l i t e - w a t e r , h e a t r e c o v e r y cycle 0.1 9

50

z e o l i t e - a m m o n i a , r e g e n e r a t i v e cycle

I

I

I

I

100

150

200

250

Td(~

Figure 22. Comparison of cooling COP for basic, heat recovery and regenerative cycles.

971

6. P R O T O T Y P E S A N D T E S T S The first adsorption machines prototypes for heating and cooling applications were realised with tube and fin heat exchangers. This type of typology was chosen recognising the low heat transfer property of the solid adsorbent and consequently to attempt increasing the heat exchanger surface at the solid adsorbent side. In these prototypes the adsorbent material, generally in beads or in little extrudates of cylindrical shape, was positioned among fins in order to completely fill the empty space. With this type of adsorbent bed it was shown that the limiting factor to obtain good performance was the slow heat transfer inside the reactor bed. This was evidenced by the low value of the global heat transfer coefficient, in fact, using a tube and fin heat exchanger, values of UA between 60 and 360 W/~ were obtained [26]. With a fin and tube heat exchanger embedded in the adsorber bed Douss [27] developed experimental tests with an activated carbon-methanol single reactor. The COP of the system was experimentally measured and its behaviour is shown in Figure 23 with respect to the maximum desorption temperature. Comparing the COP calculated with experimental tests and with the thermodynamic model the effect of heat losses from the reactor, neglected in the model is evident.

0.55

-

0.5

0.45 9 0.4

0.35

0.3 50

I

I

70

90

,

I

I

110

130

T d (~

Figure 23. Cooling COP vs. desorption temperature for carbon-methanol pair: Te = 5~ Tc = 15 ~

972 Utilising the same type of adsorber reactor, Douss [28] realised a two-bed system with internal heat recovery using the zeolite-water pair. Measured values are: COPe = 0.67 and COPh = 1.56 for cooling and heat pump application respectively. The same values measured for a machine without heat recovery are respectively COPe = 0.52 and COPh = 1.38. These results were obtained under the following operative conditions: Th = 200~ Ta = 60~ Tc = 40~ Te = 25~ From another experimental test on a different tube and fin heat exchanger [29], values of COPh = 1.2 in heat pump operation were obtained under the following operative conditions: Th = 200~ Ta = 70~ Tc = 60~ Te = 5~ From the experimental activity on this type of machine, where the solid adsorbent bed is realised in non consolidated form, many limits of the system were evidenced: 9 high adsorber weight and in particular high ratio between metal mass of heat exchanger and adsorbent mass; 9 large volume of the system; 9 low heat transfer and consequently long length of the cycle. Regarding the high weight ratio of metal and solid adsorbent, this evidently affects the COP of the machine. In fact, when the equivalent heat capacity of the metal + solid adsorbent increases, the COP decreases, as mentioned above and shown in Figure 20. Obviously higher values of metal weight derive from the necessity to improve the heat transfer, thus a compromise must be found and in this evaluation the thickness of fins and tubes can play an important role. In the above mentioned cases the specific power both per unit mass and per unit volume of adsorbent were low. In fact, values of 20-50 W/kg of adsorbent in case of chiller operation and of 50-100 W/kg of adsorbent in heat pump mode were obtained. For the above mentioned reasons research activity has been addressed to the improvement of heat transfer both through the modification of morphologic and thermo-physic characteristics of the solid adsorbent and through the design of the solid adsorbent-heat exchanger assembly. The reasons for these types of proposed solutions can be found if the heat transfer inside the adsorbent bed is accurately analysed. Referring to Figure 24, the global heat transfer coefficient between the thermal vector fluid flowing inside the heat exchanger and the solid adsorbent is represented by the following expression: U

.__

1 1 + s m + 1 + sa hf )~m hw )~a

where: 1/hf is the thermal resistance at the tube wall on the heat vector fluid side; Sm/)~mis the thermal resistance in the tube, whose wall is thick Sin, negligible with respect to the others; 1/hw is the thermal resistance at the tube wall on the adsorbent side; Sa/)~a is the thermal resistance in the adsorbent.

973

adsorbent

heat exchanger wall

Sa Sm

Figure 24. View of a section of the heat exchanger- adsorbent connection.

Regarding the improvement of the heat transfer many studies are in progress, all of which are focused on the adsorber bed and on the thermal contact between the solid adsorber and the heat exchanger surface. In fact, the parameters that control the heat transfer in the adsorber bed are the equivalent thermal conductivity of the solid adsorbent, ~.a, the wall thermal resistance between the heat exchanger surface and the adsorbent bed 1/hw, the wall thermal resistance between the heat transfer fluid and the heat exchanger surface 1/hf. The equivalent thermal conductivity of the solid bed is generally very low because the adsorbent is a porous material and also because it is available in grain or pellet forms, thus the thermal contact between the particles is very poor. The wall heat transfer between the solid adsorbent and the heat exchanger depends on the contact surface which must not be limited to points, as it is when a pellet bed is used. Finally, the wall heat transfer at the thermal fluid side depends on many factors, such as the geometry of heat exchanger, the type of fluid and the flowdynamic regime that must be used in order to have a suitable regenerative effect in the system. In the case of pellets in tube and fin heat exchanger, the contribution of each resistance to the global heat exchanger coefficient is shown in Figure 25. As can be seen, by reducing both 1/hw and Sa/~a the global heat transfer coefficient can reach the values needed for an important increase in the heat transfer and consequently in the specific power of the system. The most interesting practical solutions presented in literature to improve the heat transfer inside the adsorber bed (~.a and hw) follow two different methodologies: a) preparation of solid adsorbent composites with high thermal conductivity and external geometry well adapted for good contact with the heat exchanger wall, in this case the adsorbent composite is over few centimetres thick; b) preparation of solid adsorbent thin layer (less t h a n 5 mm) connected with the heat exchanger wall by physical or chemical method in order to guarantee a negligible thermal resistance between the heat exchanger wall and the adsorber bed.

974 0.09 0.08

~

0.07 0.06

CD r

0.05

4.a

9~t3 " 0.04 0.03 0.02 4.a

0.01 0

,

, Sa/~

a

,

1/hw

,

I

1/hf

Figure 25. Values of the thermal resistances for a pelletised bed 10 mm thick.

Obviously both methodologies could have negative consequences that must be carefully considered and minimised in order to have an economical system. In fact, following the first methodology the mass transfer through the thick bed could decrease the adsorption kinetic. With the second methodology the ratio between the metal and adsorber weight could be high with a negative effect on the performance of the machine and on the cost of the total system. Following the first methodology two types of compact composite materials were developed at CNRS in Orsay (France) [30,31]. The first consists of adsorbent zeolite powder (13 X) mixed with metallic foam and compressed. Using copper and nickel foams good equivalent thermal conductivity of the bed were obtained. The second consolidated adsorbent, is obtained by compression of 13 X-zeolite powder mixed with expanded natural graphite. In this case an anisotropic solid bed is obtained thanks to the characteristics of the oriented graphite layers. This anisotropic behaviour is also useful for the practical application of efficient regenerative cycles. Using the second methodology a research activity has been developed in the framework of the European Communities Joule program. A team comprising University of Delft (NL), Gastec Institute (NL) and CNR-TAE Institute (I) developed a solid adsorption bed consisting of zeolite directly synthesised on the heat exchanger wall [32]. In this case zeolite A was used because of its very hydrophilic nature and different types of metals (oxides) such as stainless steel, nickel, copper, titanium, etc. as well as different configurations/forms were considered. The zeolite layer thickness was optimised taking into account two opposite necessities, the first related to a practical realisation of a homogeneous

975 layer with good mechanical and chemical properties and the second connected with the performance and cost of the adsorption machine. A zeolite layer of I mm synthesised on a stainless steel tube of 0.5 mm wall thickness was finally chosen. Another type of thin layer bed has been realised at CNR-TAE Institute in Messina (Italy), which consists of pressed zeolite bricks (3-5 mm thick) bound with aluminium hydroxide and chemically tied with a thin aluminium sheet [33]. In this case the equivalent thermal conductivity of the bed is very close to the thermal conductivity of the zeolite crystals and a high hw coefficient is foreseen. At the Institute for Technical Thermodynamics, Aachen (Germany), a zeolite layer has been deposited on a metallic matrix-like honeycomb structure by a spray method and pore forming materials such as melamine or tartaric acid have been added to the mixture of zeolite, distilled water and binder [34], in order to improve the mass transfer. The organic components are completely removed during the dry process of the zeolite, increasing the macro pore fraction of the material. A thickness of 3.5 mm was obtained in this way.

7.

CONCLUSIONS

Adsorption processes are at the heart of several potential new energy technologies which can find suitable applications in the domestic sector as reversible heat pumps, and in the industrial sector as refrigerating systems and heat transformers using industrial waste heat as the primary energy source. They can also be used for technologies to be applied in the transportation sector, for automobile air conditioning or for food preservation in trucks. The use of environmental friendly materials like zeolite as adsorbent and water as refrigerant make this system very acceptable in any sector. Obviously the concrete possibility of economic and efficient machines depends on the solution to the problems still open today. From a thermodynamic point of view the recent study offers good opportunities for the regenerative cycle especially for the so called "heat wave propagation system". Nevertheless major efforts must be done to make this cycle practical. Especially taking into account that the low properties of heat and mass transfer of the adsorption bed must be contemporaneously solved without any COP decrease. The solution of these two problems mainly depends on the design of the solid adsorbent bed and its connection with heat exchanger. Finally, it is always important to look for other adsorption materials that can be efficiently used in a cycle whose highest desorption temperature is lower than 100~ in order that solar energy or traditional boilers can be used. Even the partial solution of the problems still unsolved to date, will allow the adsorption systems to find an important role as new energy technologies with high efficiency and low pollution.

976 NOMENCLATURE

A C Cp AH h K m p r(p) s Q T U

surface area heat capacity specific heat enthalpy of adsorption heat transfer coefficient heat capacity ratio mass pressure latent heat of condensation thickness thermal energy temperature global heat transfer coefficient thermal conductivity adsorbate uptake

Subscript a adsorption, adsorbent c condensation, cooling d desorption e evaporation f fluid h high, heating in input 1 low m medium, metal o dry adsorbent out output r recovery ra isosteric cooling ri isosteric heating u useful w wall

REFERENCES

1. 2. 3.

F. Meunier, Proc. Symp.: Solid sorption refrigeration, Paris (1992) 44. G. Cacciola, G. Restuccia and N. Giordano, Heat Recovery Systems and CHP, 10, 5/6, (1990) 499. D.I. Tchernev and D.T. Emerson, ASHRAE Transactions, 14 (1988) 2024.

977 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

J.J. Guilleminot, J.B. Chalfen and A. Choiser, Proc. Int. Absorption Heat Pump Conference, New Orleans January 19-21, AES-vol.31 (1994) 401. M. Pons, D. Laurent and F. Meunier, Applied Thermal Eng. Vo1.16, 5 (1996) 395. R. Lang, T.Westerfeld, A. Gerlich and K.F. Knoche, Adsorption, 2 (1996) 121. L. Pino, Yu. Aristov, G. Cacciola and G. Restuccia, Adsorption, vol. 3, 1 (1996) 33. R.E. Critoph, Solar Energy, 41, 1 (1988) 21. H.L. Turner, Proc. 3rd Int. Work. on Res. Activ. on Ad. Heat Pumps, Graz (1990). G. Cacciola and G. Restuccia, Heat Recovery Systems & CHP, 4 (1994) 409. G. Korn, Condizionamento dell'aria Riscaldamento Refrigerazione, 37, 3 (1993) 307. P. Maier-Laxhuber, M. Rothmeyer and G.Alefeld, Fluid Engineering Int. Conf. on Energy Storage, Energy Storage for Energy Management, Stratford-upon-Avon UK, (1983) 205. F. Meunier, Heat Recovery Systems & CHP, 5, 2 (1985) 133. S.V. Shelton, W.J. Wepfer and D.J. Miles, J. Energy Resources Technology, 112 (1990) 69. D.I. Tchernev and J.M. Clinch, Proc. Int. Gas Research Conference, Tokyo, (1989) 44. A. Aittom~iki and M. H~irkSnen, Proc. Symp. Solid sorption refrigeration, Paris, (1992) 56. G. Cacciola, G. Cammarata, L. Marletta and G. Restuccia, La Termotecnica, 4 (1993) 81. G. Cacciola and G. Restuccia, Int. J. of Refrigeration, 18, 2 (1995) 100. R.E. Critoph and H.L. Turner, CEC-British Gas Int. Workshop Absorption Heat Pumps, London, (1988) 89. N. Douss, Etude experimentale de cycles a cascades a adsorption solide, Ph.D Thesis, University of Paris VII (1988). T. Zanife, F.Meunier and J.B. Chalfen, Proc. of XVIII Int. Congress of Refrigeration, Montreal, (1991) 1041. M. Engler, G. Grossman and H.-M. Hellmann, Int. J. Refrig., 20, 7 (1997) 504. R. E. Critoph, Carbon, 27, 1 (1989) 63. G. Cacciola, G. Restuccia and G.H.W. van Benthem, Munchen Discussion Meeting on "Heat Transfer Enhancement by Additives", October (1994) 81. S.V. Shelton, W.J. Wepfer and D.J. Miles, Heat Recovery Systems & CHP, 9, 3 (1989) 233. G. Restuccia, V. Recupero, G. Cacciola and M. Rothmeyer, Energy the Int. J., 13, 4 (1988) 333. N. Douss and F. Meunier, Heat Recovery System & CHP, 8, 5 (1988) 383. N. Douss, F. Meunier and L. M. Sun, Ind. Eng. Chem. Res., 27 (1988) 310.

978 29. A. Brigandi, G. Cacciola, G. Maggio and G. Restuccia, 48 ~ Congresso Nazionale ATI, Taormina, Italy, (1993) 737. 30. J.J. Guilleminot, A. Choisier, J. B. Chalfen, S. Nicolas and J.L. Reymonet, Proc. Symp.: Solid sorption refrigeration, Paris (1992) 215. 31. J.J. Guilleminot, J. B. Chalfen and F. Poyelle, Proc. Int. Congress of Refrigeration, vol. IV (1995) 261. 32. G. Cacciola, G. Restuccia, A. Muller, J.C. Jansen and H. van Bekkum, Int. Absorption Heat Pump Conference 96, Montreal (1996) 609. 33. L.Pino, G. Cacciola, G. Restuccia and M. Fascetto Sivillo, Metodo per la preparazione di lamine di alluminio ricoperto con strati di materiale adsorbente, Italian Patent No. RM 96A000448 (1996). 34. K.F. Knoche and R. Lang, Proceedings of Munchen Discussion Meeting 94 on "Heat Transfer Enhancement by Additives", Munchen (1994).

Adsorption and its Applications in Industryand EnvironmentalProtection Studies in Surface Science and Catalysis,Vol. 120 A. Dabrowski (Editor) 9 1998 Elsevier Science B.V. All rights reserved.

Adsorption and adsorptive-type separations protection: A Bibliography (1967-1997)

979

for e n v i r o n m e n t a l

M. S. Ray School of Chemical Engineering, Curtin University of Technology, GPO Box U1987, P e r t h 6845, W E S T E R N AUSTRALIA

Abstract This chapter provides a bibliographic listing of published journal papers from 1967 to 1997 concerned with adsorption, adsorptive separations, and related topics with particular emphasis on environmental protection. The bibliography provides a quick and easy, but comprehensive, reference source. The references are taken from the fifty most important chemical engineering journals, but do not include papers from the chemistry journals, or books, or conference/symposium series. A listing of the journals surveyed is included at the end of this chapter. The references are arranged chronologically (and then alphabetically by first author surname) within the following subject groups:

PSA and Cyclic @stems, and Applications Liquid-Phase Adsorption Ion Exchange, Chromatography, and Related Separations The following topics, concerned with industrial applications, are included in a separate bibliography in Volume 1: Fundamental Principles of Adsorption, Theory and Models,

Adsorption Design Methods and Data, Adsorbent Materials. A general bibliography of the chemical engineering journal literature from 1967-1988 has been published by the author [ 1], and can provide access to a wider range of topics. An earlier bibliography [2] provides access to the literature prior to 1967. A complete bibliographic listing of the chemical engineering journal literature from 1989 to 1997 (with subsequent sixmonthly updates) is available on a CD-ROM database [3].

Keywords: Adsorptive separations; environmental applications; PSA; parametric pumping; ion exchange; chromatography. References 1. M.S. Ray, Chemical Engineering Bibliography (1967-1988), Noyes Publications, New Jersey, USA (1990). 2. K. Bourton, Chemical and Process Engineering Unit Operations: A Bibliographical Guide; MacDonald and Co., London (1967). 3. Engineering and Applied Science Database on CD-ROM (includes CHERUB Chemical Engineering Database, 1989-1997), published by Royal Melbourne Institute of Technology, Australia. Full details are available from the author.

980

PSA AND

CYCLIC

SYSTEMS,

AND APPLICATIONS

1967 Alexis, R.W., Upgrading hydrogen via heatless adsorption, Chem. Eng. Prog., 63(5), 69-71 (1967). Levinson, S.Z., and Orochko, D.l., Staged countercurrent multisectional contactors for continuous adsorptive treatment of petroleum products, Int. Chem. Eng., 7(4), 649-654 (1967). Silbernagel, D.R., New uses for molecular sieves in olefin plants, Chem. Eng. Prog., 63(4), 99-102 (1967). 1968 Wilhelm, R.H.; Rice, A.W.; Rolke, R.W., and Sweed, N.H., Parametric pumping, Ind. Eng. Chem. Fund., 7(3), 337-349(1968). 1969 Dellosso, L., and Winnick, J., Mixed-gas adsorption and vacuum desorption of carbon dioxide on molecular sieve, Ind. Eng. Chem. Process Des. Dev., 8(4), 469-482 (1969). Fair, J.R., Sorption processes for gas separation, Chem. Eng. (N.Y.), 14 July, 90-110 (1969). Kazakova, E.A.; Khiterer, R.Z., and Bomshtein, V.E., Purification of exhaust gases from nitric acid plants, Brit. Chem. Eng., 14(5), 667-668 (1969). Ozawa, Y., Regeneration of coked catalyst in adiabatic fixed beds at lower temperatures, Ind. Eng. Chem. Process Des. Dev., 8(3), 378-383 (1969). Pigford, R.L.; Baker, B., and Blum, D.E., An equilibrium theory of the parametric pump, Ind. Eng. Chem. Fund., 8(1), 144-149 (1969). Ponder, T.C., Adsorption systems for alkane recovery, Hydrocarbon Process., 48(10), 141 (1969). Stewart, H.A., and Heck, J.L., Pressure swing adsorption, Chem. Eng. Prog., 65(9), 78-83 (1969). Sweed, N.H., and Wilhelm, R.H., Parametric pumping (The stop-go method), Ind. Eng. Chem. Fund., 8(2), 221231 (1969). 1970 Broughton, D.B.; Neuzil, R.W.; Pharis, J.M., and Brearley, C.S., Parex process for recovering paraxylene, Chem. Eng. Prog., 66(9), 70-75 (1970). Chi, C.W., and Wasan, D.T., Fixed-bed adsorption drying, AIChE J., 16(1), 23-31 (1970). DiNapoli, R.N., Adsorption systems for LNG gas pretreatment, Hydrocarbon Process., 49(12), 93-96 (1970). Jenczewski, T.J., and Myers, A.L., Separation of gas mixtures by pulsed adsorption, Ind. Eng. Chem. Fund., 9(2), 216-221 (1970). Michelson, K.J., and Price, C.D., Molecular sieve pre-drying, Chem. Eng. Prog., 66(5), 73-74 (1970). Petukhov, S.S.; Tumanov, A.I., and Trokhina, G.A., Combined process for removal of impurities from air using synthetic zeolites, Int. Chem. Eng., 10(3), 405-409 (1970). Tan, V.A., et al., Continuous fluid bed adsorber with centrifugal separation of the solid phase, Brit. Chem. Eng., 15(10), 1295-1296 (1970). 1971 Alexis, R.W., and Dailey, L.W., Molecular sieve driers, Hydrocarbon Process., 50(5), 145-148 (1971). Baker, B., and Pigford, R.L., Cycling-zone adsorption: Quantitative theory and experimental results, Ind. Eng. Chem. Fund., 10(2), 283-292 (1971). Barnebey, H.L., Activated charcoal in the petrochemical industry, Chem. Eng. Prog., 67(11), 45-48 (1971). Bahere, C.A., Feed-gas drying with molecular sieves, Hydrocarbon Process., 50(8), 126-128 (1971). Gupta, R., and Sweed, N.H., Equilibrium theory of cycling-zone adsorption, Ind. Eng. Chem. Fund., 10(2), 280283(1971). Turnock, P.H., and Kadlec, R.H., Separation of nitrogen and methane via periodic adsorption, AIChE J., 17(2), 335-342(1971). van der Vlist, E., Oxygen and nitrogen enrichment in air by cycling zone adsorption, Sep. Sci., 6(5), 727-732 (1971). 1972 Brooking, H.L., and Walton, D.C., The specification of molecular sieve adsorption systems, Chem. Eng. (Rugby, Engl.), January, 13-18 (1972). Butts, T.J.; Gupta, R., and Sweed, N.H., Parametric pumping separations of multicomponent mixtures: An equilibrium theory, Chem. Eng. Sci., 27(5), 855-866 (1972).

981

Chen, H.T.; Rak, J.L.; Stokes, J.D., and Hill, F.B., Separations via continuous parametric pumping, AIChE J., 18(2), 356-361 (1972). Harper, C., A molecular sieve plant as the main helium purifier for a high temperature, gas cooled nuclear reactor, Chem. Eng. (Rugby, Engl.), July, 271-276 (1972). Patrick, R.R.; Schrodt, J.T., and Kermode, R.I., Thermal parametric pumping of air-sulfur dioxide, Sep. Sci., 7(4), 331-344 (1972). 1973 E1-Rifai, M.A.; Saleh, M.A., and Youssef, H.A., Steam regeneration of a solvents adsorber, Chem. Eng. (Rugby, Engl.), January, 36-38 (1973). Miller, W.C., Adsorption cuts SO2, NOx and Hg, Chem. Eng. (N.Y.), 6 August, 62-63 (1973). Otani, S., Adsorption separates xylenes, Chem. Eng. (N.Y.), 17 September, 106-107 (1973). Various, Removal of organics by adsorption (topic issue), Chem. Ind. (London), 1 September, 823-831 (1973). 1974 Bond, A., Compressed-air drying systems, Process Eng. (London), March, 52-53 (1974). Bourgeois, S.V.; Groves, F.R., and Wehe, A.H., Analysis of fixed-bed sorption: Flue gas desulfurization, AIChE J., 20(1), 94-103 (1974). Chen, H.T.; Lin, W.W.; Stokes, J.D., and Fabisiak, W.R., Separation of multicomponent mixtures via thermal parametric pumping, AIChE J., 20(2), 306-310 (1974). Chen, H.T.; Park, J.A., and Rak, J.L., Equilibrium parametric pumps, Sep. Sci., 9(1), 35-46 (1974). Gregory, R.A., Comparison of parametric pumping with conventional adsorption, AIChE J., 20(2), 294-300 (1974). Meir, D., and Lavie, R., Continuous cyclic zone adsorption, Chem. Eng. Sci., 29(5), 1133-1138 (1974). Wankat, P.C., Cyclic separation processes, Sep. Sci., 9(2), 85-116 (1974). Weaver, K., and Hamrin, C.E., Separation of hydrogen isotopes by heatless adsorption, Chem. Eng. Sci., 29(9), 1873-1882 (1974). 1975 Cummings, W.P., Save energy in adsorption, Hydrocarbon Process., 54(2), 97-98 (1975). Nandi, S.P., and Walker, P.L., Carbon molecular sieves for concentration of oxygen from air, Fuel, 54(3), 169178(1975). Raghuraman, K.S., and Johansen, T., Hydrogen by PSA process, Processing (Sutton, Engl.), October, 10-11 (1975). Wang, L.K., et al., Treatment of industrial effluents by activated carbon, J. Applied Chem. Biotechnol., 25,475502 (1975). Wankat, P.C., Multicomponent cycling zone separations, Ind. Eng. Chem. Fund., 14(2), 96-102 (1975). Wankat, P.C.; Dore, J.C., and Nelson, W.C., Cycling zone separations, Sep. Purif. Methods, 4(2), 215-266 (1975). 1976 Camero, A.A., and Sweed, N.H., Separation of nonlinearly sorbing solutes by parametric pumping, AIChE J., 22(2), 369-376 (1976). Chen, H.T.; Rastog, A.K.; Kim, C.Y., and Rak, J.L., Nonequilibrium parametric pumps, Sep. Sci., 11(4), 335346 (1976). Dote, J.C., and Wankat, P.C., Multicomponent cycling zone adsorption, Chem. Eng. Sci., 31(10), 921-928 (1976). Hsu, H.H.; Wang, K.B., and Fan, L.T., Gaseous pollutant removal by single bed cyclic adsorber with synchronous thermal contact, Sep. Sci., 11(2), 109-132 (1976). Nandi, S.P., and Walker, P.L., Separation of oxygen and nitrogen using 5A zeolite and carbon molecular sieves, Sep. Sci., 11(5), 441-454 (1976). Rice, R.G., Progress in parametric pumping, Sep. Purif. Methods, 5(1), 139-188 (1976). Sun, Y.C., and Killat, G.R., Adsorption for vapor control, Hydrocarbon Process., 55(9), 241-242 (1976). Svedberg, U.G., Numerical solution of multicolumn adsorption processes under periodic countercurrent operation, Chem. Eng. Sci., 31(5), 345-354 (1976). 1977 Grevillot, G., and Tondeur, D., Equilibrium staged parametric pumping, AIChE J., 22(6), 1055-1063 (1976); 23(6), 840-851 (1977).

982

Wankat, P.C., Fractionation by cycling zone adsorption, Chem. Eng. Sci., 32(11), 1283-1288 (1977). 1978 Heck, J.L., and Johansen, T., PSA process improves large-scale hydrogen production, Hydrocarbon Process., 57(1), 175-177 (1978). Narraway, R., PSA gas generators, Processing (Sutton, Engl.), January, 29 (1978). Nelson, W.C.; Silarski, D.F., and Wankat, P.C., Continuous flow equilibrium-staged model for cycling zone adsorption, Ind. Eng. Chem. Fund., 17(1 ), 32-38 (1978). Wankat, P.C., Continuous recuperative-mode parametric pumping, Chem. Eng. Sci., 33(6), 723-734 (1978). Wood, R., Nitrogen from PSA process, Process Eng. (London), June, 44-47 (1978). 1979 Adler, M.S., and Johnson, D.R., A flexible butylene separation process, Chem. Eng. Prog., 75(1), 77-79 (1979). Corr, F.; Dropp, F., and Rudelstorfer, E., PSA produces low-cost high-purity hydrogen, Hydrocarbon Process., 58(3), 119-122 (1979). Gay, P., PSA vs piped gas, Process Eng. (London), February, 41,43 (1979). Parmele, C.S.; O'Connell, W.L., and Basdekis, H.S., Vapor-phase adsorption cuts pollution and recovers solvents, Chem. Eng. (N.Y.), 31 December, 58-70 (1979). Rice, R.G.; Foo, S.C., and Gough, G.G., Limiting separations in parametric pumps, Ind. Eng. Chem. Fund., 18(2), 117-123 (1979). Yang, R.T., and Shen, M.S., Calcium silicates as regenerative sorbents for hot-gas desulfurization, AIChE J., 25(5), 811-819 (1979). 1980 Andrieu, J., and Smith, J.M., Rate parameters for adsorption of carbon dioxide in beds of carbon particles, AIChE J., 26(6), 944-948 (1980). Foo, S.C.; Bergsman, K.H., and Wankat, P.C., Thermal-mode cycling zone adsorption for multicomponent separations, Ind. Eng. Chem. Fund., 19(1), 86-93 (1980). Hill, F.B., Recovery of weakly adsorbed impurity by pressure swing adsorption, Chem. Eng. Commun., 7(1), 3744 (1980). Jacob, P., and Tondeur, D., Nonisothermal adsorption: Separation of gas mixtures by modulation of feed temperature, Sep. Sci. Technol., 15(8), 1563-1578 (1980). Johansson, R., and Neretnieks, I., An experimental study of adsorption on activated carbon in countercurrent flow, Chem. Eng. Sci., 35(4), 979-986 (1980). Thomas, W.J., Gas separation by adsorption, Chem. Ind. (London), 3 May, 366-372 (1980). 1981 Chan, Y.N.I.; Hill, F.B., and Wong, Y.W., Equilibrium theory of a pressure swing adsorption process, Chem. Eng. Sci., 36(2), 243-252 (1981). Rice, R.G., Adsorptive distillation, Chem. Eng. Commun., 10(1), 111-126 (1981). Vanderschuren, J., Plate efficiency of multistage fluidized-bed adsorbers, Chem. Eng. J., 21 ( 1), 1-10 (1981 ). Various, Gas handling (special report), Processing (Sutton, Engl.), July, 15-21, 48 (1981). 1982 Chihara, K.; Suzuki, M., and Smith, J.M., Cyclic regeneration of activated carbon in fluidized beds, AIChE J., 28(1), 129-134 (1982). Ezell, E.L., and Gelo, J.F., Cut cracked-gas drying costs, Hydrocarbon Process., May, 191-193 (1982). Frey, D.D., Model of adsorbent behavior applied to use of layered beds in cycling zone adsorption, Sep. Sci. Technol., 17(13), 1485-1498 (1982). Hill, F.B.; Wong, Y.W., and Chan, Y.N.I., Temperature swing adsorption for hydrogen isotope separation, AIChE J., 28(1), 1-6 (1982). Knopf, F.C., and Rice, R.G., Adsorptive distillation: Optimum solids profiles, Chem. Eng. Commun., 15(1), 109124(1982). Milewski, M., and Berak, J.M., Effect of adsorbent preparation parameters on selectivity for xylene isomers separation, Sep. Sci. Technol., 17(2), 369-374 (1982). Moseman, M.H., and Bird, G., Desiccant dehydration of natural gasoline, Chem. Eng. Prog., 78(2), 78-83 (1982). Santacesaria, E., et al., Separation of xylenes on Y-zeolites, Ind. Eng. Chem. Process Des. Dev., 21(3), 440-457 (1982).

983

Tien, C., and Wang, S.C., Dynamics of adsorption columns with bacterial growth outside adsorbents, Can. J. Chem. Eng., 60, 363-376 (1982). Wong, Y.W., and Hill, F.B., Separation of hydrogen isotopes via single-column pressure swing adsorption, Chem. Eng. Commun., 15(5), 343-356 (1982). 1983 Barrow, J.A., Proper design saves energy (adsorption dehydrators), Hydrocarbon Process., January, 117-120 (1983). Capes, P., Solvent-recovery adsorption system, Process Eng. (London), January, 33 (1983). Carter, J.W., and Wyszynski, M.L., Pressure swing adsorption drying of compressed air, Chem. Eng. Sci., 38(7), 1093-1100 (1983). Fernandez, G.F., and Kenney, C.N., Modelling of the pressure swing air separation process, Chem. Eng. Sci., 38(6), 827-834 (1983). Knaebel, K.S., and Hill, F.B., Analysis of gas purification by pressure swing adsorption: Priming the parametric pump, Sep. Sci. Technol., 18(12), 1193-1220 (1983). Knaebel, K.S., and Pigford, R.L., Equilibrium and dissipative effects in cycling zone adsorption, Ind. Eng. Chem. Fund., 22(3), 336-346 (1983). Tsai, M.C.; Wang, S.S., and Yang, R.T., Temperature-swing adsorption for hydrogen-methane separation, AIChE J., 29(6), 966-975 (1983). Wang, S.S., and Yang, R.T., Multicomponent separation by cyclic processes: A process for combined hydrogen/methane separation and acid gas removal in coal conversions, Chem. Eng. Commun., 20(1), 183190(1983). Watson, A.M., Use pressure swing adsorption for lowest cost hydrogen, Hydrocarbon Process., March, 91-95 (1983). Zuech, J.L.; Hines, A.L., and Sloan, E.D., Methane adsorption on 5A molecular sieve (4-690kPa), Ind. Eng. Chem. Process Des. Dev., 22( 1), 172-174 (1983). 1984 Ladisch, M.R.; Voloch, M.; Hong, J.; Blenkowski, P., and Tsao, G.T., Cornmeal adsorber for dehydrating ethanol vapours, Ind. Eng. Chem. Process Des. Dev., 23(3), 437-443 (1984). Lee, M.H.; Petty, L.E.; Wilson, R.H., and Galvin, C., The ultra low temperature reaction adsorption process, Chem. Eng. Prog., 80(5), 33-38 (1984). Lynch, D.T., The use of adsorption/desorption models to describe the forced periodic operation of catalytic reactors, Chem. Eng. Sci., 39(9), 1325-1328 (1984). Mills, B., and Rothery, E., Gas drying, Chem. Eng. (Rugby, Engl.), April, 19-23 (1984). Morbidelli, M.; Storti, G.; Carra, S.; Niederjaufner, G., and Pontoglio, A., Study of a separation process using a molecular sieve for chlorotoluene isomers, Chem. Eng. Sci., 39(3), 383-394 (1984). Rieke, R.D., Cycling zone adsorption: Variable-feed mode of operation, Sep. Sci. Technol., 19(4), 261-282 (1984). Wang, S.C.P., and Tien, C., Interaction between adsorption and bacterial activity in granular activated carbon columns, AIChE J., 30(5), 786-801 (1984). 1985 Anon., Applications of pressure-swing adsorption, Process Eng. (London), September, 67-71 (1985). Cen, P.L., and Yang, R.T., Separation of five-component gas mixture by pressure swing adsorption, Sep. Sci. Technol., 20(9), 725-748 (1985). Cen, P.L.; Chen, W.N., and Yang, R.T., Ternary gas mixture separation by pressure swing adsorption, Ind. Eng. Chem. Process Des. Dev., 24(4), 1201-1208 (1985). Cheng, H.C., and Hill, F.B., Separation of helium-methane mixtures by pressure swing adsorption, AIChE J., 31(1), 95-102 (1985). Costa, C., and Rodrigues, A., Design of cyclic fixed-bed adsorption procesess, AIChE J., 31(10), 1645-1665 (1985). Knaebel, K.S., and Hill, F.B., Pressure swing adsorption: Development of an equilibrium theory for gas separations, Chem. Eng. Sci., 40(12), 2351-2360 (1985). Platt, D., and Lavie, R., Pressure cyclic zone adsorption, Chem. Eng. Sci., 40(5), 733-740 (1985). Raghavan, N.S.; Hassan, M.M., and Ruthven, D.M., Numerical simulation of a PSA system, AIChE J., 31(3), 385-392; 31(12), 2008-2025 (1985).

984

Tamon, H., and Toei, R., Solar-powered adsorber dehumidifier, Ind. Eng. Chem. Process Des. Dev., 24(2), 450457 (1985). Tondeur, D., and Wankat, P.C., Gas purification by pressure swing adsorption, Sep. Purif. Methods, 14(2), 157212(1985). Tsai, M.C.; Wang, S.S.; Yang, R.T., and Desai, N.J., Temperature-swing separation of hydrogen-methane mixture, Ind. Eng. Chem. Process Des. Dev., 24(1), 57-62 (1985). Wang, J.H., and Smith, J.M., Thermal regeneration of the phenol-carbon system, AIChE J., 31(3), 496-498 (1985). Yang, R.T., and Doong, S.J., Gas separation by pressure swing adsorption: A pore-diffusion model for bulk separation, AIChE J., 31 (11), 1829-1842 (1985). 1986 Carta, G., and Pigford, R.L., Analytical solution for cycling-zone adsorption, Chem. Eng. Sci., 41(3), 511-518 (1986). Cen, P.L., and Yang, R.T., Bulk gas separation by pressure swing adsorption, Ind. Eng. Chem. Fund., 25(4), 758768 (1986). Cen, P.L., and Yang, R.T., Separation of binary gas mixture into two high-purity products by new pressure-swing adsorption cycle, Sep. Sci. Technol., 21(9), 845-864 (1986). Crabb, K.S.; Perona, J.J.; Byers, C.H., and Watson, J.S., Vacuum sorption pumping studies with pure gases on molecular sieves, AIChE J., 32(2), 255-262 (1986). Doong, S.J., and Yang, R.T., Parametric study of pressure swing adsorption process for gas separation: Criterion for pore diffusion limitation, Chem. Eng. Commun., 41, 163-180 (1986). Doong, S.J., and Yang, R.T., Bulk separation of multicomponent gas mixtures by pressure swing adsorption, AIChE J., 32(3), 397-410 (1986). Garg, D.R., and Yon, C.M., Adsorptive heat recovery drying system, Chem. Eng. Prog., 82(2), 54-60 (1986). Kayser, J.C., and Knaebel, K.S., Pressure swing adsorption: Experimental study of an equilibrium theory, Chem. Eng. Sci., 41 (11), 2931-2938 (1986). Pritchard, C.L., and Simpson, G.K., Design of an oxygen concentrator using the rapid pressure-swing adsorption principle, Chem. Eng. Res. Des., 64(6), 467-472 (1986). Raghaven, N.S.; Hassan, M.M., and Ruthven, D.M., Numerical simulation of a PSA system using a pore diffusion model, Chem. Eng. Sci., 41 (11), 2787-2794 (1986). Ray, M.S., Pressure swing adsorption: A review of UK patent literature, Sep. Sci. Technol., 21 (1), 1-38 (1986). Underwood, R.P., Model of a pressure-swing adsorption process for nonlinear adsorption equilibrium, Chem. Eng. Sci., 41(2), 409-412 (1986). 1987 Beevers, A., Adsorption for biotech separations, Processing (Sutton, Engl.), September, 41,43 (1987). Doong, S.J., and Yang, R.T., Comparison of gas separation performance by different pressure swing adsorption cycles, Chem. Eng. Commun., 54, 61-72 (1987). Doong, S.J., and Yang, R.T., Bidisperse pore diffusion model for zeolite pressure swing adsorption, AIChE J., 33(6), 1045-1049 (1987). Guo, D.; Venkat, C., and Weiss, A.H., Cyclic adsorption of styrene and of plasticizer on BPL carbon, Adsorpt. Sci. Yechnol., 4(1 ), 15-24 (1987). Hachiya, K.; Takeda, K., and Yasunaga, T., Pressure-jump method to adsorption-desorption kinetics, Adsorpt. Sci. Yechnol., 4( 1), 25-44 (1987). Lu, X.; Rothstein, D.; Madey, R., and Huang, J.C., Pressure swing adsorption for system with Freundlich isotherm, Sep. Sci. Technol., 22(6), 1547-1556 (1987). Michele, H., Purification of flue gases by dry sorbents, Int. Chem. Eng., 27(2), 183-197 (1987). O'Shea, S., et al., Gas adsorption phenomena in evacuated tubular solar collectors, Adsorpt. Sci. Technol., 4(4), 275-285 (1987). Shin, H.S., and Knaebel, K.S., Pressure swing adsorption: A theoretical study of diffusion-induced separations, AIChE J., 33(4), 654-662 (1987). Yang, R.T., Gas Separation by Adsorption Processes, Butterworth Publishers, Massachusetts, USA (1987).

1988 Basta, N., New developments in pressure swing adsorption, Chem. Eng. (N.Y.), 26 September, 26-31 (1988). Chiang, A.S.T., Arithmetic of PSA process scheduling, AIChE J., 34(11), 1910-1913 (1988).

985

Chiang, A.S.T.; Hwong, M.Y.; Lee, T.Y., and Cheng, T.W., Oxygen enrichment by pressure swing adsorption, Ind. Eng. Chem. Res., 27(1), 81-86 (1988). Chynoweth, E., PSA on-site nitrogen generation, Processing (Sutton, Engl.), February, 30 (1988). Davis, M.M.; McAvoy, R.L., and LeVan, M.D., Periodic states for thermal swing adsorption of gas mixtures, Ind. Eng. Chem. Res., 27(7), 1229-1235 (1988). Duprat, F.; Gassend, R., and Gau, G., Inductive adsorption: A new method of isomer separation, Ind. Eng. Chem. Res., 27(5), 831-836 (1988). Farooq, S.; Hassan, M.M., and Ruthven, D.M., Heat effects in pressure swing adsorption systems, Chem. Eng. Sci., 43(5), 1017-1033 (1988). Ferraz, M.C.A., Preparation of activated carbon for air pollution control, Fuel, 67(9), 1237-1241 (1988). Gollakota, S.V., and Chriswell, C.D., Study of an adsorption process using silicate for sulfur dioxide removal from combustion gases, Ind. Eng. Chem. Res., 27(1), 139-143 (1988). Kapoor, A., and Yang, R.T., Separation of hydrogen-lean mixtures for high-purity hydrogen by vacuum swing adsorption, Sep. Sci. Technol., 23(1), 153-178 (1988). Kapoor, A., and Yang, R.T., Optimization of a pressure swing adsorption cycle, Ind. Eng. Chem. Res., 27(1), 204-206 (1988). Kayser, J.C., and Knaebel, K.S., Integrated steps in pressure swing adsorption cycles, Chem. Eng. Sci., 43(11), 3015-3022 (1988). Koo, Y.M., and Wankat, P.C., Modeling of size-exclusion parametric pumping, Sep. Sci. Technol., 23(4), 413428(1988). LeVan, M.D.; Mao, R., and McLaughlin, G.P., Ethylene recovery from low-grade gas stream by adsorption on zeolites and controlled desorption, Can. J. Chem. Eng., 66(4), 686-690 (1988). Lu, X.; Rothstein, D.; Madey, R., and Huang, J.C., Pressure swing adsorption for a system with a Langmuir isotherm, Sep. Sci. Technol., 23(4), 281-292 (1988). Matz, M.J., and Knaebel, K.S., Pressure swing adsorption: Effects of incomplete purge, AIChE J., 34(9), 14861492(1988). McCormick, R.L., et al., Surface acidity studied temperature-programmed desorption of tert-butylamine, Energy Fuels, 2(6), 740-743 (1988). McKay, G., Fluidized bed adsorption of pollutants onto activated carbon, Chem. Eng. J., 39(2), 87-96 (1988). Murakami, M., and Nomura, M., Catalyst-adsorption type purifiers for ultra-pure gas supply, Gas Sep. Purif., 2(2), 95-102 (1988). Rousar, I., and Ditl, P., Optimization of pressure swing adsorption equipment, Chem. Eng. Commun., 70, 67-106 (1988). Schork, J.M., and Fair, J.R., Parametric analysis of thermal regeneration of adsorption beds, Ind. Eng. Chem. Res., 27(3), 457-469 (1988). Shin, H.S., and Knaebel, K.S., Pressure swing adsorption: Experimental study of diffusion-induced separation, AIChE J., 34(9), 1409-1416 (1988). Sircar, S., Separation of methane and carbon dioxide gas mixtures by pressure swing adsorption, Sep. Sci. Technol., 23(6), 519-530 (1988). Sircar, S., Air fractionation by adsorption, Sep. Sci. Technol., 23(14), 2379-2396 (1988). Sircar, S., and Kratz, W.C., Pressure-swing adsorption process for production of 23-50% oxygen-enriched air, Sep. Sci. Technol., 23(4), 437-450 (1988). Sircar, S., and Kratz, W.C., Simultaneous production of hydrogen and carbon dioxide from steam reformer offgas by pressure swing adsorption, Sep. Sci. Technol., 23(14), 2397-2416 (1988). Toftegard, B., and Jorgensen, S.B., Stationary profiles for periodic cycled separation columns, Ind. Eng. Chem. Res., 27(3), 481-485 (1988). Wiessner, F.G., Basics and industrial applications of pressure swing adsorption for gas separations, Gas Sep. Purif., 2(3), 115-119 (1988). 1989 Buzanowski, M.A.; Yang, R.T., and Hass, O.W., Direct observation of effects of bed pressure drop on adsorption and desorption dynamics, Chem. Eng. Sci., 44(10), 2392-2394 (1989). Davis, M.M., and LeVan, M.D., Experiments on optimization of thermal swing adsorption, Ind. Eng. Chem. Res., 28(6), 778-785 (1989). Douss, N., and Meunier, F., Experimental study of cascading adsorption cycles, Chem. Eng. Sci., 44(2), 225-236 (1989).

986

Farooq, S.; Ruthven, D.M., and Boniface, H.A., Numerical simulation of a pressure swing adsorption oxygen unit, Chem. Eng. Sci., 44(12), 2809-2816 (1989). Haggin, J., Nitrogen production for blanketing applications, Chem. Eng. News, 3 April, 27-29 (1989). Huang, C.C., and Fair, J.R., Parametric analysis of thermal swing cycles for multicomponent adsorption, AIChE J., 35(10), 1667-1677 (1989). Jozewicz, W., and Kirchgessner, D.A., Activation and reactivity of novel calcium-based sorbents for dry sulfur dioxide control in boilers, Powder Technol., 58(3), 221-230 (1989). Kawai, M., and Kaneko, T., Present state of PSA air separation in Japan, Gas Sep. Purif., 3(1), 2-6 (1989). Kayser, J.C., and Knaebel, K.S., Pressure swing adsorption: Development of an equilibrium theory for binary gas mixtures with nonlinear isotherms, Chem. Eng. Sci., 44(1), 1-8 (1989). Kubota, K.; Hata, C., and Hayashi, S., Study of simulated moving bed adsorber based on the axial dispersion model, Can. J. Chem. Eng., 67(6), 1025-1029 (1989). Kumar, R., Adsorption column blowdown: Adiabatic equilibrium model for bulk binary gas mixtures, Ind. Eng. Chem. Res., 28(11), 1677-1684 (1989). Kumar, R., and Van Sloun, J.K., Purification by adsorptive separation, Chem. Eng. Prog., 85(1), 34-40 (1989). Long, R.B., and Speight, J.G., Scale-up studies for separating heavy petroleum feedstocks by adsorption, Ind. Eng. Chem. Res., 28(10), 1503-1507 (1989). Otowa, T., et al., Methane adsorption as a calorie upgrading PSA in the SNG process, Gas Sep. Purif., 3(3), 139142(1989). Pande, D.R., and Fabiani, C., Feasibility studies on the use of a naturally occurring molecular sieve for methane enrichment from biogas, Gas Sep. Purif., 3(3), 143-147 (1989). Ruthven, D.M., and Ching, C.B., Countercurrent and simulated countercurrent adsorption separation processes (review article), Chem. Eng. Sci., 44(5), 1011-1038 (1989). Sircar, S., and Kratz, W.C., Oxygen production by pressure swing adsorption, Sep. Sci. Technol., 24(5), 429-440 (1989). Storti, G.; Masi, M.; Carra, S., and Morbidelli, M., Optimal design of multicomponent countercurrent adsorption separation processes involving nonlinear equilibria, Chem. Eng. Sci., 44(6), 1329-1346 (1989). Suh, S.S., and Wankat, P.C., New PSA process for high enrichment and recovery, Chem. Eng. Sci., 44(3), 567574(1989). Suh, S.S., and Wankat, P.C., Pressure swing adsorption process for binary gas separation with Langmuir isotherms, Chem. Eng. Sci., 44(10), 2407-2411 (1989). Suh, S.S., and Wankat, P.C., Combined cocurrent-countercurrent blowdown cycle in pressure swing adsorption, AIChE J., 35(3), 523-529 (1989). White, D.H., and Barkley, P.G., Design of pressure swing adsorption systems, Chem. Eng. Prog., 85(1), 25-33 (1989). 1990 Ackley, M.W., and Yang, R.T., Kinetic separation by pressure swing adsorption: Method of characteristics model, AIChE J., 36(8), 1229-1238 (1990). Anon., PSA nitrogen generator, Process Eng. (London), May, 23 (1990). Anon., PSA nitrogen generator, Chem. Eng. (N.Y.), September, 213-214 (1990). Banerjee, R.; Narayankhedkar, K.G., and Sukhatme, S.P., Exergy analysis of pressure swing adsorption processes for air separation, Chem. Eng. Sci., 45(2), 467-476 (1990). Du, Z.; Sarofim, A.F., and Longwell, J.P., Activation energy distribution in temperature-programmed desorption: Modeling and application to the soot-oxygen system, Energy Fuels, 4(3), 296-303 (1990). Farooq, S., and Ruthven, D.M., A comparison of linear driving force and pore diffusion models for a pressure swing adsorption bulk separation process, Chem. Eng. Sci., 45(1), 107-116 (1990). Farooq, S., and Ruthven, D.M., Continuous countercurrent flow model for a bulk PSA separation process, AIChE J., 36(2), 310-314 (1990). Hart, J.; Battrum, M.J., and Thomas, W.J., Axial pressure gradients during the pressurization and depressurization steps of a PSA gas separation cycle, Gas Sep. Purif., 4(2), 97-102 (1990). Hassan, M.M.; Fatehi, A.I., and Loughlin, K.F., Non-linear isothermal packed bed adsorbers: Comparison of Peclet and cell numbers, Gas Sep. Purif., 4(1), 50-57 (1990). Islam, M.R., and Chakma, A., Simulation of activated carbon adsorbers used in gas plants, Gas Sep. Purif., 4(2), 103-108(1990).

987

Jianyu, G., and Zhenhua, Y., Analog circuit for simulation of pressure swing adsorption, Chem. Eng. Sci., 45(10), 3063-3070 (1990). Jozewicz, W.; Chang, J.C.S., and Sedman, C.B., Bench-scale evaluation of calcium sorbents for acid gas emission control, Environ. Prog., 9(3), 137-142 (1990). Kulvaranon, S.: Findley, M.E., and Liapis, A.I., Increased separation by variable-temperature stepwise desorption in multicomponent adsorption processes, Ind. Eng. Chem. Res., 29(1), 106-115 (1990). LeVan, M.D., Multiple periodic states for thermal swing adsorption of gas mixtures, Ind. Eng. Chem. Res., 29(4), 625-631 (1990). Liow, J-L., and Kenney, C.N., The backfill cycle of the pressure swing adsorption process, AIChE J., 36(1), 5365(1990). Loureiro, J.; Costa, C., and Rodrigues, A., Dynamics of adsorptive reactors: Instantaneous non-linear adsorption and finite zero order irreversible reaction, Can. J. Chem. Eng., 68(1), 127-138 (1990). Lu, X., et al., Pressure swing adsorption for a system with a Langmuir-Freundlich isotherm, Chem. Eng. Sci., 45(4), 1097-1104 (1990). Nakamura, S., et al., Development of circulating fluidized bed adsorber, AIChE J., 36(7), 1003-1010 (1990). Richardson, J.F., Separation processes (review paper), Gas Sep. Purif., 4(1), 2-7 (1990). Rota, R., and Wankat, P.C., Intensification of pressure swing adsorption processes, AIChE J., 36(9), 1299-1312 (1990). Ruthven, D.M., and Farooq, S., Air separation by pressure swing adsorption, Gas Sep. Purif., 4(3), 141-148 (1990). Shu, C.M., et al., Experimental and computational studies on propane-propylene separation by adsorption and variable-temperature stepwise desorption, Sep. Technol., 1(1), 18-28 (1990). Sircar, S., Production of hydrogen and ammonia synthesis gas by pressure swing adsorption, Sep. Sci. Technol., 25(11), 1087-1100 (1990). Smith, O.J., and Westerberg, A.W., Mixed-integer programming for pressure swing adsorption cycle scheduling, Chem. Eng. Sci., 45(9), 2833-2842 (1990). Tezel, F.H., et al., Adsorptive separation of krypton from nitrogen: Development of a pilot unit, Can. J. Chem. Eng., 68(2), 268-277 (1990). 1991 Battrum, M.J., and Thomas, W.J., Carbon monoxide recovery by pressure swing adsorption, Chem. Eng. Res. Des., 69(2), 119-129 (1991). Bracey, W.; Akman, U., and Sunol, A.K., Fractionation of aqueous fructose-glucose mixtures by adsorption and supercritical desorption, Food Bioprod. Process., 69(C2), 97-101 (1991). Chen, A.Q., and Wankat, P.C., Analytical scaling of thermal swing adsorption, Sep. Sci. Technol., 26(12), 15751584(1991). Chen, A.Q., and Wankat, P.C., Scaling rules and intensification of thermal swing adsorption, AIChE J., 37(5), 785-789 (1991 ). Cooney, D.O., Determining external film mass transfer coefficients for adsorption columns, AIChE J., 37(8), 1270-1274 (1991). Farooq, S., and Ruthven, D.M., Numerical simulation of a kinetically controlled pressure swing adsorption bulk separation process based on a diffusion model, Chem. Eng. Sci., 46(9), 2213-2224 (1991). Garcia, A.A., Strategies for the recovery of chemicals from fermentation: A review of the use of polymeric adsorbents, Biotechnol. Prog., 7(1), 33-42 (1991 ). Gladden, L.F., Influence of pellet structure on selectivity during pressure swing adsorption separations, Chem. Eng. Sci., 46(10), 2455-2464 (1991). Haggin, J., New PSA hydrogen process, Chem. Eng. News, 13 May, 29 (1991). Hart, J., and Thomas, W.J., Gas separation by pulsed pressure swing adsorption, Gas Sep. Purif., 5(2), 125-133 (1991). Jasra, R.V.; Choudary, N.V., and Bhat, S.G.T., Separation of gases by pressure swing adsorption (a review), Sep. Sci. Technol., 26(7), 885-930 (1991). Joshi, S., and Fair, J.R., Adsorptive drying of hydrocarbon liquids, Ind. Eng. Chem. Res., 30(1), 177-185 (1991). Kasuya, F., and Tsuji, Y., High purity CO gas separation by pressure swing adsorption, Gas Sep. Purif., 5(4), 242-246(1991). Kikkinides, E.S., and Yang, R.T., Simultaneous sulfur dioxide/NOx removal and sulfur dioxide recovery from flue gas by pressure swing adsorption, Ind. Eng. Chem. Res., 30(8), 1981-1989 (1991).

988

Kluge, G., et al., Experimental investigation and mathematical modelling of the purge step in pressure swing adsorption, Gas Sep. Purif., 5(4), 235-241 (1991). Kumar, R., Adiabatic adsorption column: Pressure reduction to subatmospheric levels, Sep. Sci. Technol., 26(8), 1093-1108(1991). Kumar, R., and Golden, T.C., Thermal swing adsorption process for removing trace impurities from a multicomponent gas mixture (landfill gas), Gas Sep. Purif., 5(1 ), 21-24 (1991). Leng, L.J., and Kenney, C.N., The simple cycle of the pressure swing adsorption process, Chem. Eng. Res. Des., 69(3 ), 215-220 ( 1991 ). Malecky, M., et al., Mathematical modeling of an adsorption column with thermal regeneration, Int. Chem. Eng., 31(4), 708-714(1991). Matz, M.J., and Knaebel, K.S., Recycled thermal swing adsorption: Applied to separation of binary and ternary mixtures, Ind. Eng. Chem. Res., 30(5), 1046-1054 (1991). Mutasim, Z.Z., and Bowen, J.H., Pressure swing adsorption in non-isothermal, non-equilibrium conditions: Single adsorbate, Chem. Eng. Res. Des., 69(2), 108-118 (1991). Pechler, N., and Emig, G., Adsorptive purification of large waste gas streams, Gas Sep. Purif., 5(4), 247-251 (1991). Petkovska, M., et al., Temperature-swing gas separation with electrothermal desorption step, Sep. Sci. Technol., 26(3), 425-444 (1991). Ritter, J.A., and Yang, R.T., Air purification and vapor recovery by pressure swing adsorption: A comparison of silicalite and activated carbon, Chem. Eng. Commun., 108,289-306 (1991). Ritter, J.A., and Yang, R.T., Pressure swing adsorption: Experimental and theoretical study on air purification and vapor recovery, Ind. Eng. Chem. Res., 30(5), 1023-1032 (1991). Rodrigues, A.E.; Loureiro, J.M., and LeVan, M.D., Simulated pressurization of adsorption beds, Gas Sep. Purif., 5(2), 115-124 (1991). Ruthven, D.M., Comparison of performance of PSA and membrane systems for nitrogen production, Gas Sep. Purif., 5(1), 9-10 (1991). Scott, D.M., Similarity solutions for pressurisation and depressurisation with a two-component gas in an adsorbing bed, Chem. Eng. Sci., 46(12), 2977-2982 (1991). Smith, O.J., and Westerberg, A.W., The optimal design of pressure swing adsorption systems, Chem. Eng. Sci., 46(12), 2967-2976 (1991). Thorogood, R.M., Developments in air separation, Gas Sep. Purif., 5(2), 83-94 (1991). Ysezos, M., and Wang, X., Study on the kinetics of hazardous pollutants adsorption and desorption by biomass: Mechanistic considerations, J. Chem. Technol. Biotechnol., 50(4), 507-522 (1991). Vyas, S.N., et al., Bulk separation and purification of methane/carbon dioxide mixtures on 4A/13X molecular sieves by using pressure swing adsorption, Sep. Sci. Technol., 26(10), 1419-1430 (1991). Zhu, L.; Tu, J., and Shi, Y., Separation of carbon monoxide-carbon dioxide-nitrogen gas mixture for high-purity CO by pressure swing adsorption, Gas Sep. Purif., 5(3), 173-176 (1991). 1992 Alpay, E., and Scott, D.M., The linear driving force model for fast-cycle adsorption and desorption in spherical particle, Chem. Eng. Sci., 47(2), 499-502 (1992). Anon., TSA process for propane/propylene split, Chem. Eng. Prog., 88(3), 12 (1992). Banerjee, R.; Narayankhedkar, K.G., and Sukhatme, S.P., Exergy analysis of kinetic pressure swing adsorption processes: Comparison of different cycle configurations, Chem. Eng. Sci., 47(5), 1307-1311 (1992). Bossy, A.; Tonduer, D., and Jedrzejak, A., A non-linear equilibrium analysis of blowdown policy in pressureswing-adsorption separation, Chem. Eng. J., 48(3), 173-182 (1992). Farooq, S., and Ruthven, D.M., Effect of equilibrium selectivity in a kinetically controlled PSA separation, Chem. Eng. Sci., 47(8), 2093-2094 (1992). Inui, T., et al., Performance of iron-incorporated A-type zeolites for oxygen/nitrogen separation from air by pressure swing adsorption, Gas Sep. Purif., 6(4), 185-190 (1992). Kumar, R., et al., Gas mixture fractionation to produce two high purity products by pressure swing adsorption, Sep. Sci. Yechnol., 27(4), 509-522 (1992). Kumar, R., et al., Novel adsorption-distillation hybrid scheme for propane/propylene separation, Sep. Sci. Technol., 27(15), 2157-2170 (1992). Lemcoff, N.O., and LaCava, A.I., Effect of regeneration pressure level in kinetically controlled pressure swing adsorption, Gas Sep. Purif., 6(1), 9-14 (1992).

989

Lu, Z.P., et al., Intraparticle convection effect on pressurization and blowdown of adsorbers, AIChE J., 38(6), 857-867 (1992). Lu, Z.P., et al., Dynamics of pressurization and blowdown of an adiabatic bed, Gas Sep. Purif., 6(1), 15-24 (1992). Lu, Z.P., et al., Intraparticle diffusion/convection models for pressurization and blowdown of adsorption beds with Langmuir isotherm, Sep. Sci. Technol., 27(14), 1857-1874 (1992). Mutasim, Z.Z., and Bowen, J.H., Multicomponent pressure swing adsorption for non-isothermal non-equilibrium conditions, Chem. Eng. Res. Des., 70(4), 346-353 (1992). Portsmouth, R.L., and Gladden, L.F., Transport in zeolites: Application to pressure and temperature swing process, Chem. Eng. Res. Des., 70(2), 186-188 (1992). Rao, M.B., and Sircar, S., Production of motor fuel grade alcohol by concentration swing adsorption, Sep. Sci. Technol., 27(14), 1875-1888 (1992). Rodrigues, M.I., et al., Dynamic modelling, simulation and control of continuous adsorption recycle extraction, Chem. Eng. Sci., 47(1), 263-270 (1992). Shah, G.C., Improve activated carbon bed adsorber operations, Hydrocarbon Process., 71 (11), 61-63 (1992). Smith, O.J., and Westerberg, A.W., Acceleration of cyclic steady state convergence for pressure swing adsorption models, Ind. Eng. Chem. Res., 31(6), 1569-1573 (1992). Smith, O.J., and Westerberg, A.W., The optimal design of pressure swing adsorption systems, Chem. Eng. Sci., 47(15), 4213-4217 (1992). Storti, G., et al., Performance of a six-port simulated moving-bed pilot plant for vapor-phase adsorption separations, Sep. Sci. Technol., 27(14), 1889-1916 (1992). Various, Application of gas and liquid phase adsorption (topic issue), Sep. Sci. Technol., 27(14), 1822-2040 (1992). Zhong, G.M., et al., Pressurization of a single-component gas in an adsorption column, Chem. Eng. Sci., 47(3), 543-550 (1992). 1993 Aitani, A.M., Sour natural gas drying, Hydrocarbon Process., 72(4), 67-73 (1993). Alpay, E.; Kenney, C.N., and Scott, D.M., Simulation of rapid pressure swing adsorption and reaction processes, Chem. Eng. Sci., 48(18), 3173-3186 (1993). Baron, G., Modelling PSA processes, Gas Sep. Purif., 7(2), 111-118 (1993). Eldridge, R.B., Olefin/paraffin separation technology: A review, Ind. Eng. Chem. Res., 32(10), 2208-2212 (1993). Ellis, D.I.; Heggs, P.J., and Lawrie, M.G., The removal of Freon-12 from air using pressure swing adsorption, Chem. Eng. Res. Des., 71(2), 169-180 (1993). Farooq, S.; Rathor, M.N., and Hidajat, K., A predictive model for a kinetically controlled pressure swing adsorption separation process, Chem. Eng. Sci., 48(24), 4129-4142 (1993). Fish, B.B.; Cart, R.W., and Aris, R., Design and performance of a simulated countercurrent moving-bed separator, AIChE J., 39(11), 1783-1790 (1993). Georgiev, J.K., On the possibility of effective helium extraction from air, Sep. Sci. Technol., 28(15), 2401-2418 (1993). Germerdonk, R., and Wang, A., Pollutant adsorption during activated carbon and steam regeneration in technical columns (experiments, modelling): Influence of radial maldistribution during adsorption and desorption, Chem. Eng. Process., 32(6), 359-368 (1993). Germerdonk, R., and Wang, A., Pollutant adsorption during activated carbon and steam regeneration in technical columns (experiments, modelling): Influence of process parameters on exit concentration during steam regeneration, Chem. Eng. Process., 32(6), 369-378 (1993). Golden, T.C., et al., Carbon-based oxygen selective dessicants for use in nitrogen PSA, Gas Sep. Purif., 7(4), 274-278(1993). Gong, S.Y., and Lee, W.K., The effect of the time length of pressure changing steps on concentration in the gas phase in the pressure swing adsorption process, Sep. Sci. Yechnol., 28(11), 1995-2006 (1993). Gouvalias, G.S., and Markatos, N.C., Mathematical modeling of heat and mass transfer in packed-bed adsorbers/regenerators, AIChE J., 39(11), 1799-1809 (1993). Guan, J., and Ye, Z., Analog curcuit for simulation of pressure swing adsorption: Kinetic model, Chem. Eng. Sci., 48(15), 2821-2823 (1993).

990

Huang, C.C.; Hwu, T.L., and Hsai, Y.S., Regeneration of a fixed bed of activated carbon adsorbed organic vapour by using hot nitrogen purge, Sep. Sci. Technol., 28(15), 2431-2448 (1993). Ivo, R., and Pavel, D., Pressure swing adsorption: Analytical solution for optimum purge, Chem. Eng. Sci., 48(4), 723-734 (1993). Kikkinides, E.S.; Yang, R.T., and Cho, S.H., Concentration and recovery of carbon dioxide from flue gas by pressure swing adsorption, Ind. Eng. Chem. Res., 32(11), 2714-2720 (1993). Lu, Z., et al., Pressure swing adsorption processes: Intraparticle diffusion/convection models, Ind. Eng. Chem. Res., 32(11), 2740-2751 (1993). Lu, Z.P., et al., Pressurization and blowdown of adsorption beds: Effect of the momentum and equilibrium relations on isothermal operation, Chem. Eng. Sci., 48(9), 1699-1708 (1993). Lu, Z.P., et al., Simulation of a three-step one-column pressure swing adsorption process, AIChE J., 39(9), 14831496(1993). Meunier, F., Thermal swing adsorption refrigeration (heat pump), Sep. Yechnol., 3(3), 143-150 (1993). Ng, M.; Schork, J.M., and Fabergas, K.R., The mass transfer zone in nitrogen PSA columns, Gas Sep. Purif., 7(3), 159-166 (1993). Ruthven, D.M.; Xu, Z., and Farooq, S., Sorption kinetics in PSA systems, Gas Sep. Purif., 7(2), 75-82 (1993). Schoofs, G.R., Operation of adsorbers for purifying ammonia synthesis make-up gas, Ind. Eng. Chem. Res., 32(4), 613-619 (1993). Schoofs, G.R., and Petit, P., Repressurization of adsorption purifiers for cryogenic air separation, Chem. Eng. Sci., 48(4), 753-760 (1993). Schork, J.M.; Srinivasan, R., and Auvil, S.R., A shortcut computational method for designing nitrogen PSA adsorbents, Ind. Eng. Chem. Res., 32(10), 2226-2235 (1993). Schulte-Schulze-Berndt, A., and Krabiell, K., Nitrogen generation by pressure swing adsorption on carbon molecular sieves, Gas Sep. Purif., 7(4), 253-258 (1993). Scott, D.M., Effects of bed pressure drop on adsorption and desorption with Langmuir isotherms, Chem. Eng. Sci., 48(17), 3001-3006 (1993). Shirley, A.I., and LaCava, A.I., Novel pressurization methods in pressure swing adsorption systems for the generation of high-purity gas, Ind. Eng. Chem. Res., 32(5), 906-910 (1993). Sircar, S., Recent trends in pressure swing adsorption: Production of multiple products from a multicomponent feed gas, Gas Sep. Purif., 7(2), 69-74 (1993). Sircar, S., and Hanley, B.F., Fractionated vacuum swing adsorption process for air separation, Sep. Sci. Technol., 28(17), 2553-2566 (1993). Sundaram, N., A noniterative solution for periodic steady states in gas purification pressure swing adsorption, Ind. Eng. Chem. Res., 32(8), 1686-1691 (1993). Takeguchi, T.; Tanakulrungsank, W., and Inui, T., Separation and/or concentration of carbon dioxide from carbon dioxide/nitrogen gaseous mixture by PSA using metal-incorporated microporous crystals with high surface area, Gas Sep. Purif., 7(1), 3-10 (1993). Tondeur, D., and Chlendi, M., Front analysis and cycle policy in PSA operations, Gas Sep. Purif., 7(2), 105-110 (1993). Various, Papers from the Symposium on Gas Separation by PSA (14 February 1992, The Netherlands), Gas Sep. Purif., 7(2), 67-1 t8 (1993). Wankat, P.C., Feed purge cycles in pressure swing adsorption, Sep. Sci. Technol., 28(17), 2567-2586 (1993). Westgate, P.J., and Ladisch, M.R., Air drying using corn grits as the sorbent in a pressure swing adsorber, AIChE J., 39(4), 720-723 (1993). 1994 Alpay, E.; Kenney, C.N., and Scott, D.M., Adsorbent particle size effects in the separation of air by rapid pressure swing adsorption, Chem. Eng. Sci., 49(18), 3059-3076 (1994). Anon., Reduce the risk of adsorbent bed fires during vapor recovery, Chem. Eng. (N.Y.), December, 127 (1994). Basta, N., New sorbents and adsorption processes, Chem. Eng. (N.Y.), November, 39-43 (1994). Chiang, A.S.T., et al., Experimental study on a four-bed PSA air separation process, AIChE J., 40(12), 19761982 (1994). Chou, C.; Wu, C.L., and Chiang, A.S.T., A complementary pressure swing adsorption process configuration for air separation, Sep. Technol., 4(2), 93-103 (1994). Chou, C.T., and Huang, W.C., Incorporation of a valve equation into the simulation of a pressure swing adsorption process, Chem. Eng. Sci., 49(1), 75-84 (1994).

991

Chou, C.T., and Huang, W.C., Simulation of a four-bed pressure swing adsorption process for oxygen enrichment, Ind. Eng. Chem. Res., 33(5), 1250-1258 (1994). Crittenden, B.D., et al., Dynamics of pressurization and depressurization during pressure swing adsorption, Chem. Eng. Sci., 49(16), 2657-2670 (1994). Farooq, S.; Rathor, M.N., and Hidajat, K., A modified continuous countercurrent flow model for a PSA separation process, Chem. Eng. Res. Des., 72(1), 114-118 (1994). Kirkby, N.F., and Morgan, J.E.P., A theoretical investigation of pressure swing reaction, Chem. Eng. Res. Des., 72(4), 541-550 (1994). Kovvali, S.A.; Vemury, S., and Admassu, W., Modeling of multicomponent countercurrent gas permeators, Ind. Eng. Chem. Res., 33(4), 896-903 (1994). Kumar, R., Pressure swing adsorption process: Performance optimum and adsorbent selection, Ind. Eng. Chem. Res., 33(6), 1600-1605 (1994). Kumar, R., et al., A versatile process simulator for adsorptive separations, Chem. Eng. Sci., 49(18), 3115-3126 (1994). Lu, Z.P., and Rodrigues, A.E., Pressure swing adsorption reactors: Simulation of three-step one-bed process, AIChE J., 40(7), 1118-1137 (1994). Naumov, V.A.; Vorotynthev, V.B., and Kusiy, G.V., Non-isothermal regeneration of molecular sieve adsorbers in a cyclic process of air purification from moisture and carbon dioxide, Gas Sep. Purif., 8(3), 167-174 (1994). Oliveira, F., et al., Two adaptive grid methods for fixed bed systems simulation, Comput. Chem. Eng., 18(3), 227-244 (1994). Rahman, A.K.M.S.; Hassan, M.M., and Loughlin, K.F., Unsteady state simulation of "SORBEX" system with nonlinear adsorption isotherms, Sep. Technol., 4(1), 27-37 (1994). Rajniak, P., and Yang, R.T., Hysteresis-dependent adsorption-desorption cycles: Generalization for isothermal conditions, AIChE J., 40(6), 913-924 (1994). Reiss, G., Status and development of oxygen generation process on molecular sieve zeolites, Gas Sep. Purif., 8(2), 95-100 (1994). Reunanen, J., et al., Reliable numerical calculation of the breakthrough curve of an adsorption column, Chem. Eng. Process., 33(3), 117-124 (1994). Ruthven, D.M., and Farooq, S., Concentration of a trace component by pressure swing adsorption, Chem. Eng. Sci., 49(1 ), 51-60 (1994). Ruthven, D.M., and Inui, T., Letter and reply concerning "Carbon dioxide/nitrogen separation by PSA using metal-incorporated crystals", Gas Sep. Purif., 8(1), 53 (1994). Salasivam, M., and Balakrishnan, A.R., Experimental investigations on the thermal effects in packed bed liquid desiccant dehumidifiers, Ind. Eng. Chem. Res., 33(6), 1636-1640 (1994). Scott, D.M., The linear driving force model for cyclic adsorption and desorption: The effect of shape, Chem. Eng. Sci., 49(6), 914-916 (1994). Shirley, A.I., and Kikkinides, E.S., Comments and reply on "Concentration and recovery of carbon dioxide from flue gas by pressure swing adsorption", Ind. Eng. Chem. Res., 33(11), 2879-2881 (1994). Sorrento, L., The proven process of carbon adsorption, Chem. Eng. (N.Y.), July, 94-95 (1994). Tsuru, T., and Hwang, S.T., Production of high-purity oxygen by continuous membrane column combined with PSA oxygen generator, Ind. Eng. Chem. Res., 33(2), 311-316 (1994). 1995 Acharya, D.R., and Jain, R., Recent advances in molecular sieve unit design for air separation plants, Sep. Sci. Technol., 30(18), 3489-3508 (1995). Barbosa Mota, J.P., et al., A simulation model of a high-capacity methane adsorptive storage system, Adsorption, 1(1), 17-28 (1995). Chiang, A.S.T., and Hong, M.C., Radial-flow rapid pressure swing adsorption, Adsorption, 1(2), 153-164 (1995). Chlendi, M., and Tondeur, D., Dynamic behaviour of layered columns in pressure swing adsorption, Gas Sep. Purif., 9(4), 231-242 (1995). Chlendi, M.; Tondeur, D., and Rolland, F., A method to obtain a compact representation of process performances from a numerical simulator: Example of pressure swing adsorption for pure hydrogen production, Gas Sep. Purif., 9(2), 125-136 (1995).

992

Chue, K.T.; Kim, J.N., and Yang, R.T., Comparison of activated carbon and zeolite 13X for carbon dioxide recovery from flue gas by pressure swing adsorption, Ind. Eng. Chem. Res., 34(2), 591-598 (1995). Crittenden, B.D., et al., Pressure, concentration and temperature profiles in a 5A zeolite adsorbent bed during pressurisation and depressurisation with air, Chem. Eng. Sci., 50(9), 1417-1428 (1995); corrigendum, 50(16), 2677 (1995). Diagne, D.; Goto, M., and Hirose, T., Parametric studies on carbon dioxide separation and recovery by a dual reflux PSA process consisting of both rectifying and stripping sections, Ind. Eng. Chem. Res., 34(9), 30833089 (1995). Fatehi, A.I.; Loughlin, K.F., and Hassan, M.M., Separation of methane-nitrogen mixtures by pressure swing adsorption using a carbon molecular sieve, Gas Sep. Purif., 9(3), 199-204 (1995). Ferreira, L.M., and Rodrigues, A.E., Adsorptive separation by thermal parametric pumping: Modeling and simulation, Adsorption, 1(3), 213-232 (1995). Ferreira, L.M., and Rodrigues, A.E., Adsorptive separation by thermal parametric pumping: Experimental study of the purification of aqueous phenolic solutions at pilot scale, Adsorption, 1(3), 233-252 (1995). Gilleskie, G.L.; Parker, J.L., and Cussler, E.L., Gas separations in hollow-fiber adsorbers, AIChE J., 41(6), 14131425(1995). Hartzog, D.G., and Sircar, S., Sensitivity of PSA process performance to input variables, Adsorption, 1(2), 133152(1995). Hassan, M.M.; Shamsur Rahman, A.K.M., and Loughlin, K.F., Modelling of simulated moving bed adsorption system: A more precise approach, Sep. Technol., 5(2), 77-90 (1995). Heggs, P.J.; Ellis, D.I., and Ismail, M.S., Prediction of flow distributions and pressure changes in multi-layered annular packed beds, Gas Sep. Purif., 9(4), 243-252 (! 995). Keller, G.E., Adsorption: Building upon a solid foundation, Chem. Eng. Prog., 91(10), 56-67 (1995). Kim, J.N., et al., Production of high-purity nitrogen from air by pressure swing adsorption on zeolite X, Sep. Sci. Technol., 30(3), 347-368 (1995). Knaebel, K.S., For your next separation consider adsorption, Chem. Eng. (N.Y.), November, 92-102 (1995). Kumar, R., Effect of variable feed concentration on the performance of pressure swing adsorption process, Adsorption, 1(3), 203-212 (1995). Kumar, R.; Guro, D.E., and Schmidt, W.P., A new concept to increase recovery from hydrogen PSA: Processing different pressure feed streams in a single unit, Gas Sep. Purif., 9(4), 271-276 (1995). Kvamsdal, H.M., and Hertzberg, T., Optimization of pressure swing adsorption systems" The effect of mass transfer during the blowdown step, Chem. Eng. Sci., 50(7), 1203-1212 (1995). LeVan, M.D., Pressure swing adsorption: Equilibrium theory for purification and enrichment, Ind. Eng. Chem. Res., 34(8), 2655-2660 (1995). LeVan, M.D., and Croft, D.T., Determination of periodic states of pressure swing adsorption cycles, Gas Sep. Purif., 9(1), 13-16 (1995). Ming, F., et al., Identification and prediction of protein adsorption breakthrough, desorption, and fractionation in a packed column using a neural network, Sep. Sci. Technol., 30(7), 1397-1406 (1995). Reichhold, A., and Hofbauer, H., Internally circulating fluidized bed for continuous adsorption and desorption, Chem. Eng. Process., 34(6), 521-528 (1995). Shin, H., Separation of a binary gas mixture by pressure swing adsorption: Comparison of different PSA cycles, Adsorption, 1(4), 321-330 (1995). Shirley, A.I., and LaCava, A.I., PSA performance of densely packed adsorbent beds, AIChE J., 41 (6), 1389-1394 (1995). Sircar, S., and Hanley, B.F., Production of oxygen-enriched air by rapid pressure swing adsorption, Adsorption, 1(4), 313-320 (1995). Storti, G.; Baciocchi, R., and Morbidelli, M., Design of optimal operating conditions of simulated moving bed adsorption separation units, Ind. Eng. Chem. Res., 34(1), 288-301 (1995). Stradella, L., and Damien, L.B., Energetics of the adsorption-desorption cycles of different gases and vapours on cadmium sulphide, Adsorpt. Sci. Yechnol., 12(3), 181.-190 (1995). Sundaram, N., Axial dispersion effects on gas purification PSA periodic steady state, Chem. Eng. Commun., 133, 207-225 (1995). Yang, J., et al., Bulk separation of hydrogen mixtures by a one-column PSA process, Sep. Technol., 5(4), 239250 (1995).

993

1996 Acharya, D.; Fitch, F., and Jain, R., Some issues in operating adsorption pre-purification systems for cryogenic air separation, Sep. Sci. Technol., 31 (16), 2171-2182 (1996). Agrawal, A., and Bums, M.A., Recuperative parametric pumping in adsorptive membranes, AIChE J., 42(1), 131-146 (1996). Anderson, L.E., et al., Synthesis and optimization of a new starch-based adsorbent for dehumidification of air in a pressure-swing dryer, Ind. Eng. Chem. Res., 35(4), 1180-1187 (1996). Baronskaya, N.A., et al., Ethylene recovery from the gas product of methane oxidative coupling by temperature swing adsorption, Gas Sep. Purif., 10(1), 85-88 (1996). Bhaumik, S.; Majumdar, S., and Sircar, K.K., Hollow-fiber membrane-based rapid pressure swing adsorption, AIChE J., 42(2), 409-421 (1996). Chiang, A.S.T., An analytical solution to equilibrium PSA cycles, Chem. Eng. Sci., 51(2), 207-216 (1996). Davies, R.; Hewerdine, S., and Chapman, J., Fatigue cracking of adsorber on hydrogen PSA unit, Ammonia Plant Safety, 36, 148-159 (1996). Diagne, D.; Goto, M., and Hirose, T., Numerical analysis of a dual refluxed PSA process during simultaneous removal and concentration of carbon dioxide dilute gas from air, J. Chem. Technol. Biotechnol., 65(1), 29-38 (1996). Follin, S.; Goetz, V., and Guillot, A., Influence of microporous characteristics of activated carbons on the performance of an adsorption cycle for refrigeration, Ind. Eng. Chem. Res., 35(8), 2632-2639 (1996). Hassan, M.M.; Loughlin, K.F., and Biswas, M.E., Optimization of continuous countercurrent adsorption systems, Sep. Technol., 6(1 ), 19-28 (1996). Hayashi, S.; Kawai, M., and Kaneko, T., Dynamics of high purity oxygen PSA, Gas Sep. Purif., 10(1), 19-24 (1996). Kumar, R., Vacuum swing adsorption process for oxygen production: A historical perspective and review, Sep. Sci. Technol., 31 (7), 877-894 (1996). Kvamsdal, H.M., and Hertzberg, T., A preliminary design study of a multicomponent PSA gas separation system, Chem. Eng. Process., 35(3), 2 i3-224 (1996). LaCava, A.I., and Lemcoff, N.O., Single bed pressure swing adsorption process to generate high purity nitrogen, Gas Sep. Purif., 10(2), 113-116 (1996). Lee, H., et al., Adsorption process dynamics with vacuum purge and atmospheric blowdown, Sep. Sci. Technol., 31(12), 1741-1770(1996). Liu, Y., and Ritter, J.A., Pressure swing adsorption/solvent vapor recovery: Process dynamics and parametric study, Ind. Eng. Chem. Res., 35(7), 2299-2312 (1996). Mahle, J.J.; Friday, D.K., and LeVan, M.D., Pressure swing adsorption for air purification: Temperature cycling and role of weakly adsorbed carrier gas, Ind. Eng. Chem. Res., 35(7), 2342-2354 (1996). Mazzotti, M.; Storti, G., and Morbidelli, M., Robust design of countercurrent adsorption separation: Nonstoichiometric systems, AIChE J., 42(10), 2784-2796 (1996). Pacalowska, B.; Whysall, M., and Narasimhan, M.V., Improve hydrogen recovery from refinery offgases, Hydrocarbon Process., 75(11), 55-59 (1996). Ratto, M.; Lodi, G., and Costa, P., Sensitivity analysis of a fixed-bed gas-solid TSA: The problem of design with uncertain models, Sep. Technol., 6(4), 235-246 (1996). Ruthven, D.M., and Thaeron, C., Performance of a parallel passage adsorbent contactor, Gas Sep. Purif., 10(1), 63-74(1996). Schweiger, T.A.J., A design method for adsorption bed capacity: Steam-regenerated adsorbers, Ind. Eng. Chem. Res., 35(6), 1929-1934 (1996). Sircar, S., Production of oxy-rich air by RPSA for combustion use, Adsorption, 2(4), 323-326 (1996). Smith, D.L., Optimize solid bed adsorption systems, Hydrocarbon Process., 75(5), 129-132 (1996). Sun, L.M.; Le Quere, p., and LeVan, M.D., Numerical simulation of diffusion-limited PSA process models by finite difference methods, Chem. Eng. Sci., 51(24), 5341-5352 (1996). Suzuki, M., et al., Piston-driven ultra-rapid pressure swing adsorption, Adsorption, 2(2), 111-120 (1996). 1997 Boger, T." Salden, A., and Eigenberger, G., A combined vacuum and temperature swing adsorption process for the recovery of amine from foundry air, Chem. Eng. Process., 36(3), 231-242 (1997). Brennsteiner, A., et al., Environmental pollution control devices based on novel forms of carbon: Heavy metals, Energy Fuels, 11(2), 348-353 (1997).

994

Cao, Y., and Fenouil, L.A., Comments and reply on "Design of entrained-flow and moving-, packed-, and fluidized-bed sorption systems: Grain-model kinetics for hot coal-gas desulfurization with limestone", Ind. Eng. Chem. Res., 36(11), 5037-5039 (1997). Han, C., and Harrison, D.P., Multicycle performance of a single-step process for hydrogen production, Sep. Sci. Yechnol., 32(1 ), 681-697 (1997). Huang, W.C., and Chou, C.T., A moving finite element simulation of a pressure swing adsorption process, Comput. Chem. Eng., 21(3), 301-316 (1997). Kvamsdal, H.M., and Hertzberg, T., Optimization of PSA systems: Studies on cyclic steady-state convergence, Comput. Chem. Eng., 21 (8), 819-832 (1997). Liu, Y., and Ritter, J.A., Fractional factorial design study of a pressure swing adsorption-solvent vapor recovery process, Adsorption, 3(2), 151-164 (1997). Liu, Y., and Ritter, J.A., Evaluation of model approximations in simulating pressure swing adsorption-solvent vapor recovery, Ind. Eng. Chem. Res., 36(5), 1767-1778 (1997). Lizzio, A.A., and Pollack, N.R., Symposium on production and use of carbon-based materials for environmental cleanup: An introduction, Energy Fuels, 11(2), 249 (1997). Lu, Z.P., and Ching, C.B., Dynamics of simulated moving-bed adsorption separation processes, Sep. Sci. Technol., 32(12), 1993-2010 (1997). Malek, A., and Farooq, S., Study of a six-bed pressure swing adsorption process, AIChE J., 43(10), 2509-2523 (1997). Mazzotti, M.; Storti, G., and Morbidelli, M., Robust design of countercurrent adsorption separation processes: Desorbent in the feed, AIChE J., 43(1), 64-72 (1997). Pezolt, D.J.; Collick, S.J., and Johnson, H.A., Pressure swing adsorption for VOC recovery at gasoline loading terminals, Environ. Prog., 16( 1), 16-19 (1997). Pigorini, G., and LeVan, M.D., Equilibrium theory for pressure swing adsorption: Purification and enrichment in layered beds, Ind. Eng. Chem. Res., 36(6), 2296-2305 (1997). Pigorini, G., and LeVan, M.D., Equilibrium theory for pressure swing adsorption: Separation and purification in two-component adsorption, Ind. Eng. Chem. Res., 36(6), 2306-2319 (1997). Ruthven, D.M., and Thaeron, C., Performance of a parallel passage adsorbent contactor, Sep. Purif. Technol., 12(1), 43-60 (1997). Serbezov, A., and Sotirchos, S.V., Multicomponent transport effects in sorbent particles under pressure swing conditions, Ind. Eng. Chem. Res., 36(8), 3002-3012 (1997). Sheng, P., and Costa, C.A.V., Modelling of selectivity inverted two-column thermal direct mode parametric pumping, Sep. Purif. Yechnol., 12(1), 81-95 (1997). Shirley, A.I., and Lemcoff, N.O., High-purity nitrogen by pressure-swing adsorption, AIChE J., 43(2), 419-424 (1997). Singh, K., and Jones, J., Numerical simulation of air separation by piston-driven pressure swing adsorption, Chem. Eng. Sci., 52(18), 3133-3146 (1997). Singh, R.P., and Singh, D., Effect of cosolvent (acetone) on the adsorption and movement of cypermethrin in Indian soils, Adsorpt. Sci. Technol., 15(2), 135-145 (1997). Smith, E.H., Wave front analysis for design of fixed-bed adsorbers, Chem. Eng. Commun., 159, 17-38 (1997). Subramanian, D., and Ritter, J.A., Equilibrium theory for solvent vapor recovery by pressure swing adsorption: Analytical solution for process performance, Chem. Eng. Sci., 52(18), 3147-3160 (1997). Warmuzinski, K., and Tanczyk, M., Multicomponent pressure swing adsorption: Modelling of large-scale PSA installations, Chem. Eng. Process., 36(2), 89-100 (1997). Yang, J.; Lee, C.H., and Chang, J.W., Separation of hydrogen mixtures by a two-bed pressure swing adsorption process using zeolite 5A, Ind. Eng. Chem. Res., 36(7), 2789-2798 (1997). Zumkeller, H.J., and Bart, H.J., Influence of residual loadings on mass transfer and efficiency in sorption cycles, Chem. Eng. Process., 36(6), 459-468 (1997).

995

LIQUID-PHASE

ADSORPTION

1967 Gehrhardt, H.M., and Kyle, B.G., Fixed-bed liquid-phase drying with molecular sieve adsorbent, Ind. Eng. Chem. Process Des. Dev., 6(3), 265-267 (1967). Morton, E.L., and Murrill, P.W., Analysis of liquid-phase adsorption fractionation in fixed beds, AIChE J., 13(5), 965-972 (1967). Roberts, P.V., and York, R., Adsorption of normal paraffins from binary liquid solutions by 5A molecular sieve adsorbent, Ind. Eng. Chem. Process Des. Dev., 6(4), 516-525 (1967). 1970 Miller, C.O.M., and Clump, C.W., Liquid-phase adsorption study of rate of diffusion of phenol from aqueous solution into activated carbon, AIChE J., 16(2), 169-172 (1970). 1971 Erskine, D.B., and Schuliger, W.G., Activated carbon processes for liquids, Chem. Eng. Prog., 67(11), 41-44 (1971). Misic, D.M., and Smith, J.M., Adsorption of benzene in carbon slurries, Ind. Eng. Chem. Fund., 10(3), 380-389 (1971). Sircar, S., and Myers, A.L., Thermodynamic consistency test for adsorption from binary liquid mixtures on solids, AIChE J., 17(1), 186-190 (1971). 1972 Kim, Y.S., and Zeitlin, H., Separation of zinc and copper from seawater by adsorption colloid flotation, Sep. Sci., 7(1), 1-12 (1972). Radke, C.J., and Prausnitz, J.M., Adsorption of organic solutes from dilute aqueous solution on activated carbon, Ind. Eng. Chem. Fund., 11(4), 445-451 (1972). Radke, C.J., and Prausnitz, J.M., Thermodynamics of multi-solute adsorption from dilute liquid solutions, AIChE J., 18(4), 761-768 (1972). Sircar, S.; Novosad, J., and Myers, A.L., Adsorption from liquid mixtures on solids, Ind. Eng. Chem. Fund., 11(2), 249-254 (1972). 1973 Butler, P., Activated carbon for organics removal from aqueous solutions, Process Eng. (London), June, 134-135 (1973). Furusawa, T., and Smith, J.M., Fluid-particle and intraparticle mass-transport adsorption rates in slurries, Ind. Eng. Chem. Fund., 12(2), 197-203 (1973). Furusawa, T., and Smith, J.M., Mass-transfer adsorption rates in slurries by chromatography, Ind. Eng. Chem. Fund., 12(3), 360-364 (1973). Minka, C., and Myers, A.L., Adsorption from ternary liquid mixtures on solids, AIChE J., 19(3), 453-459 (1973). Sircar, S., and Myers, A.L., Prediction of adsorption at liquid-solid interface from adsorption isotherms of pure unsaturated vapors, AIChE J., 19(1), 159-166 (1973). 1974 Dernini, S.; DeSantis, R., and Pasquinucci, A., Liquid-phase adsorption of xylenes by molecular sieves, Chem. Process., June, 20, 22 (1974). Furusawa, T., and Smith, J.M., Intraparticle mass transport in slurries by dynamic adsorption studies, AIChE J., 20(1), 88-93 (1974). Komiyama, H.; Furusawa, T., and Smith, J.M., Effectiveness factors for adsorption in slurries, Ind. Eng. Chem. Fund., 13(3), 293-296 (1974). Malati, M.A., and Mazza, R.J., Mechanism of adsorption of alkali metal ions on silica, Powder Technol., 9, 107110(1974). 1975 Deineko, P.S., et al., Zeolite activation for liquid-phase adsorptive separation of hydrocarbons, Int. Chem. Eng., 15(2), 339-341 (1975). Prober, R.; Pyeha, J.J., and Kidon, W.E., Interaction of activated carbon with dissolved oxygen, AIChE J., 21(6), 1200-1204 (1975).

996

Schlunder, E.U., et al., Scale-up of activated carbon columns for water purification based on batch-test results, Chem. Eng. Sci., 30(5), 529-548 (1975). Skrivanek, J., and Hostomsky, J., Model of isothermal adsorption in a cascade of continuous flow stirred-tank reactors, Chem. Eng. Commun., 2( 1), 109-114 (1975). 1976 Foti, G., and Nagy, L.G., Critical evaluation of the linear approximation of the individual adsorption isotherms of binary liquid mixtures on solid surfaces, Period. Polytech. Chem. Eng., 20, 107-114 (1976). Niiyama, H., and Smith, J.M., Adsorption of nitric oxide in aqueous slurries of activated carbon, AIChE J., 22(6), 961-970 (1976). 1977 Alexander, F., and McKay, G., Kinetics of the removal of basic dye from effluent using silica, Chem. Eng. (Rugby, Engl.), April, 243-246 (1977). Hsieh, J.S.C.; Turian, R.M., and Tien, C., Multicomponent liquid-phase adsorption in fixed beds, AIChE J., 23(3), 263-275 (1977). Myers, A.L., and Moser, F., Slurry sorption separations, Chem. Eng. Sci., 32(5), 529-534 (1977). Niiyama, H., and Smith, J.M., Adsorption rates of oxygen in aqueous slurries of activated carbon, AIChE J., 23(4), 592-596 (1977). Rizzo, J.L., and Shepherd, A.R., Treating industrial wastewater with activated carbon, Chem. Eng. (N.Y.), 3 January, 95-100 (1977). 1978 Anon., Carbon regeneration in wastewater treatment, Processing (Sutton, Engl.), July, 36 (1978). Chester, A., Wastewater treatment by carbon adsorption, Processing (Sutton, Engl.), February, 69-70 (1978). Jossens, L., et al., Thermodynamics of multi-solute adsorption from dilute aqueous solutions, Chem. Eng. Sci., 33(8), 1097-1106 (1978). Mahajan, O.P.; Youssef, A., and Walker, P.L., Surface-treated activated carbon for removal of ammonia from water, Sep. Sci. Technol., 13(6), 487-500 (1978). Ramachandran, P.A., and Smith, J.M., Adsorption of hydrogen sulfide in a slurry reactor, Ind. Eng. Chem. Fund., 17(1), 17-23 (1978). 1979 Brunovska, A., and Brunovsky, P., Optimal temperature control of a stirred adsorber, Chem. Eng. Sci., 34(3), 379-386 (1979). Kars, R.L.; Best, R.J., and Drinkenburg, A.A.H., Sorption of propane in slurries of active carbon in water, Chem. Eng. J., 17(3), 201-210 (1979). 1980 Hutchins, R.A., Liquid-phase adsorption: Maximising performance, Chem. Eng. (N.Y.), 25 February, 101-110 (1980). Mahajan, O.P.; Castilla, C.M., and Walker, P.L., Surface-treated activated carbon for removal of phenol from water, Sep. Sci. Yechnol., 15(10), 1733-1752 (1980). McKay, G., Surface mass transfer during acid dye removal from water using carbon, Chem. Eng. (Rugby, Engl.), April, 219-221,224 (1980). McKay, G., and Allen, S.J., Surface mass transfer processes using peat as adsorbent for dyestuffs, Can. J. Chem. Eng., 58, 521-526 (1980). McKay, G., and Poots, V.J.P., Kinetics and diffusion processes in colour removal from effluent using wood as an adsorbent, J. Chem. Technol. Biotechnol., 30, 279-292 (1980). Sylvester, N.D., and Dianat, S., Oxygen absorption and adsorption in aqueous slurries of powdered activated carbon, Ind. Eng. Chem. Process Des. Dev., 19(1), 199-201 (1980). 1981 Bhattacharyya, D., et al., Separation of toxic organotin compounds from aqueous solution by adsorption, Sep. Sci. Yechnol., 16(5), 495-504 (1981). Fritz, W.; Merk, W., and Schlunder, E.U., Competitive adsorption of two dissolved organics onto activated carbon, Chem. Eng. Sci., 36(4), 721-758 (1981). Fritz, W.; Merk, W., and Schlunder, E.U., Simultaneous adsorption of organic solutes in water by activated carbon, Int. Chem. Eng., 21(3), 384-399 (1981).

997

Goto, S.; Goto, M., and Teshima, H., Simplified evaluations of mass-transfer resistances from batch-wise adsorption and ion-exchange data, Ind. Eng. Chem. Fund., 20(4), 368-375 (1981). Gupta, R.K.; Kunzru, D., and Saraf, D.N., Liquid-phase adsorption of binary and ternary systems of n-paraffins on LMS-5A, Ind. Eng. Chem. Fund., 20(1), 28-34 (1981). Hasanain, M.A., and Hines, A.L., Application of the adsorption potential theory to adsorption of carboxylic acids from aqueous solutions onto a macroreticular resin, Ind. Eng. Chem. Process Des. Dev., 20(4), 621-625 (1981). Jaroniec, M., and Derylo, A., Simple relationships for predicting multi-solute adsorption from dilute aqueous solutions, Chem. Eng. Sci., 36(6), 1017-1020 (1981). McKay, G., Design models for adsorption systems in wastewater treatment, J. Chem. Technol. Biotechnol., 31, 717-731 (1981). McKay, G., and McConvey, I.F., External mass transfer of basic and acidic dyes on wood, J. Chem. Technol. Biotechnol., 31,401-408 (1981). Perineau, F., and Gaset, A., Adsorption of ionic and non-ionic surfactants by wool charring waste, J. Chem. Technol. Biotechnol., 31,395-400 (1981). 1982 Choudhary, V.R., and Vaidya, S.H., Adsorption of copper nitrate from solution on silica gel, J. Chem. Technol. Biotechnol., 32, 888-892 (1982). Ghim, Y.S., and Chang, H.N., Adsorption characteristics of glucose and fructose in ion-exchange resin columns, Ind. Eng. Chem. Fund., 21(4), 369-374 (1982). Hashimoto, K.; Miura, K., and Watanabe, T., Kinetics of thermal regeneration reaction of activated carbons used in wastewater treatment, AIChE J., 28(5), 737-746 (1982). Heitkamp, D., and Wagener, K., Uranium recovery by adsorption from seawater, Ind. Eng. Chem. Process Des. Dev., 21(4), 781-784 (1982). Kaneko, S., Adsorption of several dyes from aqueous solutions on silica-containing complex-oxide gels, Sep. Sci. Technol., 17(13), 1499-1510 (1982). Mansour, A.; von Rosenberg, D.U., and Sylvester, N.D., Numerical solution of liquid-phase multicomponent adsorption in fixed beds, AIChE J., 28(5), 765-772 (1982). McKay, G., Adsorption of dyestuffs from aqueous solutions with activated carbon, J. Chem. Technol. Biotechnol., 32,759-780 (1982). Pedram, E.; Hines, A.L.; Poulson, R.E., and Cooney, D.O., Adsorption of organics from a true in-situ oil-shale retort water on activated carbon in packed beds, Chem. Eng. Commun., 15(5), 291-304 (1982). Pedram, E.O.; Hines, A.L., and Cooney, D.O., Kinetics of adsorption of organics from an above-ground oil-shale retort water, Chem. Eng. Commun., 19(1), 167-176 (1982). Perineau, F.; Molinier, J., and Gaset, A., Adsorption of ionic dyes on charred plant material, J. Chem. Technol. Biotechnol., 32, 749-758 (1982). Wang, S.C., and Tien, C., Multicomponent liquid-phase adsorption in fixed beds, AIChE J., 28(4), 565-573 (1982). 1983 Asteimer, L., et al., Development of sorbers for recovery of uranium from seawater, Sep. Sci. Technol., 18(4), 307-340 (1983). Borwanker, R.P., and Wasan, D.T., Kinetics of adsorption of surface active agents at gas-liquid surfaces, Chem. Eng. Sci., 38(10), 1637-1650 (1983). Keaton, M.M., and Bourke, M.J., Activated carbon system cuts foaming and amine losses, Hydrocarbon Process., August, 71-73 (1983). McKay, G., Adsorption of dyestuffs from aqueous solutions using activated carbon, J. Chem. Technol. Biotechnol., 33A(4), 196-218 (1983). McKay, G., Adsorption of acidic and basic dyes onto activated carbon in fluidised beds, Chem. Eng. Res. Des., 61, 29-36 (1983). Sheindorf, C.; Rebhun, M., and Sheintuch, M., Prediction of breakthrough curves from fixed-bed adsorbers with Freundlich-type multisolute isotherm, Chem. Eng. Sci., 38(2), 335-342 (1983). Sutikno, T., and Himmelstein, K.J., Desorption of phenol from activated carbon by solvent regeneration, Ind. Eng. Chem. Fund., 22(4), 420-425 (1983).

998

1984 Andrews, G.F., and Elkcechen, S., Solid adsorbents in batch fermentations, Chem. Eng. Commun., 29(1), 139152 (1984). Baba, Y., et al., Adsorption equilibrium of copper and cadmium on silica gel from ammoniacal solutions, Sep. Sci. Technol., 19(6), 417-428 (1984). Belfort, G., et al., Selective adsorption of organic homologues onto activated carbon from dilute aqueous solutions, AIChE J., 30(2), 197-207 (1984). Blasinski, H., and Kazmierczak, J., Surface diffusion on active carbon in adsorption from solutions, Chem. Eng. Commun., 25(1), 351-362 (1984). Broughton, D.B., Production-scale adsorptive separations of liquid mixtures by simulated moving-bed technology, Sep. Sci. Technol., 19(11), 723-732 (1984). Jaroniec, M.; Dabrowski, A., and Toth, J., Multilayer single-solute adsorption from dilute solutions on energetically heterogeneous solids., Chem. Eng. Sci., 39(1), 65-70 (1984). Larson, A.C., and Tien, C., Multicomponent liquid-phase adsorption in batch, Chem. Eng. Commun., 27(5), 339379(1984). Mathews, A.P., and Weber, W.J., Modeling and parameter evaluation for adsorption in slurry reactors, Chem. Eng. Commun., 25(1), 157-172 (1984). McKay, G., Batch adsorption of dyestuffs from aqueous solution using activated carbon, Chem. Eng. J., 27(3), 187-196 (1983); 28(2), 95-104 (1984). McKay, G., Adsorption of basic dye onto silica from aqueous solution-solid diffusion model, Chem. Eng. Sci., 39(1), 129-138 (1984). McKay, G., Two-resistance mass transfer models for adsorption of dyestuffs from aqueous solutions using activated carbon, J. Chem. Technol. Biotechnol., 34A(6), 294-310 (1984). McKay, G., Analytical solution for adsorption of basic dye on silica, AIChE J., 30(4), 692-697 (1984). McKay, G., Mass transfer processes during the adsorption of solutes in aqueous solutions in batch and fixed bed adsorbers, Chem. Eng. Res. Des., 62, 235-246 (1984). McKay, G., and Allen, S.J., Pore diffusion model for dye adsorption onto peat in batch adsorbers, Can. J. Chem. Eng., 62(3), 340-345 (1984). McKay, G.; Allen, S.J.; McConvey, I.F., and Waiters, J.H.R., External mass transfer and homogeneous solidphase diffusion effects during adsorption of dyestuffs, Ind. Eng. Chem. Process Des. Dev., 23(2), 221-226 (1984). Milestone, N.B., and Bibby, D.M., Adsorption of alcohols from aqueous solution by ZSM-5, J. Chem. Technol. Biotechnol., 34A(2), 73-79 (1984). Rivera-Utrilla, J., Adsorption of radionuclides on activated carbons from aqueous solutions, J. Chem. Technol. Biotechnol., 34A(5), 243-250 (1984). Tsezos, M., and Noh, S.H., Extraction of uranium from seawater using biological-origin adsorbents, Can. J. Chem. Eng., 62(4), 559-561 (1984). 1985 Asfour, H.M.; Nassar, M.M.; Fadali, O.A., and E1-Geundi, M.S., Adsorption of textile dyes, J. Chem. Technol. Biotechnol., 35A(1), 21-35 (1985). Baba, Y.; Doi, S.; Inoue, K., and Nakamori, I., Adsorption rate of copper (II) from aqueous ammonium nitrate solution on silica gel, Solvent Extr. Ion Exch., 3(5), 741-752 (1985). Bui, S.; Verykios, X., and Mutharasan, R., Removal of ethanol from fermentation broths by adsorption, Ind. Eng. Chem. Process Des. Dev., 24(4), 1209-1213 (1985). Cloutier, J.N.; Leduy, A., and Ramalho, R.S., Peat adsorption of herbicide 2,4-D from wastewaters, Can. J. Chem. Eng., 63(2), 250-257 (1985). Desai, N.J., and Do, D.D., Adsorption of organic solutes into activated carbons (batch studies), Chem. Eng. Commun., 39, 101-126 (1985). Friedrich, M.; Seidel, A., and Gelbin, D., Measuring adsorption rates from an aqueous solution, AIChE J., 31(2), 324-327 (1985). Hashimoto, K.; Miura, K., and Kyotani, S., Regeneration of activated carbons used in wastewater treatment by a moving-bed regenerator, AIChE J., 31(12), 1986-1996 (1985). Hinds, B.C., and McCoy, B.J., Perturbation theory and moments for nonlinear Freundlich desorption in continuous-flow well-stirred vessel, Chem. Eng. Commun., 37(1), 265-274 (1985).

999

Jayaraj, K., and Tien, C., Characterization of adsorption affinity of unknown substances in aqueous solutions, Ind. Eng. Chem. Process Des. Dev., 24(4), 1230-1239 (1985). Kosmulski, M.; Jaroniec, M., and Szczypa, J., Liquid-solid interfaces: Kinetics of isotope exchange, Adsorpt. Sci. Technol., 2(2), 97-120 (1985). Laszlo, K.; Nagy, L.G.; Foti, G., and Schay, G., Investigation of HPLC packings by liquid mixture adsorption, Period. Polytech. Chem. Eng., 29(2), 73-86 (1985). Malmary, G.; Perineau, F.; Molinier, J., and Geset, A., Continuous process for dye removal from liquid effluents using carbonised wool waste, J. Chem. Technol. Biotechnol., 35A(8), 431-437 (1985). Mansour, A.R.; Shahalam, A.B., and Sotari, M.A., Parametric sensitivity study of multicomponent adsorption in agitated tanks, Sep. Sci. Technol., 20(1), 1-20 (1985). McConvey, I.F., and McKay, G., Mass transfer model for adsorption of basic dyes on woodmeal in agitated batch adsorbers, Chem. Eng. Process., 19(5), 267-276 (1985). McKay, G., Adsorption of dyestuffs from aqueous solutions using activated carbon, AIChE J., 31(2), 335-339 (1985). McKay, G., and McConvey, I.F., Adsorption of acid dye onto woodmeal by solid diffusional mass transfer, Chem. Eng. Process., 19(6), 287-296 (1985). Sircar, S., Adsorption of dilute hydrocarbon solutes from aqueous solutions on hetergeneous adsorbents, Adsorpt. Sci. Technol., 2(1), 1-8 (1985). 1986 Akser, M.; Wan, R.Y., and Miller, J.D., Gold adsorption from alkaline aurocyanide solution by neutral polymeric adsorbents, Solvent Extr. Ion Exch., 4(3), 531-546 (1986). Bindal, R.C., and Misra, B.M., Separation of binary liquid systems by sorption: Comparison with pervaporation, Sep. Sci. Yechnol., 21(10), 1047-1058 (1986). Ching, C.B.; Ho, C., and Ruthven, D.M., An improved adsorption process for the production of high-fructose syrup, AIChE J., 32(11), 1876-1880 (1986). Cornel, P.; Sontheimer, H.; Summers, R.S., and Roberts, P.V., Sorption of dissolved organics from aqueous solution by polystyrene resins, Chem. Eng. Sci., 41(7), 1791-1810 (1986). Dabrowski, A.; Jaroniec, M., and Oscik, J., Surface activity coefficients in adsorption at liquid-solid interface, Adsorpt. Sci. Yechnol., 3(4), 221-228 (1986). Garbacz, J.K.; Biniak, S., and Swiatkowski, A., Description of adsorption of iodine from non-aqueous solutions on active carbon, Adsorpt. Sci. Technol., 3(2), 61-64 (1986). Goworek, J., Adsorption of alcohols from multicomponent solutions onto silica gel, Adsorpt. Sci. Technol., 3(3), 141-148(1986). Hidajat, K.; Ching, C.B., and Ruthven, D.M., Numerical simulation of a semi-continuous countercurrent adsorption unit for fructose-glucose separation, Biochem. Eng. J., 33(3), B55-B62 (1986). Lee, H.S., and Ihm, S.K., Statistical thermodynamic analysis on the sorption selectivity for binary liquid mixture on solid surface, Chem. Eng. Commun., 48, 21-34 (1986). Omichi, H., et al., New type of amidoxime-group-containing adsorbent for recovery of uranium from seawater, Sep. Sci. Technol., 20(2), 163-178 (1985); 21(6), 563-574 (1986). Seidel,, A., et al., Equilibrium adsorption of two-component organic solutes from aqueous solutions on activated carbon, Adsorpt. Sci. Yechnol., 3(3), 189-200 (1986). Sircar, S., and Myers, A.L., Liquid adsorption operations: Equilibrium, kinetics, column dynamics, and applications, Sep. Sci. Technol., 21(6), 535-562 (1986). Tsezos, D.E.; Baird, M.H.I., and Shemilt, L.W., Adsorptive treatment with microbial biomass of 226Racontaining wastewaters, Biochem. Eng. J., 32(2), B29-B38 (1986). Utrilla, J.R.; Garcia, M.A.F.; Mingorance, M.D., and Toledo, I.B., Adsorption of lead on activated carbons, J. Chem. Technol. Biotechnol., 36(2), 47-52 (1986). Van Deventer, J.S.J., Kinetic model for reversible adsorption of gold cyanide on activated carbon, Chem. Eng. Commun., 44, 257-274 (1986). 1987 AI-Zaid, K., et al., Adsorption of aromatic compounds and mixtures from solutions on molecular sieve 13X, Adsorpt. Sci. Technol., 4(3), 185-210 (1987). Cooney, D.O., An experiment in liquid-phase adsorption fundamentals, Chem. Eng. Educ., 21(4), 200-203 (1987).

1000

Garbacz, J.K., et al., Theoretical equations for adsorption from binary non-electrolyte solutions with limited miscibility, Adsorpt. Sci. Yechnol., 3(4), 253-260 (1986); 4(1), 105-112 (1987). Gomez-Jimenez, L.; Rodriguez, A.G.; Gonzalez, J.L., and Guijosa, A.N., Study of kinetics of adsorption by activated carbons of2,4,5-trichlorophenoxyacetic acid from aqueous solution, J. Chem. Technol. Biotechnol., 37(4) 271-280 (1987); 38(1), 1-14 (1987). Gonzalez-Pradas, E., et al., Adsorption of para-substituted anilines from cyclohexane solution on sepiolite, Adsorpt. Sci. Technol., 4(1), 79-86 (1987). Gonzalez-Pradas, E.; Sanchez, M.V.; Viciana, M.S.; Rey-Bueno, F.D., and Rodriguez, A.G., Adsorption of thiram from aqueous solution on activated carbon and sepiolite, J. Chem. Technol. Biotechnol., 39(1), 19-28 (1987). Goworek, J., Adsorption of alcohols from ternary solutions in benzene and n-heptane on silica gel, Adsorpt. Sci. Technol., 4(1), 113-120 (1987). Goworek, J., Effects of surface heterogeneity in adsorption from binary solutions on modified silica gel, Adsorpt. Sci. Yechnol., 4(4), 224-229 (1987). Helmy, A.K.; Ferreiro, E.A., and DeBussetti, S.G., Adsorption of acid on charcoal, Adsorpt. Sci. Technol., 4(3), 211-216(1987). Khare, S.K.; Panday, K.K.; Srivastava, R.M., and Singh, V.N., Removal of Victoria Blue from aqueous solution by fly ash, J. Chem. Technol. Biotechnol., 38(2), 99-104 (1987). King, C.J., et al., Use of adsorbents for recovery of acetic acid from aqueous solutions, Sep. Purif. Methods, 16(1), 31-102 (1987). Kumar, S.; Upadhyay, S.N., and Upadhya, Y.D., Removal of phenols by adsorption on fly ash, J. Chem. Technol. Biotechnol., 37(4), 281-290 (1987). Kuo, J.F., et al., Kinetics of adsorption of organics from water produced during in-situ tar sands experiments, Chem. Eng. Commun., 50, 201-212 (1987). Marutovsky, R.M., and Bulow, M., Sorption kinetics of multicomponent gaseous and liquid mixtures on porous sorbents (review paper), Gas Sep. Purif., 1(2), 66-76 (1987). McConvey, I.F., Sorption of vitamin B12 from aqueous solution, Chem. Eng. Res. Des., 65(3), 231-233 (1987). McKay, G., Mass transport processes for adsorption of dyestuffs onto chitin, Chem. Eng. Process., 21(1), 41-52 (1987). McKay, G., and AI Duri, B., Simplified model for equilibrium adsorption of dyes from mixtures using activated carbon, Chem. Eng. Process., 22(3), 145-156 (1987). McKay, G.; Otterburn, M.S., and Aga, J.A., Pore diffusion and external mass transport during dye adsorption on to Fuller's earth and silica, J. Chem. Technol. Biotechnol., 37(4), 247-256 (1987). Price, P.E., and Danner, R.P., Prediction of multicomponent liquid adsorption equilibria, AIChE J., 33(4), 551557 (1987). Van Deventer, J.S.J., Parametric sensitivity study of adsorption in periodic countercurrent cascade of stirred tanks, Sep. Sci. Technol., 22(7), 1737-1760 (1987). 1988 AI Duri, B., and McKay, G., Basic dye adsorption on carbon using a solid-phase diffusion model, Chem. Eng. J., 38(1), 23-32 (1988). Friedrich, M.; Seidel, A., and Gelbin, D., Kinetics of adsorption of phenol and indole from aqueous solutions on activated carbons, Chem. Eng. Process., 24(1), 33-38 (1988). Kobuke, Y., et al., Composite fiber adsorbent for rapid uptake of uranyl frown seawater, Ind. Eng. Chem. Res., 27(8), 1461-1466 (1988). Koske, P.H.; Ohlrogge, K., and Peinemann, K.V., Uranium recovery from seawater by adsorption, Sep. Sci. Yechnol., 23(12), 1929-1940 (1988). Lee, C.S., and Belfort, G., Thermodynamics of multiorganic solute adsorption from dilute aqueous solution, Ind. Eng. Chem. Res., 27(6), 951-955 (1988). McKay, G., and A1-Duri, B., Prediction of bisolute adsorption isotherms using single-component data for dye adsorption onto carbon, Chem. Eng. Sci., 43(5), 1133-1143 (1988). McKay, G., and AI Duri, B., Branched-pore model applied to adsorption of basic dyes on carbon, Chem. Eng. Process., 24(1), 1-14 (1988). McKay, G., and McAleavey, G., Ozonation and carbon adsorption in three-phase fluidized bed for colour removal from peat water, Chem. Eng. Res. Des., 66(6), 531-536 (1988).

1001

Omichi, H.; Katakai, A., and Okamoto, J., Simulation of adsorption of uranium from seawater using liquid-film mass transfer controlling model, Sep. Sci. Technol., 23(10), 1133-1144 (1988). Omichi, H.; Katakai, A., and Okamoto, J., Effect of ultrasonic irradiation on recovery of uranium from seawater with adsorbents, Sep. Sci. Technol., 23(14), 2445-2450 (1988). Price, P.E., and Danner, R.P., Extension and evaluation of the Minka and Myers theory of liquid adsorption, Ind. Eng. Chem. Res., 27(3), 506-512 (1988). Sircar, S., and Myers, A.L., Determination of surface area and pore volume of adsorbents from adsorption isotherms of binary liquid mixtures, Chem. Eng. Sci., 43(12), 3259-3263 (1988). 1989 AI-Ameeri, R.S., and Owaysi, F.A., Improved purification process for liquid n-paraffins by selective adsorption on type-X zeolites, Ind. Eng. Chem. Res., 28(6), 809-814 (1989). Allen, S.J.; McKay, G., and Khader, K.Y.H., Equilibrium adsorption isotherms for basic dyes onto lignite, J. Chem. Technol. Biotechnol., 45(4), 291-302 (1989). Antonucci, V., et al., Effect of reduction on the adsorptive behaviour of the titanium dioxide/water interface, Adsorpt. Sci. Technol., 6(2), 52-63 (1989). Boniak, S.; Kazmierczak, J., and Swiatkowski, A., Adsorption of phenol from aqueous solutions on activated carbons with different oxygen contents, Adsorpt. Sci. Technol., 6(4), 182-191 (1989). Choudhary, V.R.; Mamman, A.S., and Nayak, V.S., Mass transfer of liquid cumene in ZSM-5 zeolites using novel volumetric apparatus, Ind. Eng. Chem. Res., 28(8), 1241-1246 (1989). Fernandez-Colinas, J.; Denoyel, R., and Rouquerol, J., Adsorption of iodine from aqueous solutions onto activated carbons: Correlation with nitrogen adsorption at 77K, Adsorpt. Sci. Technol., 6(1), 18-26 (1989). Garcia, A.A., and King, C.J., Use of basic polymer sorbents for acetic acid recovery from dilute aqueous solution, Ind. Eng. Chem. Res., 28(2), 204-213 (1989). Hasany, S.M.; Rehman, H., and Rashid, A., Adsorption of microamounts of silver on manganese dioxide from acid solutions, Sep. Sci. Technol., 24(15), 1363-1376 (1989). Hsiao, H.C.; Yih, S.M., and Li, M.H., Adsorption equilibrium of xylene isomers and p-diethylbenzene in the liquid phase on a Y-zeolite, Adsorpt. Sci. Technol., 6(2), 64-82 (1989). Johnson, R.L., and Chandler, B.V., Kinetic studies of adsorption of bitter principles and titratable acid from grapefruit juice, J. Chem. Technol. Biotechnol., 44(3), 225-236 (1989). Kasaoka, S., et al., Design of molecular-sieve carbon: Studies on the adsorption of various dyes in the liquid phase, Int. Chem. Eng., 29(4), 734-742 (1989). Leyva-Ramos, R., Effect of temperature and pH on adsorption of anionic detergent on activated carbon, J. Chem. Technol. Biotechnol., 45(3), 231-240 (1989). Lin, Y.S., and Ma, Y.H., Comparative chromatographic study of liquid adsorption and diffusion in microporous and macroporous adsorbents, Ind. Eng. Chem. Res., 28(5), 622-630 (1989). Okazaki, M., et al., Prediction of the breakthrough curve of a packed bed adsorber used for treatment of unknown multi-solute wastewater, Chem. Eng. Process., 26(3), 247-256 (1989). Scalabrin, G., Analysis of liquid desiccant desorption by ambient air at low temperatures, Chem. Eng. Process., 25(1), 1-14 (1989). Siri, G.J.; Galan, M.A., and McCoy, B.J., Moment analysis for stirred-tank batch adsorption with nonlinear isotherm, Comput. Chem. Eng., 13(6), 661-666 (1989). Soto, A.M., and Machuca, R.A., Adsorption of gold-thiourea complex on activated carbon, J. Chem. Technol. Biotechnol., 44(3), 219-224 (1989). Veeraraghavan, S.; Fan, L.T., and Mathews, A.P., Modeling adsorption in liquid-solid fluidized beds, Chem. Eng. Sci., 44(10), 2333-2344 (1989). 1990 Browne, T.E., and Cohen, Y., Aqueous-phase adsorption of trichloroethene and chloroform onto polymeric resins and activated carbon, Ind. Eng. Chem. Res., 29(7), 1338-1346; 29(12), 2402 (1990). Cadena, F.; Garcia, R., and Peters, R.W., Adsorption of benzene from aqueous solutions by bentonite treated with quaternary amines, Environ. Prog., 9(4), 245-253 (1990). Coughlin, R.W.; Deshaies, M.R., and Davis, E.M., Chitosan in crab shell wastes purifies electroplating wastewater, Environ. Prog., 9(1), 35-39 (1990). Duprat, F., and Gau, G., Model of competitive adsorption of alkylpyridines on silica at saturation conditions, Ind. Eng. Chem. Res., 29(7), 1424-1431 (1990).

1002

Gonzalez-Pradas, E.; Valverde-Garcia, A., and Sanchez, M.V., Removal of aromatic amines from aqueous solution by activated sepiolite, J. Chem. Technol. Biotechnol., 47(1), 15-22 (1990). Heitkamp, D., and Wagener, K., Kinetics of adsorption of uranium from seawater by humic acids, Sep. Sci. Technol., 25(5), 535-546 (1990). Holstvoogd, R.D., and van Swaaij, W.P.M., The influence of adsorption capacity on enhanced gas absorption in activated carbon slurries, Chem. Eng. Sci., 45(1), 151-162 (1990). Jain, A.K.; Jasra, R.V., and Bhat, S.G.T., Liquid-phase adsorption of olefin/paraffin mixtures on ion-exchanged X-zeolite, Sep. Sci. Technol., 25(4), 489-505 (1990). Ji, W.R., and Hou, Y.C., Prediction of equilibrium data of adsorptions from liquid mixtures, Ind. Eng. Chem. Res., 29(4), 560-564 (1990). Karve, S., and Juvekar, V.A., Gas absorption into slurries containing fine catalyst particles, Chem. Eng. Sci., 45(3), 587-594 (1990). Kawabata, N.; Sumiyoshi, K., and Tanaka, M., Selective adsorption of cationic surfactants on cross-linked poly(p-hydroxystyrene), Ind. Eng. Chem. Res., 29(9), 1889-1893 (1990). Kobuke, Y., et al., Recovery of uranium from seawater by composite fiber adsorbent, Ind. Eng. Chem. Res., 29(8), 1662-1668 (1990). Li, M.H., and Hsiao, H.C., Thermodynamics of adsorption from liquid mixtures of xylenes and p-diethylbenzene on a modified K-BaY zeolite, Adsorpt. Sci. Technol., 7(1), 9-27 (1990). Mansour, A.W., et al., An accurate numerical solution of biospecific adsorption in a stirred bath, Sep. Sci. Technol., 25(3), 347-356 (1990). Matz, M.J., and Knaebel, K.S., Criteria for selection of an adsorbent for a temperature swing process: Applied to purification of an aliphatic solvent contaminated with aromatic solutes, Sep. Sci. Technol., 25(9), 961-984 (1990). Mehra, A., Gas absorption in slurries of finite-capacity microphases, Chem. Eng. Sci., 45(6), 1525-1538 (1990). Srinivasan, M.P.; Smith, J.M., and McCoy, B.J., Supercritical fluid desorption from activated carbon, Chem. Eng. Sci., 45(7), 1885-1896 (1990). Tan, C.S., and Liou, D.C., Adsorption equilibrium of toluene from supercritical carbon dioxide on activated carbon, Ind. Eng. Chem. Res., 29(7), 1412-1416 (1990). Thurner, F., and Stietz, M., Determination of adsorption isotherms on solvent-wet adsorbents by the flow method, Int. Chem. Eng., 30(1), 36-44 (1990). Ying, W.C.; Dietz, E.A., and Woehr, G.C., Adsorptive capacities of activated carbon for organic constituents of wastewaters, Environ. Prog., 9( 1), 1-9 (1990). Yoshida, H.; Shimizu, K., and Kataoko, T., Adsorption of amine and paints on H-form resin from electrodeposition wastewater, AIChE J., 36(12), 1815-1821 (1990). 1991 AI-Ameeri, R.S.; Owaysi, F., and Hassan, M., Separation of individual n-alkanes from an isooctane solution by adsorption on crystalline urea: Equilibrium study, Ind. Eng. Chem. Res., 30(1), 202-206 (1991). AI-Duri, B., and McKay, G., Prediction of binary system for kinetics of batch adsorption using basic dyes onto activated carbon, Chem. Eng. Sci., 46(1), 193-204 (1991). AI-Duri, B., et al., Modelling of multicomponent equilibrium for the adsorption of basic dyes onto bagasse pith, Process Safety Environ. Prot., 69(B4), 246-254 (1991). Antonjuk, A.A., et al., Determination of intraparticle kinetic parameters for adsorption of binary solutions on activated carbon, Chem. Eng. Sci., 46(4), 1035-1040 (1991). Blasinski, H., et al., Application of activated carbon to decolouration in the sugar industry: Determination of the physical and chemical properties of activated carbon, Adsorpt. Sci. Technol., 7(4), 220-227 (1991). Blasinski, H., et al., Application of activated carbon to decolouration in the sugar industry: Investigations of equilibrium and adsorption kinetics, Adsorpt. Sci. Technol., 7(4), 228-238 (1991). Borthakur, S., and Srivastava, R.C., Kinetics of removal of p-toluene sulphonic acid from concentrated solution by granular activated carbon, J. Chem. Technol. Biotechnol., 51 (4), 497-506 (1991). Cooney, D.O., The importance of axial dispersion in liquid-phase fixed-bed adsorption operations, Chem. Eng. Commun., 110, 217-232 (1991 ). Dada, E.A., and Wenzel, L.A., Estimation of the adsorbent capacities from the adsorption isotherm of binary liquid mixtures on solids, Ind. Eng. Chem. Res., 30(2), 396-402 (1991). EI-Geundi, M.S., Adsorption equilibria of basic dyestuffs onto maize cob, Adsorpt. Sci. Technol., 7(3), 114-123 (1991).

1003

EI-Geundi, M.S., External mass transport processes during the adsorption of basic dyestuffs onto maize cob, Adsorpt. Sci. Technol., 7(3), 124-132 (1991). Jordi, R.G.; Young, B.D., and Bryson, A.W., Gold adsorption on activated carbon and the effect of suspended solids and dissolved silicon dioxide, Chem. Eng. Commun., 102, 127-148 (1991). Larsen, E.S., and Pilat, M.J., Design and testing of a moving bed VOC adsorption system, Environ. Prog., 10(1), 75-82 (1991). Lee, J.Y.; Westgate, P.J., and Ladisch, M.R., Water and ethanol sorption phenomena on starch, AlChE J., 37(8), 1187-1195 (1991). Macasek, F.; Keltos, D., and Matel, L., Optimization of batch adsorption of cesium and strontium by zeolite from water solutions, Solvent Extr. Ion Exch., 9(5), 865-874 (1991). McKay, G., and AI-Duri, B., Multicomponent dye adsorption onto carbon using a solid diffusion mass-transfer model, Ind. Eng. Chem. Res., 30(2), 385-395 (1991). Moon, H.; Park, H.C., and Tien, C., Adsorption of unknown substances from aqueous solutions, Chem. Eng. Sci., 46(1), 23-32 (1991). Morooka, S., et al., Modeling of an adsorption unit packed with amidoxime fiber balls for the recovery of uranium from seawater, Ind. Eng. Chem. Res., 30(1), 190-196 (1991 ). Nassar, M.M., and EI-Geundi, M.S., Comparative cost of colour removal from textile effluents using natural adsorbents, J. Chem. Technol. Biotechnol., 50(2), 257-264 (1991). Petersen, F.W., and van Deventer, J.S.J., The influence of pH, dissolved oxygen and organics on the adsorption of metal cyanides on activated carbon, Chem. Eng. Sci., 46(12), 3053-3066 (1991). Sowerby, B., and Crittenden, B.D., A vapour phase adsorption and desorption model for drying the ethanol-water azeotrope in small columns, Chem. Eng. Res. Des., 69(1), 3-13 (1991). Takeda, T., et al., Adsorption and elution in hollow-fiber-packed bed for recovery of uranium from seawater, Ind. Eng. Chem. Res., 30(1), 185-190 (1991). Urano, K., and Tachikawa, H., Process development for removal and recovery of phosphorus from wastewater by a new adsorbent: 1. Preparation method and adsorption capability of a new adsorbent, Ind. Eng. Chem. Res., 30(8), 1893-1896 (1991). Urano, K., and Tachikawa, H., Process development for removal and recovery of phosphorus from wastewater by a new adsorbent: 2. Adsorption rates and breakthrough curves, Ind. Eng. Chem. Res., 30(8), 1897-1899 (1991). Yadava, K.P.; Tyagi, B.S., and Singh, V.N., Effect of temperature on the removal of lead(II) by adsorption on china clay and wollastonite, J. Chem. Technol. Biotechnol., 51(1), 47-60 (1991). Yenkie, M.K.N., and Natarajan, G.S., Adsorption equilibrium studies of some aqueous aromatic pollutants on granular activated carbon samples, Sep. Sci. Technol., 26(5), 661-674 (1991). 1992 Balkose, D., and Baltacioglu, H., Adsorption of heavy metal cations from aqueous solutions by wool fibers, J. Chem. Technol. Biotechnol., 54(4), 393-398 (1992). Bhutani, M.M.; Mitra, A.K., and Kumari, R., Sorption and radiochemical study of Cr(VI) ions at a stannic oxide/solution interface, Adsorpt. Sci. Technol., 8(t), 44-56 (1992). Bruckner, P., et al., Adsorption and immersion of benzene in active carbons, Adsorpt. Sci. Technol., 8(1), 57-68 (1992). Choudhary, V.R.; Nayak, V.S., and Mamman, A.S., Diffusion of straight- and branched-chain liquid compounds in H-ZSM-5 zeolite, Ind. Eng. Chem. Res., 31(2), 624-628 (1992). Do, D.D.; Hu, S.G., and Nguyen, T.S., Separation of dipeptides on a reverse-phase column: Effect of non-linear intrinsic adsorption kinetics, Biochem. Eng. J., 49(3), B41-B49 (1992). Duprat, F., Model of adsorption equilibrium of pyridine ternary mixture at saturation conditions, Ind. Eng. Chem. Res., 31(8), 1907-1913 (1992). E1-Geundi, M.S., Homogeneous surface diffusion model for the adsorption of basic dyestuffs onto natural clay in batch adsorbers, Adsorpt. Sci. Technol., 8(4), 217-225 (1992). Furlan, L.T.; Chaves, B.C., and Santana, C.C., Separation of liquid mixtures of p-xylene and o-xylene in Xzeolites: The role of water content on the adsorbent selectivity, Ind. Eng. Chem. Res., 31(7), 1780-1784 (1992). Garcia-Delgado, R.A.; Cotoruelo-Minguez, L., and Rodriguez, J., Equilibrium study of single-solute adsorption of anionic surfactants with polymeric XAD resins, Sep. Sci. Technol., 27(7), 975-988 (1992).

1004

Garcia-Delgado, R.A.; Cotoruelo-Minguez, L., and Rodriguez, J.J., Adsorption of anionic surfactant mixtures by polymeric resins, Sep. Sci. Technol., 27(8), 1065-1076 (1992). Garrido, A.; Garcia, P., and Brenes, M., The recycling of table olive brine using ultrafiltration and activated carbon adsorption, J. Food Eng., 17(4), 291-305 (1992). Hasany, S.M., and Saeed, M.M., A kinetic and thermodynamic study of silver sorption onto manganese dioxide from acid solutions, Sep. Sci. Technol., 27(13), 1789-1800 (1992). Ito, Y., et al., Uranium adsorption characteristics of a circulating fluidized-bed adsorber, AIChE J., 38(6), 879886 (1992). Juang, R.S., and Su, J.Y., Sorption of copper and zinc from aqueous sulfate solutions with bis(2-ethylhexyl)phosphoric acid-impregnated macroporous resin, Ind. Eng. Chem. Res., 31(12), 2774-2779 (1992). Juang, R.S., and Su, J.Y., Separation of zinc and copper from aqueous sulfate solutions using bis(2-ethylhexyl)phosphoric acid-impregnated macroporous resin, Ind. Eng. Chem. Res., 31(12), 2779-2783 (1992). Kago, T., et al., Preparation and performance of amidoxime fiber adsorbents for recovery of uranium from seawater, Ind. Eng. Chem. Res., 31 (1), 204-209 (1992). Leitao, A., et al., Modeling of solid-liquid adsorption: Effects of adsorbent loads on model parameters, Can. J. Chem. Eng., 70(4), 690-698 (1992). Lin, S.H., Concentration-dependent diffusion of dye in adsorptive dyeing systems, J. Chem. Technol. Biotechnol., 54(4), 387-392 (1992). Lu, C.S., and Huang, S.D., Removal of organophosphorus pesticides from aqueous solution by using adsorptive bubble separation techniques, Sep. Sci. Technol., 27(13), 1733-1742 (1992). McKay, G.; Kelly, J.C., and McConvey, I.F., The adsorption of pollutants from aqueous effluents using a tworesistance mass-transfer model, Adsorpt. Sci. Technol., 8(1), 13-33 (1992). Medrzycka, K.B., The effect of surfactant adsorption on the evaporation of volatile hydrocarbons from their aqueous solutions, Sep. Sci. Technol., 27(8), 1077-1092 (1992). Mellah, A., et al., Adsorption of organic matter from a wet phosphoric acid using activated carbon: Equilibrium study, Chem. Eng. Process., 31(3), 191-194 (1992). Milonjic, S.K.; Boskovic, M.R., and Ceranic, T.S., Adsorption of uranium(VI) and zirconium(IV) from acid solutions on silica gel, Sep. Sci. Technol., 27(12), 1643-1660 (1992). Narsimhan, G., and Uraizee, F., Kinetics of adsorption of globular proteins at an air-water interface, Biotechnol. Prog., 8(3), 187-196 (1992). Payne, G.F., and Maity, N., Solute adsorption from water onto a "modified" sorbent in which the hydrogen binding site is protected from water: Thermodynamics and separations, Ind. Eng. Chem. Res., 31(8), 20242033 (1992). Pizzio, L.R.; Caceres, C.V., and Blanco, M.N., Parameters for the adsorption of tungsten from meta-tungstate solution on to alumina, Adsorpt. Sci. Technol., 8(3), 142-152 (1992). Radeke, K.H., and Hartmann, G., On the temperature dependence of adsorption of organic materials from aqueous solution, Adsorpt. Sci. Technol., 8(3), 153-156 (1992). Reed, B.E., and Nonavinakere, S.K., Metal adsorption by activated carbon: Effect of complexing ligands, competing adsorbates, ionic strength, and background electrolyte, Sep. Sci. Technol., 27(14), 1985-2000 (1992). Roy, D.; Valsaraj, K.T., and Kottai, S.A., Separation of organic dyes from wastewater by using colloidal gas aphrons, Sep. Sci. Technol., 27(5), 573-588 (1992). Rudisill, E.N.; Hacskaylo, J.J., and LeVan, M.D., Coadsorption of hydrocarbons and water in BPL activated carbon, Ind. Eng. Chem. Res., 31(4), 1122-1130 (1992). Saleem, M., et al., Selective adsorption of uranium on activated charcoal from electrolytic aqueous solutions, Sep. Sci. Yechnol., 27(2), 239-254 (1992). Sanciolo, P.; Harding, I.H., and Mainwaring, D.E., The removal of chromium, nickel, and zinc from electroplating wastewater by adsorbing colloid flotation with a sodium dodecylsulfate/dodecanoic acid mixture, Sep. Sci. Technol., 27(3), 375-388 (1992). Saska, M., et al., Continuous separation of sugarcane molasses with a simulated moving-bed adsorber: Adsorption equilibria, kinetics, and application, Sep. Sci. Technol., 27(13), 1711-1732 (1992). Silem, A., et al., Adsorption of organic matter from a wet phosphoric acid using activated carbon: Batch-contact time study and linear driving force models, Can. J. Chem. Eng., 70(3), 491-498 (1992). Sircar, S., and Rao, M.B., Kinetics and column dynamics for adsorption of bulk liquid mixtures, AIChE J., 38(6), 811-820 (1992).

1005

Tinge, J.T., and Drinkenburg, A.A.H., Absorption of gases into activated carbon-water slurries in a stirred cell, Chem. Eng. Sci., 47(6), 1337-1346 (1992). Urano, K., and Tachikawa, H., Process development for removal and recovery of phosphorus from wastewater by a new adsorbent: Desorption of phosphate and regeneration of adsorbent, Ind. Eng. Chem. Res., 31 (6), 15101513 (1992). Urano, K.; Tachikawa, H., and Kitajima, M., Process development for removal and recovery of phosphorus from wastewater by a new adsorbent: Recovery of phosphate and aluminum from desorbing solution, Ind. Eng. Chem. Res., 31(6), 1513-1515 (1992). Valsaraj, K.T., Adsorption of trace hydrophobic compounds from water on surfactant-coated alumina, Sep. Sci. Technol., 27(12), 1633-1642 (1992). Zamzow, M.J., and Murphy, J.E., Removal of metal cations from water using zeolites, Sep. Sci. Technol., 27(14), 1969-1984 (1992). 1993 Bakoyannakis, D.N., et al., Studies of alizarine adsorption from solution on to aluminium hydroxide gels, J. Chem. Technol. Biotechnol., 58(3), 247-254 (1993). Chanda, M., and Rempel, G.L., Poly(4-vinylpyridine) gel-coated on silica: High capacity and fast kinetics in uranyl sulfate recovery, Ind. Eng. Chem. Res., 32(4), 726-732 (1993). Chatzopoulos, D.; Varma, A., and Irvine, R.L., Activated carbon adsorption and desorption of toluene in the aqueous phase, AIChE J., 39(12), 2027-2041 (1993). E1-Geundi, M.S., Pore diffusion model for the adsorption of basic dyestuffs onto natural clay in batch adsorbers, Adsorpt. Sci. Technol., 9(2), 109-120 (1993). EI-Geundi, M.S., Branched-pore kinetic model for basic dyestuff adsorption onto natural clay, Adsorpt. Sci. Technol., 9(3), 199-211 (1993). El-Guendi, M.S., and Aly, I.H., Equilibrium studies during the adsorption of acid dyestuffs into maize cob, Adsorpt. Sci. Technol., 9(3), 121-129 (1993). EI-Naggar, I.M., et al., Sorption behavior of uranium and thorium on cryptomelane-type hydrous manganese dioxide from aqueous solution, Solvent Extr. Ion Exch., 11(3), 521-540 (1993). Fox, I., and Malati, M.A., An investigation of phosphate adsorption by clays and its relation to the problems of eutrophication of the River Stour in Kent, J. Chem. Yechnol. Biotechnol., 57(2), 97-108 (1993). Gonzalez-Pradas, E., et al., Removal of 3-(3,4-dichlorophenyl)-l,1 dimethylurea from aqueous solution by natural and activated bentonite, J. Chem. Technol. Biotechnol., 56(1), 67-72 (1993). Goto, A., et al., A test of uranium recovery from seawater with a packed bed of amidoxime fiber adsorbent, Sep. Sci. Technol., 28(6), 1273-1286 (1993). Gusler, G.M.; Browne, T.E., and Cohen, Y., Sorption of organics from aqueous solution onto polymeric resins, Ind. Eng. Chem. Res., 32(11), 2727-2735 (1993). Hawash, S.; Farah, J.Y., and EI-Geundi, M.S., Investigation of nickel ion removal by means of activated clay, Adsorpt. Sci. Technol., 9(4), 244-257 (1993). Kapoor, A., and Viraraghavan, T., Adsorption of mercury from wastewater by fly ash, Adsorpt. Sci. Technol., 9(3), 130-147 (1993). Lameloise, M.L., and Viard, V., Modelling and simulation of a glucose-fructose simulated moving bed adsorber, Food Bioprod. Process., 71 (C 1), 27-32 (1993). Lee, S.Y., et al., Multicomponent liquid-phase diffusion and adsorption in porous catalyst particles, Chem. Eng. Sci., 48(3), 595-608 (1993). Leitao, A., and Rodrigues, A., Modelling of solid-liquid adsorption: Effects of adsorbent heterogeneity, Chem. Eng. J., 51(3), 159-166 (1993). Lin, S.H., Adsorption of disperse dye by powdered activated carbon, J. Chem. Technol. Biotechnol., 57(4), 387391 (1993). Lin, S.H., Adsorption of disperse dye by various adsorbents, J. Chem. Technol. Biotechnol., 58(2), 159-164 (1993). Luo, C.S., and Huang, S.D., Adsorption of copper ion with metal hydroxide from ammonia solution, Sep. Sci. Yechnol., 28(6), 1253-1272 (1993). Mattuschka, B., and Straube, G., Biosorption of metals by a waste biomass, J. Chem. Technol. Biotechnol., 58(1), 57-64 (1993). Otu, E.O.; Byerley, J.J., and Robinson, C.W., Kinetic modelling of gold cyanide multi-cycle adsorption and elution using activated carbon in the presence of foulants, Can. J. Chem. Eng., 71(6), 925-933 (1993).

1006

Petersen, F.W., and Kruger, S., The adsorption of gold cyanide onto porous adsorbents: Relation between liquidphase concentration, suspended solids, and mass-transfer mechanisms, Sep. Sci. Technol., 28(10), 1849-1858 (1993). Petersen, F.W.; van Deventer, J.S.J., and Lorenzen, L., The interaction between metal cyanides, fine particles and porous adsorbents in an agitated slurry, Chem. Eng. Sci., 48(16), 2919-2926 (1993). Pizzio, L.R.; Caceres, C.V., and Blanco, M.N., Tungsten adsorption on to alumina from ammonium meta- and para-tungstate solutions: UV-visible spectra of the solutions, Adsorpt. Sci. Technol., 9(1), 36-47 (1993). Quach, T.; Koch, D.F.A., and Lawson, F., Adsorption of gold cyanide on gangue minerals, Chem. Eng. Aust., 18(3), 6-9 (1993). Rao, M.B., and Sircar, S., Concentration-thermal swing adsorption process for separation of bulk liquid mixtures, Sep. Sci. Technol., 28(10), 1837-1848 (1993). Rauf, M.A.; Hussain, M.T., and Hasany, S.M., Adsorption of europium on manganese dioxide from binary mixtures of aqueous sulfuric acid and methanol, Sep. Sci. Technol., 28(13), 2237-2246 (1993). Reunanen, J., et al., Column adsorption in multi-solute water, Chem. Eng. Process., 32(5), 291-300 (1993). Rivera-Utrilla, J., et al., Removal of tannic acid from aqueous solutions by activated carbons, Chem. Eng. J., 52(1), 37-40 (1993). Rorrer, G.L.; Hsien, T.Y., and Way, J.D., Synthesis of porous-magnetic chitosan beads for removal of cadmium ions from waste water, Ind. Eng. Chem. Res., 32(9), 2170-2178 (1993). Saleem, M., et al., Selective adsorption of europium on activated charcoal from aqueous solutions, Adsorpt. Sci. Technol., 9( 1), 1-16 (1993). Saleem, M., et al., Surface characterization and thermodynamics of adsorption of Pr, Nd and Er on alumina from aqueous solution, Adsorpt. Sci. Technol., 9(1), 17-29 (1993). Simpson, E.J., et al., Sorption equilibrium isotherms for volatile organics in aqueous solution: Comparison of head-space gas chromatography and on-line UV stirred cell results, Ind. Eng. Chem. Res., 32(10), 2269-2276 (1993). Sircar, S., Gibbsian thermodynamics and column dynamics for adsorption of liquid mixtures, Ind. Eng. Chem. Res., 32(10), 2430-2437 (1993). Yang, O.B., et al., Use of activated carbon fiber for direct removal of iodine from acetic acid solution, Ind. Eng. Chem. Res., 32(8), 1692-1697 (1993). Yoshida, H.; Okamoto, A., and Kataoka, T., Adsorption of acid dye on cross-linked chitosan fibers: Equilibria, Chem. Eng. Sci., 48(12), 2267-2272 (1993). Yu, M.C., and Middleman, S., Air entrapment during liquid infiltration of porous media, Chem. Eng. Commun., 123, 61-70 (1993). 1994 Abbasi, W.A., and Streat, M., Adsorption of uranium from aqueous solutions using activated carbon, Sep. Sci. Technol., 29(9), 1217-1230 (1994). Akman, U., and Sunol, A.K., Equilibrium theory for exsorption: A gas-liquid-adsorbent mass-transfer operation, Chem. Eng. Sci., 49(21), 3555-3564 (1994). Bohra, P.M., et al., Adsorptive recovery of water soluble essential oil components, J. Chem. Technol. Biotechnol., 60(1), 97-102 (1994). Brewster, M.D., and Passmore, R.J., Use of electrochemical iron generation for removing heavy metals from contaminated groundwater, Environ. Prog., 13(2), 143-148 (1994). Budinova, T.K., et al., Removal of metal ions from aqueous solution by activated carbons obtained from different raw materials, J. Chem. Technol. Biotechnol., 60(2), 177-182 (1994). Butani, M.M., and Kumari, R., Surface and charge characteristics of the oxide/solution interface towards chromium(VI) sorption, Adsorpt. Sci. Yechnol., 11 (3), 145-154 (1994). Cooney, D.O., and Xi, Z., Activated carbon catalyzes reactions of phenolics during liquid-phase adsorption, AIChE J., 40(2), 361-364 (1994). Dabrowski, A., et al., Application of the Dubinin-Radushkevich equation for describing adsorption from solutions on to various carbons, Adsorpt. Sci. Technol., 10, 105-122 (1994). Egawa, H., et al., Recovery of uranium from seawater: Development of amidoxime resins with high sedimentation velocity for passively driven fluidized bed adsorbers, Ind. Eng. Chem. Res., 33(3), 657-661 (1994). Ellis, J., and Korth, J., Removal of nitrogen compounds from hydrotreated shale oil by adsorption on zeolite, Fuel, 73(10), 1569-1573 (1994).

1007

Eltekova, N.A., and Eltekov, Y.A., Application of the DR and DS equations to benzene adsorption from water solutions, Adsorpt. Sci. Technol., 10, 203-210 (1994). Harmon, T.C., and Roberts, P.V., The effect of equilibration time on desorption rate measurements with chlorinated alkenes and aquifer particles, Environ. Prog., 13(1), 1-8 (1994). Hassan, N.M., The adsorption of long-chain n-paraffin from isooctane solution on crystalline urea, Sep. Technol., 4(1), 62-64 (1994). Hassan, N.M.; AI-Ameeri, R.S., and Oweysi, F.A., Adsorption of n-paraffin wax from isooctane solution on crystalline urea, Sep. Sci. Technol., 29(7), 897-906 (1994). Ho, Y.S.; Wase, D.A.J., and Forster, C.F., The adsorption of divalent copper ions from aqueous solution by Sphagnum moss peat, Process Safety Environ. Prot., 72(B3), 185-194 (1994). Hobday, M.D., et al., The use of low-rank coal-based adsorbents for the removal of nitrophenol from aqueous solution, Fuel, 73(12), 1848-1854 (1994). Holtzapple, M.T., and Brown, R.F., Conceptual design for a process to recover volatile solutes from aqueous solutions using silicalite, Sep. Technol., 4(4), 213-229 (1994). Holtzapple, M.T.; Flores, K.L., and Brown, R.F., Recovery of volatile solutes from dilute aqueous solutions using immobilized silicalite, Sep. Technol., 4(4), 230-238 (1994). Jain, A.K., and Gupta, A.K., Adsorptive drying of isopropyl alcohol on 4A molecular sieves: Equilibrium and kinetic studies, Sep. Sci. Technol., 29(11), 1461-1472 (1994). Kabay, N., Preparation of amidoxime-fiber adsorbents based on poly(methacrylonitrile) for recovery of uranium from seawater, Sep. Sci. Technol., 29(3), 375-384 (1994). Kabil, M.A., and Ghazy, S.E., Separation of some dyes from aqueous solutions by flotation, Sep. Sci. Technol., 29(18), 2533-2539 (1994). Kataoka, T.; Muto, A., and Nishiki, T., Adsorption equilibria of soluble silica into an OH-type strong anion exchange resin from a dilute solution, Chem. Eng. Res. Des., 72(6), 777-782 (1994). Kesraoui-Ouki, S.; Cheeseman, C.R., and Perry, R., Natural zeolite utilisation in pollution control: A review of applications to metal's effluents, J. Chem. Technol. Biotechnol., 59(2), 121-126 (1994). Kusakabe, K.; Goto, A., and Morooka, S., Kinetics of uranium adsorption from seawater with imidedioxime adsorbent, Sep. Sci. Technol., 29(12), 1567-1578 (1994). Leyva-Ramos, R., and Geankoplis, C.J., Diffusion in liquid-filled pores of activated carbon: Pore volume diffusion, Can. J. Chem. Eng., 72(2), 262-271 (1994). Nassar, M.M., and EI-Geundi, M.S., Studies of the dimensionless mass-transfer coefficient during the adsorption of basic and acid dyes on to bagasse pith, Adsorpt. Sci. Technol., 11(2), 73-82 (1994). Nirdosh, I.; Vogl, A.K., and Carroll, S.M., Removal of 23~ and other metals from sulphuric acid leach uranium mill solution by solvent extraction, adsorption and precipitation, Dev. Chem. Eng. Mineral Process., 2(2), 171-180(1994). Parker, W.J.; Bell, J.P., and Melcer, H., Modelling the fate of chlorinated phenols in wastewater treatment plants, Environ. Prog., 13(2), 98-104 (1994). Peng, F.F., and Di, P., Removal of arsenic from aqueous solution by adsorbing colloid flotation, Ind. Eng. Chem. Res., 33(4), 922-928 (1994). Periasamy, K., and Namasivayam, C., Process development for removal and recovery of cadmium from wastewater by a low-cost adsorbent: Adsorption rates and equilibrium studies, Ind. Eng. Chem. Res., 33(2), 317-320 (1994). Pizzio, L.R.; Cacares, C.V., and Blanco, M.N., Adsorption of tungsten and alumina from sodium tungstate solutions: Estimation of equilibrium and kinetic parameters, Adsorpt. Sci. Technol., 11(3), 133-144 (1994). Pradas, E.G., et al., Adsorption of cadmium and zinc from aqueous solution on natural and activated bentonite, J. Chem. Technol. Biotechnol., 59(3), 289-296 (1994). Pradas, E.G., et al., Adsorption of chlorophyll-a from acetone solution on natural and activated bentonite, J. Chem. Yechnol. Biotechnol., 61(2), 175-178 (1994). Rao, M.B., and Sircar, S., Liquid-phase adsorption of bulk ethanol-water mixtures by alumina, Adsorpt. Sci. Technol., 10, 93-104 (1994). Rauf, M.A., et al., Adsorption studies of europium on manganese dioxide from aqueous sulphuric acid and 1propanol mixtures, Adsorpt. Sci. Technol., 11(3), 155-160 (1994). Rauf, M.A., et al., Adsorption of europium onto titanium oxide from aqueous solutions, Adsorpt. Sci. Technol., 11(3), 187-191 (1994).

1008

Reed, B.E.; Arunachalam, S., and Thomas, B., Removal of lead and cadmium from aqueous waste streams using granular activated carbon (GC) columns, Environ. Prog., 13(1), 60-64 (1994). Sekiguchi, K., et al., Effect of seawater temperature on uranium recovery from seawater using amidoxime adsorbents, Ind. Eng. Chem. Res., 33(3), 662-666 (1994). Stang, M.; Karbstein, H., and Schubert, H., Adsorption kinetics of emulsifiers at oil-water interfaces and their effect on the mechanical emulsification, Chem. Eng. Process., 33(5), 307-312 (1994). Taha, F., et al., Electrokinetic properties of cassiterite and quartz particles in the presence of cationic, anionic and non-ionogenic surfactants, Adsorpt. Sci. Technol., 11(3), 161 - 174 (1994). Takatsuji, W., and Yoshida, H., Removal of organic acids from wine by adsorption on weakly basic ion exchangers: Equilibria for single and binary systems, Sep. Sci. Technol., 29(11), 1473-1490 (1994). Taylor, R.M., and Kuennen, R.W., Removing lead in drinking water with activated carbon, Environ. Prog., 13(1), 65-71 (1994). Yang, X.; Tsai, G.J., and Tsao, G.T., Enhancement of in-situ adsorption on the acetone-butanol fermentation by Clostridium acetobutylicum, Sep. Yechnol., 4(2), 81-92 (1994). Yoshida, H.; Nishihara, H., and Kataoka, T., Adsorption of BSA on DEAE-dextran: Equilibria, Sep. Sci. Technol., 29(17), 2227-2244 (1994). Young, D.F., and Ball, W.P., A-priori simulation of tetrachloroethane transport through aquifer material using an intraparticle diffusion model, Environ. Prog., 13(1), 9-20 (1994). Zouboulis, A.I.; Lazaridis, N.K., and Zamboulis, D., Powdered activated carbon separation from water by foam flotation, Sep. Sci. Technol., 29(3), 385-400 (1994). 1995 Ahmad, H., et al., Thermodynamics of the adsorption of cobalt on lead dioxide from aqueous solution, Adsorpt. Sci. Yechnol., 12(2), 139-150 (1995). AI-Asheh, S., and Duvnjak, Z., Adsorption of copper and chromium by Aspergillus carbonarius, Biotechnol. Prog., 11 (6), 638-642 (1995). Batabyal, D.; Sahu, A., and Chaudhuri, S.K., Kinetics and mechanism of removal of 2,4-dimethyl phenol from aqueous solutions with coal fly ash, Sep. Technol., 5(4), 179-186 (1995). Blasinski, H.; Kazmierczak, J., and Wolborska, A., Kinetics of adsorption from single and binary solutions on activated carbons with chemically different surfaces, Adsorpt. Sci. Technol., 12(4), 297-306 (1995). Brandani, S., and Ruthven, D.M., Analysis of ZLC desorption curves for liquid systems, Chem. Eng. Sci., 50(13), 2055-2060 (1995). Carriere, P.P.E.; Reed, B.E., and Cline, S.R., Retention and release of lead by a silty loam and a fine sandy loam: Kinetics, Sep. Sci. Technol., 30(18), 3471-3488 (1995). Celik, M.S., A method for isolating precipitation from adsorption in surfactant solid systems, Adsorpt. Sci. Technol., 12(1), 19-26 (1995). Chang, Z., et al., A study on the adsorption of gold(III) and macroporous crosslinked polyacrylate (MET) resins: Adsorption equilibrium, Sep. Sci. Technol., 30(17), 3299-3312 (1995). Chang, Z., et al., A study on the adsorption of gold(III) with macroporous crosslinked polyacrylate MET resins: Particle diffusion process, Sep. Sci. Technol., 30(18), 3509-3522 (1995). Chang, Z., et al., Study on the adsorption of gold(III) with macroporous crosslinked polyacrylate MET resins: Liquid diffusion process, Sep. Sci. Technol., 30(19), 3681-3696 (1995). Chatzopoulos, D., and Varma, A., Aqueous-phase adsorption and desorption of toluene in activated carbon fixed beds: Experiments and model, Chem. Eng. Sci., 50(1), 127-142 (1995). Choudhary, V.R.; Mayadevi, S., and Singh, A.P., Simple apparatus for the gravimetric adsorption of liquid vapors on solid catalysts/adsorbents, Ind. Eng. Chem. Res., 34(1), 413-415 (1995). Costa, E.T.H.; Winkler-Hechenleitner, A.A., and Gomez-Pineda, E., Removal of cupric ions from aqueous solutions by contact with corncobs, Sep. Sci. Technol., 30(12), 2593-2602 (1995). Halhouli, K.A.; Darwish, N.A., and A1-Dhoon, N.M., Effects of pH and inorganic salts on the adsorption of phenol from aqueous systems on activated decolorizing charcoal, Sep. Sci. Technol., 30(17), 3313-3324 (1995). Hasany, S.M., and Chaudhary, M.H., Removal of cobalt from aqueous solutions using Haro River sand, Adsorpt. Sci. Yechnol., 12(4), 307-316 (1995). Hassan, N.M., et al., Adsorption decontamination of radioactive waste solvent by activated alumina and bauxites, Sep. Sci. Yechnol., 30(11), 2403-2420 (1995).

1009

Juang, R.S., and Chang, H.L., Column sorption of citric acid from aqueous solutions using tri-n-octylamineimpregnated macroporous resins, Sep. Sci. Technol., 30(6), 917-932 (1995). Kahraman, F., Solvent-in-pulp extraction of boron from slurries, Solvent Extr. Ion Exch., 13(5), 901-922 (1995). Karadag, E.; Saraydin, D., and Guven, O., Behaviors of acrylamide/itaconic acid hydrogels in uptake of uranyl ions from aqueous solutions, Sep. Sci. Technol., 30(20), 3747-3760 (1995). Khalfaoui, B.; Meniai, A.H., and Borja, R., Removal of copper from industrial wastewater by raw charcoal obtained from reeds, J. Chem. Technol. B iotechnol., 64(2), 153-156 (1995). Kim, B.T., et al., Adsorption of radionuclides from aqueous solutions by inorganic adsorbents, Sep. Sci. Yechnol., 30(16), 3165-3182 (1995). Kim, K.R.; Lee, K.J., and Bae, J.H., Characteristics of cobalt adsorption on prepared titanium dioxide and Fe-TiO adsorbents in high temperature water, Sep. Sci. Technol., 30(6), 963-980 (1995). Kumar, S., et al., Sorption of platinum, palladium, iridium, and gold complexes on polyaniline, Solvent Extr. Ion Exch., 13(6), 1097-1122 (1995). Kwok, W.; Hayes, R.E., and Nasr-EI-Din, H.A., Modelling dynamic adsorption of an anionic surfactant on Berea sandstone with radial flow, Chem. Eng. Sci., 50(5), 769-784 (1995). Leitao, A., and Rodrigues, A., The simulation of solid-liquid adsorption in activated carbon columns using estimates of intraparticle kinetic parameters obtained from continuous stirred tank reactor experiments, Chem. Eng. J., 58(3), 239-244 (1995). Leyva-Ramos, R., et al., Adsorption of trivalent chromium from aqueous solutions onto activated carbon, J. Chem. Technol. Biotechnol., 62(1), 64-67 (1995). Lin, S.H., and Hsu, F.M., Liquid-phase adsorption of organic compounds by granular activated carbon and activated carbon fibers, Ind. Eng. Chem. Res., 34(6), 2110-2116 (1995). Mandjiny, S.; Zouboulis, A.I., and Matis, K.A., Removal of cadmium from dilute solutions by hydroxyapatite: Sorption studies, Sep. Sci. Technol., 30(15), 2963-2978 (1995). Morawe, B.; Vogelpohl, A., and Ramteke, D.S., Activated carbon column performance studies of biologically treated landfill leachate, Chem. Eng. Process., 34(3), 299-304 (1995). Nag, A., Utilization of charred sawdust as an adsorbent of dyes, toxic salts and oil from water, Process Safety Environ. Prot., 73(B4), 299-304 (1995). Namasivayam, C., and Ranganathan, K., Removal of lead(II) by adsorption onto 'waste' iron(III)/chromium(III) hydroxide from aqueous solution and radiator manufacturing industry wastewater, Ind. Eng. Chem. Res., 34(3), 869-873 (1995). Namasivayam, C., and Senthilkumar, S., Recycling of industrial solid wastes: 'Waste' Fe(III)/Cr(llI) hydroxide as an adsorbent for the removal of toxic ions and dyes from wastewater, Adsorpt. Sci. Technol., 12(4), 293-296 (1995). Payne, G.F., and Ramakrishnan, S., Coupling ion pair extraction with adsorption for the separation of acidic solutes from water, Ind. Eng. Chem. Res., 34(2), 575-584 (1995). Periasamy, K., and Namasivayam, C., Adsorption of Pb(II) by peanut hull carbon from aqueous solution, Sep. Sci. Technol., 30(10), 2223-2238 (1995). Qadeer, R., and Hanif, J., The isosteric heat of adsorption of Sr2+, Ce 3+, Sm 3+, Gd 3+, Th 4+ and UO22+ ions on activated charcoal, Adsorpt. Sci. Technol., 11(4), 201-208 (1995). Rauf, M.A.; Hasany, S.M., and Ellahi, I., Selective adsorption studies of ytterbium on sand from aqueous solution, Adsorpt. Sci. Technol., 12(4), 317-322 (1995). Rauf, M.A.; Hussain, M.T., and Hasany, S.M., Europium removal at trace concentrations by manganese dioxide from slightly acidic mixtures of water and tetrahydrofuran, Sep. Sci. Technol., 30(1), 117-124 (1995). Sag, Y., and Kutsal, T., Copper(II) and nickel(II) adsorption by Rhizopus arrhizus in batch stirred reactors in series, Chem. Eng. J., 58(3), 265-274 (1995). Saraydin, D.; Karadag, E., and Guven, O., Adsorptions of some heavy metal ions in aqueous solutions by acrylamide/maleic acid hydrogels, Sep. Sci. Yechnol., 30(17), 3287-3298 (1995). Schweiger, T.A.J., Effects of water residues on solvent adsorption cycles, Ind. Eng. Chem. Res., 34(1), 283-287 (1995). Sheu, E.Y.; Storm, D.A., and Shields, M.B., Adsorption kinetics of asphaltenes at toluene/acid solution interface, Fuel, 74(10), 1475-1479 (1995). Taha, F., et al., Effect of cationic and anionic surfactants on the electrokinetic potentials of cassiterite and quartz in the presence ofpolyvalent cations, Adsorpt. Sci. Technol., 12(1), 7-18 (1995).

1010

Teirlinck, P.A.M., and Petersen, F.W., Factors influencing the adsorption of gold-iodide onto activated carbon, Sep. Sci. Technol., 30(16), 3129-3142 (1995). Tinge, J.T., and Drinkenburg, A.A.H., The enhancement of the physical absorption of gases in aqueous activated carbon slurries, Chem. Eng. Sci., 50(6), 937-942 (1995). van Deventer, J.S.J., and van der Merwe, P.F., Kinetic model for the decomposition of cyanide during the elution of gold from activated carbon, Sep. Sci. Technol., 30(6), 883-898 (1995). Wang, Z., Experimental errors due to liquid-phase adsorption in a batch adsorber, Chem. Eng. Sci., 50(15), 2491-2494 (1995). Yang, J.H.K.; Burban, J.H., and Cussler, E.L., Copper selective adsorption with a microemulsion-based resin, AIChE J., 41(5), 1165-1170 (1995). Yeh, R.Y.L., and Thomas, A., Color removal from dye wastewaters by adsorption using powdered activated carbon: Mass transfer studies, J. Chem. Technol. Biotechnol., 63(1), 48-54 (1995). Yeh, R.Y.L., and Thomas, A., Colour difference measurement and color removal from dye wastewaters using different adsorbents, J. Chem. Technol. Biotechnol., 63(1), 55-59 (1995). Zarraa, M.A., A study on the removal of chromium(VI) from waste solutions by adsorption on to sawdust in stirred vessels, Adsorpt. Sci. Technol., 12(2), 129-138 (1995). 1996 Abuziad, N.S., and Nakhla, G.F., Design and economic aspects of activated carbon adsorption in relation to the effect of dissolved oxygen, Environ. Prog., 15(2), 128-134 (1996). Aksenenko, E.V., and Tarasevich, Y.I., Quantum chemical study of the interaction of water molecules with a partially oxidized graphite surface, Adsorpt. Sci. Yechnol., 14(6), 383-392 (1996). AI Mansi, N.M., Decolorizing wastewater in a fixed bed using natural adsorbents, Sep. Sci. Technol., 31(14), 1989-1996 (1996). Alemany, L.J., et al., Removal of phenol from aqueous solution by adsorption on to coal fly ash, Adsorpt. Sci. Yechnol., 13(6), 527-536 (1996). Balkose, D., et al., Flexible poly(vinyl chloride)/zeolite composites for dye adsorption from aqueous solutions, Sep. Sci. Technol., 31(9), 1279-1290 (1996). Belfer, S., and Binman, S., Gold recovery from cyanide solutions with a new fibrous polymer adsorbent, Adsorption, 2(3), 237-244 (1996). Bhummasobhana, A., et al., Surfactant-enhanced carbon regeneration in liquid-phase application, Sep. Sci. Yechnol., 31(5), 629-642 (1996). Chen, J.; Yiacoumi, S., and Blaydes, T.G., Equilibrium and kinetic studies of copper adsorption by activated carbon, Sep. Yechnol., 6(2), 133-146 (1996). Chibowski, S., Investigation of the mechanism of polymer adsorption on a metal oxide/water solution interface, Adsorpt. Sci. Technol., 14(3), 179-188 (1996). Choi, S.J., and Choi, Y.H., Removal of direct red from aqueous solution by foam separation techniques of ion and adsorbing colloid flotation, Sep. Sci. Technol., 31 (15), 2105-2116 (1996). Claessens, R., and Baron, G.V., Measurement of liquid phase multicomponent adsorption in a synzyme partial oxidation catalyst, Chem. Eng. Sci., 51 (10), 1869-1878 (1996). Da Costa, A.C.A., and De Franca, F.P., Cadmium uptake by biosorbent seaweeds: Adsorption isotherms and some process conditions, Sep. Sci. Technol., 31 (17), 2373-2394 (1996). Darwish, N.A.; Halhouli, K.A., and AI-Dhoon, N.M., Adsorption of phenol from aqueous systems onto spent oil shale, Sep. Sci. Yechnol., 31(5), 705-714 (1996). Dasmahapatra, G.P., et al., Studies on separation characteristics of hexavalent chromium from aqueous solution by fly ash, Sep. Sci. Technol., 31(14), 2001-2009 (1996). Dunne, J.A.; Myers, A.L., and Kofke, D.A., Simulation of adsorption of liquid mixtures of nitrogen and oxygen in a model faujasite cavity at 77.5 K, Adsorption, 2(1), 41-50 (1996). E1-Geundi, M.S., Adsorption kinetics of cationic dyestuffs on to natural clay, Adsorpt. Sci. Technol., 13(4), 295303 (1996). EI-Nabarawy, T.; Fagal, G.A., and Khalil, L.B., Removal of ammonia from aqueous solution using activated carbons, Adsorpt. Sci. Technol., 13(1), 7-14 (1996). Farhadpour, F.A., and Bono, A., Sorptive separation of ethanol-water mixtures with a bi-dispersed hydrophobic molecular sieve, silicalite: Determination of the controlling mass transfer mechanism, Chem. Eng. Process., 35(2), 141-156 (1996).

1011

Farhadpour, F.A., and Bono, A., Sorptive separation of ethanol-water mixtures with a bi-dispersed hydrophobic molecular sieve, silicalite: Measurement and theoretical analysis of column dynamics, Chem. Eng. Process., 35(2), 157-168 (1996). Friese, T.; Ulbig, P., and Schulz, S., Development of a universal group contribution model for single-component and multicomponent adsorption of liquids on various adsorbents (UGCMA), Ind. Eng. Chem. Res., 35(6), 2032-2038(1996). Ganguly, S.K., and Goswami, A.N., Surface diffusion kinetics in the adsorption of acetic acid on activated carbon, Sep. Sci. Technol., 31(9), 1267-1278 (1996). Gardea-Torresdey, J.L., et al., Uptake of copper ions from solution by different populations of Medicago sativa (Alfalfa), Solvent Extr. Ion Exch., 14(1), 119-140 (1996). Gierak, A., Application of activated carbon for the sorption of some heavy metals from aqueous solution and their determination by atomic spectroscopy, Adsorpt. Sci. Technol., 14(1), 47-58 (1996). Gorbachevska, K.R., et al., Peculiarities associated with the adsorption of alkyne peroxides from dimethyl formamide/water solutions on to the mercury electrode in the presence of tetraethylammonium cations, Adsorpt. Sci. Technol., 14(4), 229-238 (1996). Goto, M., et al., Impulse response analysis for adsorption of ethyl acetate on activated carbon in supercritical carbon dioxide, Sep. Sci. Technol., 31(12), 1649-1662 (1996). Grzegorczyk, D.S., and Carta, G., Adsorption of amino acids on porous polymeric adsorbents: Equilibrium, Chem. Eng. Sci., 51 (5), 807-818 (1996). Grzegorczyk, D.S., and Carta, G., Adsorption of amino acids on porous polymeric adsorbents Intraparticle mass transfer, Chem. Eng. Sci., 51 (5), 819-826 (1996). Gu, B., et al., Efficient separation and recovery of technetium-99 from contaminated groundwater, Sep. Technol., 6(2), 123-132 (1996). Gupta, G.S., and Shukla, S.P., An inexpensive adsorption technique for the treatment of carpet effluents by low cost materials, Adsorpt. Sci. Technol., 13(1), 15-26 (1996). Guy, D.W.; Crawford, R.J., and Mainwaring, D.E., The wetting behaviour of several organic liquids in water on coal surfaces, Fuel, 72(2), 238-242 (1996). Hankins, N.P.; O'Haver, J.H., and Harwell, J.H., Modeling effects of pH and counterions on surfactant adsorption at the oxide/water interface, Ind. Eng. Chem. Res., 35(9), 2844-2855 (1996). Huggahalli, M., and Fair, J.R., Prediction of equilibrium adsorption of water onto activated carbon, Ind. Eng. Chem. Res., 35(6), 2071-2074 (1996). Janusz, W., The structure of the electrical double layer at the LiChrospher-type adsorbent/aqueous electrolyte solution interface, Adsorpt. Sci. Technol., 14(3), 151-162 (1996). Juang, R.S., and Swei, S.L., Effect of dye nature on its adsorption from aqueous solutions onto activated carbon, Sep. Sci. Technol., 31(15), 2143-2158 (1996). Juang, R.S., et al., Liquid-phase adsorption of phenol and its derivatives on activated carbon fibers, Sep. Sci. Technol., 31(14), 1915-1932 (1996). Kmetty, K., et al., Production of high-purity CH-products by liquid adsorption on molecular sieve materials, Comput. Chem. Eng., 20 (Suppl. B), S 1521-S 1526 (1996). Lee, K.C., and Ku, Y., Removal of chlorophenols from aqueous solution by anion-exchange resins, Sep. Sci. Technol., 13(18), 2557-2578 (1996). Liang, L.; Gu, B., and Yin, X., Removal of technetium-99 from contaminated groundwater with sorbents and reductive materials, Sep. Technol., 6(2), 111-122 (1996). Lin, M.C., and Liu, J.C., Adsorbing colloid flotation of As(V): Feasibility of utilizing streaming current detector, Sep. Sci. Technol., 31 (11 ) 1629-1642 (1996). Lin, M.C., and Liu, J.C., Adsorbing colloid flotation of Pb(II): Feasibility of utilizing streaming current detector, Sep. Sci. Technol., 31(17), 2335-2350 (1996). Magdy, Y.H., The adsorption of mixed dyes (acidic and basic) on to hardwood in a fixed bed, Adsorpt. Sci. Technol., 13(5), 367-376 (1996). Marton, G., et al., Purification of nuclear power plant decontamination solutions by preparative scale reactive adsorption, Chem. Eng. Sci., 51(11), 2655-2660 (1996). Marzal, P., et al., Cadmium and zinc adsorption onto activated carbon: Influence of temperature, pH and metal/carbon ratio, J. Chem. Technol. Biotechnol., 66(3), 279-285 (1996). Mathur, J.N., et al., Recovery of neptunium from highly radioactive waste solutions of Purex origin using tributyl phosphate, Sep. Sci. Technol., 31(15), 2045-2064 (1996).

1012

Matsuyama, H., et al., Selective separation of rare earth metals by solvent extraction in the presence of new hydrophilic chelating polymers functionalized with ethylenediamine-tetraacetic acid: Adsorption properties, Sep. Sci. Technol., 31(5), 687-704 (1996). McKay, G.; El Geundi, M., and Nassar, M.M., Pore diffusion during the adsorption of dyes onto bagasse pith, Process Safety Environ. Prot., 74(B4), 277-288 (1996). Membrez, J.; Infelta, P.P., and Renken, A., Use of the Laplace transform technique for simple kinetic parameters evaluation: Application to the adsorption of a protein on porous beads, Chem. Eng. Sci., 51(19), 4489-4498 (1996). Meszaros, R.; Nagy, M., and Veress, G., Study of adsorption from dilute aqueous solutions on to activated carbon: Novel representation of experimental data, Adsorpt. Sci. Technol., 13(5), 327-340 (1996). Meszaros, R.; Nagy, M., and Veress, G., Study of adsorption from dilute aqueous solutions on to activated carbon: Some exotic adsorption isotherms, Adsorpt. Sci. Technol., 13(5), 341-354 (1996). Mittal, A.K., and Venkobachar, C., Uptake of cationic dyes by sulfonated coal: Sorption mechanism, Ind. Eng. Chem. Res., 35(4), 1472-1474 (1996). Mousa, H., and Qasaimeh, M., Experimental investigation of the de-inking of recycled newspaper using plastic particles, Sep. Sci. Technol., 31(8), 1093-1104 (1996). Murad, S., Molecular dynamics simulations of osmosis and reverse osmosis in solutions, Adsorption, 2(1), 95110(1996). Mustafa, S., et al., Exchange of some divalent metal cations and their effect on the dissociation of iron(Ill) phosphate, Adsorpt. Sci. Technol., 13(4), 241-260 (1996). Nassar, M.M.; Hamoda, M.F., and Radwan, G.H., Utilization of palm-fruit bunch particles for the adsorption of dyestuff wastes, Adsorpt. Sci. Technol., 13(1), 1-6 (1996). Orumwense, F.F.O., Removal of lead from water by adsorption on a kaolinitic clay, J. Chem. Technol. Biotechnol., 65(4), 363-369 (1996). Qadeer, R.; Hanif, I., and Hanif, J., Effect of different cations on the adsorption of Dy3+, Gd 3+, Eu 3. and Sm> ions on activated charcoal from aqueous solutions, Adsorpt. Sci. Technol., 13(1), 27-30 (1996). Rao, M.B., and Sircar, S., Adsorption from binary immiscible liquid mixture, AIChE J., 42(4), 1191-1194 (1996). Rautiu, R., and White, D.A., The sorption of cationic species on hydrous tin dioxide, Solvent Extr. Ion Exch., 14(4), 721-738 (1996). Saraydin, D.; Karadag, E., and Guven, O., Adsorption of some basic dyes by acrylamide-maleic acid hydrogels, Sep. Sci. Technol., 31(3), 423-434 (1996). Saraydin, D.; Karadag, E., and Guven, O., Behaviors of acrylamide/maleic acid hydrogels in uptake of some cationic dyes from aqueous solutions, Sep. Sci. Technol., 31 (17), 2359-2372 (1996). Seki, H., and Suzuki, A., Adsorption of lead ions on composite biopolymer adsorbent, Ind. Eng. Chem. Res., 35(4), 1378-1382 (1996). Siddiqui, F.A., and Franses, E.I., Surface tension and adsorption synergism for solutions of binary surfactants, Ind. Eng. Chem. Res., 35(9), 3223-3232 (1996). Suzuki, Y., et al., Biological activated carbon treatment of effluent water from wastewater treatment processes of plating industries, Sep. Technol., 6(2), 147-154 (1996). Talu, O., and Meunier, F., Adsorption of associating molecules and micropores and application to water on carbon, AlChE J., 42(3), 809-819 (1996). Vanjara, A.K., and Dixit, S.G., Mixed micelle formation and co-adsorption of anionic/non-ionic surfactant mixtures on ruffle, Adsorpt. Sci. Technol., 13(5), 377-396 (1996). Vanjara, A.K., and Dixit, S.G., Adsorption of cationic surfactants on ruffle, Adsorpt. Sci. Technol., 13(5), 397408 (1996). Vasylechko, V.O., et al., Adsorption of copper on transcarpathian mordenite, Adsorpt. Sci. Technol., 14(5), 267278 (1996). Wasay, S.A.; Tokunaga, S., and Park, S.W., Removal of hazardous anions from aqueous solutions by La(III)and Y(III)-impregnated alumina, Sep. Sci. Technol., 31 (10) 1501-1514 (1996). 1997 Aksu, Z.; Acikel, U., and Kutsal, T., Application of multicomponent adsorption isotherms to simultaneous biosorption of iron(III) and chromium(VI) on C. vulgaris, J. Chem. Technol. Biotechnol., 70(4), 368-378 (1997).

1013

AI-Haj Ali, A., and EI-Bishtawi, R., Removal of lead and nickel ions using zeolite tuff, J. Chem. Technol. Biotechnol., 69(1 ), 27-34 (1997). Allen, S.J., et al., The adsorption of pollutants by peat, lignite and activated chars, J. Chem. Technol. Biotechnol., 68(4), 442-452 (1997). Anjoh, N.; Yamazaki, T., and Ozawa, S., UV-Vis spectroscopic study on effects of pressure for adsorption of pnitrotoluene at liquid-solid interface, Adsorption, 3(2), 173-186 (1997). Anon., Roll-press briquetters for recycling waste sludge help steelmakers end landfilling and cut scrap costs, Powder Handling Process., 9(1), 64-67 (1997). Atia, A.A., and Radwan, N.R.E., Adsorption of different surfactants on kaolinite, Adsorpt. Sci. Technol., 15(8), 619-626 (1997). Bossrez, S.; Remacle, J., and Coyette, J., Adsorption of nickel on Enterococcus hirae cell walls, J. Chem. Technol. Biotechnol., 70(1), 45-50 (1997). Bostick, D.T.; Arnold, W.D., and Guo, B., The evaluation of sodium-modified chabazite zeolite and resorcinolformaldehyde resin for the treatment of contaminated process wastewater, Sep. Sci. Technol., 32(1), 793-811 (1997). Bouchard, C.R.; Jolicoeur, J.; Kouadio, P., and Britten, M., Study of humic acid adsorption on nanofiltration membranes by contact angle measurements, Can. J. Chem. Eng., 75(2), 339-345 (1997). Bulewicz, E.M.; Kozak, A., and Kowalski, Z., Treatment of chromic tannery wastes using coal ashes from fluidized bed combustion of coal, Ind. Eng. Chem. Res., 36(10), 4381-4384 (1997). Chen, J.P., and Yiacoumi, S., Biosorption of metal ions from aqueous solutions, Sep. Sci. Technol., 32(1), 51-69 (1997). Choudhary, V.R., and Choudhary, T.V., Entrance of straight and branched chain compounds from their bulk liquid phase into H-ZSM-5 zeolite, Chem. Eng. Sci., 52(20), 3543-3552 (1997). Choudhary, V.R.; Nayak, V.S., and Choudhary, T.V., Single-component sorption/diffusion of cyclic compounds from their bulk liquid phase in H-ZSM-5 zeolite, Ind. Eng. Chem. Res., 36(5), 1812-1818 (1997). Chu, W.; Yang, X., and Wu, Y., Adsorption of dodecatungstosilicic acid onto activated carbons from aqueous and acidic media, Adsorpt. Sci. Technol., 15(1 ), 1-14 (1997). Condoret, J.S., et al., Prediction of water adsorption curves for heterogeneous biocatalysis in organic and supercritical solvents, Chem. Eng. Sci., 52(2), 213-220 (1997). Dahal, M.P.; Lawrance, G.A., and Maeder, M., Variation in the adsorption of lead(II) by a range of electrolytic manganese dioxides: Chemometric examination of correlation with physical properties, Adsorpt. Sci. Technol., 15(8), 583-592 (1997). Dahl, I.M.; Myhrvold, E.; Slagtern, A., and Stocker, M., Adsorption of lower alcohols from water solutions on high silica zeolites, mesoporous MCM-41 and AIPO4-5, Adsorpt. Sci. Technol., 15(4), 289-300 (1997). Deorkar, N.V., and Tavlarides, L.L., Zinc, cadmium, and lead separation from aqueous streams using solid-phase extractants, Ind. Eng. Chem. Res., 36(2), 399-406 (1997). Desai,T.R., and Dixit, S.G., Adsorption from mixtures of cationic/non-ionic surfactants onto polystyrene surface, Adsorpt. Sci. Yechnol., 15(5), 391-405 (1997). Dias, N.L., and Gushikem, Y., 2-Mercaptoimidazole covalently bonded to a silica gel surface for the selective separation of mercury(II) from an aqueous solution, Sep. Sci. Technol., 32(15), 2535-2545 (1997). E1-Nabarawy, T.; Mostafa, M.R., and Youssef, A.M., Activated carbons tailored to remove different pollutants from gas streams and from solution, Adsorpt. Sci. Technol., 15(1), 59-68 (1997). Feng, M.; Mei, J., and Hu, S., Selective removal of iron from grape juice using an iron(III) chelating resin, Sep. Purif. Technol., 11(2), 127-136 (1997). Gonzalez-Pradas, E.; Villafranca-Sanchez, M., and Campo, A.G., Removal of l,l'-dimethyl-4,4'bipyridyl dichloride from aqueous solution by natural and activated bentonite, J. Chem. Technol. Biotechnol., 69(2), 173-178 (1997). Grzegorczyk, D.S., and Carta, G., Frequency response of liquid-phase adsorption on polymeric adsorbents, Chem. Eng. Sci., 52(10), 1589-1608 (1997). Gupta, V.K.; Rastogi, A., and Dwivedi, M.K., Process development for the removal of zinc and cadmium from wastewater using slag: A blast furnace waste material, Sep. Sci. Technol., 32(17), 2883-2912 (1997). Gupta, V.K.; Srivastava, S.K., and Mohan, D., Equilibrium uptake, sorption dynamics, process optimization, and column operations for the removal and recovery of malachite green from wastewater using activated carbon and activated slag, Ind. Eng. Chem. Res., 36(6), 2207-2218; 36(12), 5545 (1997).

1014

Halhouli, K.A.; Darwish, N.A., and AI-Jahmany, Y.Y., Effects of temperature and inorganic salts on the adsorption of phenol from multicomponent systems onto a decolorizing carbon, Sep. Sci. Technol., 32(18), 3027-3036 (1997). Hasany, S.M., and Ikram, M., Uptake of traces of selenite by manganese dioxide from aqueous solutions, Sep. Sci. Yechnol., 32(12), 1945-1957 (1997). Hashim, M.A.; Chu, K.H.; Phang, S.M., and Ong, G.S., Adsorption equilibria of cadmium on algal biomass, Adsorpt. Sci. Yechnol., 15(6), 445-454 (1997). Hsu, Y.C.; Chiang, C.C., and Yu, M.F., Adsorption behavior of basic dyes on activated clay, Sep. Sci. Technol., 32(15), 2513-2534 (1997). Huang, J.G., and Liu, J.C., Enhanced removal of As(V) from water with iron-coated spent catalyst, Sep. Sci. Yechnol., 32(9), 1557-1570 (1997). Jasra, R.V.; Choudary, N.V., and Bhat, S.G.T., Liquid-phase sorption of higher alkanes and alkenes in zeolite NaZSM-5 at 10, 30 and 50~ Sep. Sci. Technol., 32(9), 1571-1588 (1997). Juang, R.S.; Tseng, R.L., and Wu, F.C., Adsorption behavior of reactive dyes from aqueous solutions on chitosan, J. Chem. Technol. Biotechnol., 70(4), 391-399 (1997). Khokhlova, T.D.; Nikitin, Y.S. and Detistova, A.L., Modification of silicas and their investigation by dye adsorption, Adsorpt. Sci. Technol., 15(5), 333-340 (1997). Knapp, J.S.; Zhang, F.M., and Tapley, K.N., Decolourisation of orange-II by a wood-rotting fungus, J. Chem. Yechnol. Biotechnol., 69(3), 289-296 (1997). Leahy, J.J., and Hughes, M.A., The rheology of peat/solvent slurries, J. Chem. Technol. Biotechnol., 70(2), 193197 (1997). Le Cloirec, P.; Brasquet, C., and Subrenat, E., Adsorption onto fibrous activated carbon: Applications to water treatment, Energy Fuels, 11(2), 331-336 (1997). Lee, C.K.; Low, K.S., and Chung, L.C., Removal of some organic dyes by hexane-extracted spent bleaching earth, J. Chem. Technol. Biotechnol., 69(1), 93-99 (1997). Lee, H.W.; Kim, K.J., and Fane, A.G., Removal of phenol by adsorption on powdered activated carbon in a continuous flow stirred cell membrane system, Sep. Sci. Technol., 32(11), 1835-1849 (1997). Lee, J.H.; Song, D.I., and Jeon, Y.W, Adsorption of organic phenols onto dual organic cation Montmorillonite from water, Sep. Sci. Yechnol., 32(12), 1975-1992 (1997). Liebenberg, S.P., and VanDeventer, J.S.J., Evaluating a dynamic model for the competitive elution of gold and base metals from activated carbon, Sep. Sci. Technol., 32(11), 1787-1804 (1997). Lin, S.H., and Chert, Y.W., Liquid-phase adsorption of 1,1-dichloro-l-fluoroethane by various adsorbents, Ind. Eng. Chem. Res., 36(10), 4347-4352 (1997). Matatov-Meytal, Y.I., and Sheintuch, M., Abatement of pollutants by adsorption and oxidative catalytic regeneration, Ind. Eng. Chem. Res., 36(10), 4374-4380 (1997). Mayer, A.F.; Hartmann, R., and Deckwer, W.D., Diffusivities of clavulanic acid in porous sorption systems with ion pairing, Chem. Eng. Sci., 52(24), 4561-4568 (1997). McKay, G., and Porter, J.F., Equilibrium parameters for the sorption of copper, cadmium and zinc ions onto peat, J. Chem. Technol. Biotechnol., 69(3), 309-320 (1997). McKay, G., and Porter, J.F., A comparison of Langmuir-based models for predicting multicomponent metal ion equilibrium sorption isotherms on peat, Process Safety Environ. Prot., 75(B3), 171-180 (1997). McKay, G.; E1-Geundi, M., and Nassar, M.M., Equilibrium studies for the adsorption of dyes on bagasse pith, Adsorpt. Sci. Yechnol., 15(4), 251-270 (1997). Miller, C.J.; Olson, A.L., and Johnson, C.K., Cesium absorption from acidic solutions using ammonium molybdophosphate on a polyacrylonitrile support (AMP-PAN), Sep. Sci. Technol., 32(1), 37-50 (1997). Miyabe, K., and Takeuchi, S., Model for surface diffusion in liquid-phase adsorption, AIChE J., 43(11), 29973006 (1997). Mostafa, M.R., Adsorption of mercury, lead and cadmium ions on modified activated carbons, Adsorpt. Sci. Technol., 15(8), 551-558 (1997). Myasoedova, G.V.; Shcherbinina, N.I., and Zakhartchenko, E.A., Sorption of platinum group metals and gold chlorocomplexes by amine polymeric sorbents, Solvent Extr. Ion Exch., 15(6), 1107-1118 (1997). Nassar, M.M., The kinetics of basic dye removal using palm-fruit bunch, Adsorpt. Sci. Technol., 15(8), 609-618 (1997). Nassar, M.M., and Magdy, Y.H., Removal of different basic dyes from aqueous solutions by adsorption on palmfruit bunch particles, Chem. Eng. J., 66(3), 223-226 (1997).

1015

Oludipe, J.O., Studies on the sorption of metal ions from aqueous media: Effect of hydrogen peroxide on the sorption kinetics, Adsorpt. Sci. Technol., 15(5), 361-372 (1997). Otu, E.O., Elution of gold from activated carbon using supercritical carbon dioxide, Sep. Sci. Technol., 32(6), 1107-1114 (1997). Ouki, S.K., and Neufeld, R.D., Use of activated carbon for the recovery of chromium from industrial wastewaters, J. Chem. Technol. Biotechnol., 70(1), 3-8 (1997). Paajanen, A.; Lehto, J., and Santapakka, T., Sorption of cobalt on activated carbons from aqueous solutions, Sep. Sci. Technol., 32(1 ), 813-826 (1997). Petersen, F.W., and VanDeventer, J.S.J., Competitive adsorption of gold cyanide and organic compounds onto porous adsorbents, Sep. Sci. Technol., 32(13), 2087-2103 (1997). Qadeer, R., and Saleem, M., Adsorption of UO22+ ions on activated charcoal: pH effect, Adsorpt. Sci. Technol., 15(5), 373-376 (1997). Safarik, I.; Nymburska, K., and Safarikova, M., Adsorption of water-soluble organic dyes on magnetic charcoal, J. Chem. Technol. Biotechnol., 69(1 ), 1-4 (1997). Safarik, I., and Safarikova, M., Copper phthalocyanine dye immobilized on magnetite particles: An efficient adsorbent for rapid removal of polycyclic aromatic compounds from water solutions and suspensions, Sep. Sci. Technol., 32(14), 2385-2392 (1997). Sakoda, A.; Nomura, T., and Suzuki, M., Activated carbon membrane for water treatments: Application to decolorization of coke furnace wastewater, Adsorption, 3(1 ), 93-105 (1997). Seco, A., et al., Absorption of heavy metals from aqueous solutions onto activated carbon in single Cu and Ni systems and in binary Cu-Ni, Cu-Cd and Cu-Zn systems, J. Chem. Technol. Biotechnol., 68(1), 23-30 (1997). Shethna, H.K., and Towler, G.P., Design of mixed-solvent processes for chemisorption with ultrahigh recovery, Ind. Eng. Chem. Res., 36(12), 5307-5320 (1997). Song, K.C.; Lee, H.K., and Moon, H., Simultaneous removal of the radiotoxic nuclides CS 137 and 1129 from aqueous solution, Sep. Purif. Technol., 12(3), 215-228 (1997). Sun, G., and Xu, X., Sunflower stalks as adsorbents for color removal from textile wastewater, Ind. Eng. Chem. Res., 36(3), 808-812 (1997). Taiwei, C.; Jinzhou, D., and Zuyi, T., Static and kinetic studies of the adsorption/desorption of H+ and OH- ions on alumina in aqueous sodium nitrate solution, Adsorpt. Sci. Technol., 15(5), 349-360 (1997). Taiwei, C.; Jinzhou, D., and Zuyi, T., Adsorption of H+, OH- and background electrolyte ions on alumina: Point of zero charge, triple layer model (TLM) parameters and thermodynamic parameters, Adsorpt. Sci. Technol., 15(6), 455-464 (1997). Tan, H.K.S., New algorithms for the computation of column dynamics of multicomponent liquid-phase adsorption, Adsorption, 3(2), 137-150 (1997). Tsezos, M.; Georgousis, Z., and Remoudaki, E., Ionic competition effects in a continuous uranium biosorptive recovery process, J. Chem. Technol. Biotechnol., 70(2), 198-206 (1997). Various, Production and use of carbon-based materials for environmental cleanup (symposium papers), Energy Fuels, 11(2), 249-353 (1997). Wang, R.C.; Kuo, C.C., and Shyu, C.C., Adsorption of phenols onto granular activated carbon in a liquid-solid fluidized bed, J. Chem. Technol. Biotechnol., 68(2), 187-194 (1997). Wang, Z., and Govind, R., Biofiltration of isopentane in peat and compost packed beds, AIChE J., 43(5), 13481356 (1997). White, D.A., and Asfarsiddique, A., Removal of manganese and iron from drinking water using hydrous manganese dioxide, Solvent Extr. Ion Exch., 15(6), 1133-1145 (1997). White, D.A., and Bussey, R.L., Water sorption properties of modified clinoptilolite, Sep. Purif. Technol., 11(2), 137-141 (1997). Williams, C.J., and Edyvean, R.G.J., Optimization of metal adsorption by seaweeds and seaweed derivatives, Process Safety Environ. Prot., 75(B 1), 19-26 (1997). Ziolkowska, D., and Garbacz, J.K., Adaptation of single gas adsorption equations for the description of adsorption from non-aqueous liquid solutions of iodine onto active carbons, Adsorpt. Sci. Technol., 15(3), 155-164 (1997). Zouboulis, A.I.; Matis, K.A., and Lanara, B.G., Removal of cadmium from dilute solutions by hydroxyapatite: Flotation studies, Sep. Sci. Technol., 32(10), 1755-1767 (1997).

1016

ION EXCHANGE,

CHROMATOGRAPHY,

AND RELATED

SEPARATIONS

1967 Copeland, J.P.; Henderson, C.L., and Marchello, J.M., Influence of resin selectivity on film diffusion-controlled ion exchange, AIChE J., 13(3), 449-452 (1967). Gilwood, M.E., Saving capital and chemicals with countercurrent ion exchange, Chem. Eng. (N.Y.), 18 December, 83-88 (1967). Klein, G.; Tondeur, D., and Vermeulen, T., Multicomponent ion exchange in fixed beds, Ind. Eng. Chem. Fund., 6(3), 339-364 (1967). Lai, C.L., and Roth, J.A., Dynamic simulation of gas chromatographic column, Chem. Eng. Sci., 22(10), 12991304 (1967). Solt, G.S., Continuous countercurrent ion exchange, Brit. Chem. Eng., 12(10), 1582-1586 (1967). Tallmadge, J.A., Ion exchange treatment of mixed electroplating wastes, Ind. Eng. Chem. Process Des. Dev., 6(4), 419-423 (1967). 1968 George, D.R.; Riley, J.M., and Ross, J.R., Potassium recovery by chemical precipitation and ion exchange, Chem. Eng. Prog., 64(5), 96-99 (1968). Hall, G.R.; Streat, M., and Creed, G.R.B., Ion exchange in nuclear chemical processes, Trans. IChemE, 46, T53T59 (1968). Lifshutz, N., and Dranoff, J.S., Inversion of concentrated sucrose solutions in fixed beds of ion exchange resin, Ind. Eng. Chem. Process Des. Dev., 7(2), 266-269 (1968). Michalson, A.W., High quality water via ion exchange, Chem. Eng. Prog., 64(10), 67-73 (1968). Ryan, J.M.; Timmins, R.S., and O'Donnell, J.F., Production-scale gas chromatography, Chem. Eng. Prog., 64(8), 53-59(1968). Schneider, P., and Smith, J.M., Adsorption rate constants from chromatography, AIChE J., 14(5), 762-771 (1968). Tuichiev, I.S.; Rizaev, N.U.; Merenkov, K.V., and Yusipov, M.M., Hydrodynamic properties of ion exchange resins during fluidization, Int. Chem. Eng., 8(2), 221-223 (1968). Turner, J.C.R., and Snowdon, C.B., Liquid-side mass-transfer coefficients in ion exchange: Nernst-Planck model, Chem. Eng. Sci., 23(3), 221-230; 23(9), 1099-1104 (1968). Turner, J.C.R.; Snowdon, C.B.; Jones, D.C., and Ward, J.W.C., Estimation of ion-exchange equilibrium diagrams involving weakly dissociated electrolytes, Trans. IChemE, 46, T232-T235 (1968). 1969 Cooke, J.P., Understanding a gas chromatograph, Chem. Eng. (N.Y.), 10 March, 134-144 (1969). Copeland, J.P., and Marchello, J.M., Film-diffusion controlled ion-exchange with a selective resin, Chem. Eng. Sci., 24(9), 1471-1474 (1969). Eberly, P.E., Diffusion studies in zeolites and related solids by gas chromatography, Ind. Eng. Chem. Fund., 8(1), 25-30(1969). Kadlec, V., and Matejka, Z., Mixed-bed deionisation by weak electrolyte ion-exchange resins regenerated in-situ by carbon dioxide, J. Appl. Chem., 19, 352-355 (1969). Kuong, J.F., Maximising ion-exchanger throughput, Chem. Eng. (N.Y.), 15 December, 160 (1969). Pollio, F.X.; Kunin, R., and Petralia, J.W., Treat sour water by ion exchange, Hydrocarbon Process., 48(5), 124126 (1969). Timmins, R.S.; Mir, L., and Ryan, J.M., Large-scale chromatography: New separation tool, Chem. Eng. (N.Y.), 19 May, 170-178 (1969). 1970 Campbell, D.O., and Buxton, S.R., Rapid ion exchange separations, Ind. Eng. Chem. Process Des. Dev., 9(1), 89-99 (1970). McGovern, T.J., and Dranoff, J.S., Sucrose inversion by partially deactivated ion-exchange resin beds, AIChE J., 16(4), 536-538 (1970). Streat, M., and Brignal, W.J., Representation of ternary ion exchange equilibria, Trans. IChemE, 48, T151-T155 (1970). Turner, J.C.R., and Snowdon, C.B., Liquid-side mass transfer coefficients in ion exchange, Chem. Eng. Sci., 25(11), 1673-1678 (1970).

1017

Weber, O.W.; Miller, I.F., and Gregor, H.P., Absorption of carbon dioxide by weak-base ion exchange resins, AIChE J., 16(4), 609-614 (1970). 1971 Colwell, C.J., and Dranoff, J.S., Nonlinear equilibrium anJ axial mixing effects in intraparticle diffusioncontrolled sorption by ion-exchange resin beds, Ind. Eng. Chem. Fund., 10(1), 65-70 (1971). Danes, F., Batch process application to ion-exchange unit operation, Chem. Eng. Sci., 26(8), 1277-1288 (1971). Gardiner, W.C., and Munoz, F., Mercury removal from waste effluent via ion exchange, Chem. Eng. (N.Y.), 23 August, 57-59 (1971 ). Gondo, S.; Itai, M., and Kusunoki, K., Computational and experimental studies on a moving ion-exchange bed, Ind. Eng. Chem. Fund., 10(1), 140-146 (1971). Heines, V., A history of chromatography, Chemtech, May, 280-285 (1971). Hsu, H.W., Optimum adsorbent volume in liquid adsorption chromatography, Sep. Sci., 6(5), 645-652 (1971). Kunin, R., and Downing, D.G., Ion-exchange system boasts more pulling power, Chem. Eng. (N.Y.), 28 June, 67-69 (1971). Maldacker, T.A., and Rogers, L.B., Effect of loading on separation efficiency using steric exclusion chromatography, Sep. Sci., 6(6), 747-758 (1971). Mir, L., Comparison of static bed and moving bed chromatography, Sep. Sci., 6(4), 515-536 (1971). Moreland, A.K., and Rogers, L.B., Effects of slow mass transfer using microporous adsorbents in gas-solid chromatography, Sep. Sci., 6(1 ), 1-24 (1971 ). Smuts, T.W.; Jordaan, J.T., and Pretorius, V., Phenomenological plate height equation for packed chromatographic columns, Sep. Sci., 6(5), 653-684 (1971). Various, Gel permeation chromatography (topic issue), Sep. Sci., 6(1), 47-164; 6(2), 207-330 (1971). 1972 Cloete, C.E., and de Clerk, K., Distillation vs. chromatography: Comparison based on purity index, Sep. Sci., 7(4), 449-456 (1972). Conrard, P.; Caude, M., and Rosset, R., Separation of close species on ion exchangers, Sep. Sci., 7(5), 465-490 (1972). Dodds, J.A., and Tondeur, D., Design of cyclic fixed-bed ion-exchange operations, Chem. Eng. Sci., 27(6), 1267-1282; 27(12), 2291-2298 (1972). Golden, L.S., and Irving, J., Osmotic and mechanical strength of ion-exchange resins, Chem. Ind. (London), 4 November, 837-844 (1972). Holliday, D.C., Continuous ion exchange: Design and development, Chem. Ind. (London), 16 September, 717723 (1972). Parker, K.J., Ion exchange in the sugar industry, Chem. Ind. (London), 21 October, 782-790 (1972). Qureshi, M.; Qureshi, S.Z.; Gupta, J.P., and Rathore, H.S., Progress in ion-exchange studies on insoluble salts of polybasic metals, Sep. Sci., 7(6), 615-630 (1972). 1973 Buys, T.S., and de Clerk, K., Effect of temperature on production rate in chromatography, Sep. Sci., 8(5), 551566(1973). Conrard, P.; Caude, M., and Rosset, R., Separation of close species on ion exchangers, Sep. Sci., 8(1), 1-10; 8(2), 269-278 (1973). Johnson, J.F.; Macphail, M.G.; Cooper, A.R., and Bruzzone, A.R., Effect of column length on chromatographic fractionation of polymers, Sep. Sci., 8(5), 577-584 (1973). Lal, B.B., and Douglas, W.J.M., Techniques for measuring sorption of water by ion-exchange resin spheres, lnd. Eng. Chem. Fund., 12(3), 381-384 (1973). Letan, R., Continuous ion-exchanger, Chem. Eng. Sci., 28(3), 981-985 (1973). Martin, J.R., and Johnson, J.F., Cost-efficiency comparisons of some polymer chromatographic fractionation techniques, Sep. Sci., 8(5), 619-622 (1973). Meares, P., Characteristics and uses of ion exchange membranes, Chem. Ind. (London), 1 December, 103-107 (1973). Metzger, V.G.; Barford, R.A., and Rothbart, H.L., Chromatography and countercurrent distribution, Sep. Sci., 8(2), 143-160 (1973). Millar, J.R., Fundamentals of ion exchange, Chem. Ind. (London), 5 May, 409-413 (1973).

1018

Ouano, A.C., and Barker, J.A., Computer simulation of linear gel permeation chromatography, Sep. Sci., 8(6), 673-700(1973). Stevens, B., Chromatographic refining unit, Process Eng. (London), March, 82-84 (1973). Weiss, G.H., and Dishon, M., Resolution in nonuniform chromatographic systems, Sep. Sci., 8(3), 337-344 (1973). Williams, R.C., Ion exchange resins in power stations, Chem. Ind. (London), 19 May, 465-470 (1973). 1974 Braud, C., and Selegny, E., Interrelation of swelling and selectivity of ion-exchange resins, Sep. Sci., 9(1), 13-26 (1974). Bull, P.S.; Evans, J.V., and Nicholson, F.D., Condensate polishing performance of powdered ion-exchange resins, J. Appl. Chem. Biotechnol., 24, 475-486 (1974). Caude, M.; Conrard, P., and Rosset, R., Displacement development on ion exchangers, Sep. Sci., 9(4), 269-286 (1974). Dodds, J.A., and Tondeur, D., Design of cyclic fixed-bed ion-exchange operations, Chem. Eng. Sci., 29(2), 611620(1974). Kirchner, J.G., Thin-layer chromatography, Chemtech, February, 79-82 (1974). Lal, B.B., and Douglas, W.J.M., Equilibrium water sorption and volumetric behavior of ion-exchange resin spheres, Ind. Eng. Chem. Fund., 13(3), 223-227 (1974). Nikelly, J.G., Porous-layer open-tubular gas chromatography columns, Sep. Purif. Methods, 3(2), 423-441 (1974). Scott, C.D., High-pressure ion-exchange chromatography applied to separation of complex biochemical mixtures, Sep. Purif. Methods, 3(2), 263-298 (1974). Singhal, R.P., Separation and analysis of nucleic acids and their constituents by ion-exclusion and ion-exchange column chromatography, Sep. Purif. Methods, 3(2), 339-398 (1974). Various, Liquid-liquid extraction and ion exchange in analytical chemistry (topic issue), Chem. Ind. (London), 17 August, 639-647 (1974). Various, Chromatographic separations (topic issue), Sep. Purif. Methods, 3(1), 1-86, 133-244 (1974). Whitlock, L.R., and Siggia, S., Fusion reaction gas chromatography, Sep. Purif. Methods, 3(2), 299-338 (1974). 1975 Bolto, B.A., Sirotherm ion-exchange desalination, Chemtech, May, 303-307 (1975). Braud, C., and Selegny, E., Interrelation of swelling and selectivity of ion-exchange resins, Sep. Sci., 10(1), 47110; 10(2), 175-244; 10(3), 331-358 (1975). Farkas, E.J., and Himsley, A., Fundamental aspects of behavior of ion exchange equipment, Can. J. Chem. Eng., 53,575-585 (1975). Kataoka, T.; Nishiki, T., and Ueyama, K., Mass transfer with liquid anion exchange, Chem. Eng. J., 10(3), 189196(1975). Pawlowski, L., and Zytomirski, S., Influence of ion exchange capacity and total concentration of solution of ions of different valency on their chromatographic separation, Sep. Sci., 10(1 ), 33-38 (1975). Prengle, H.W., et al., Recycle wastewater by ion exchange, Hydrocarbon Process., 54(4), 173-184 (1975). Rendell, M., Future of large-scale chromatography, Process Eng. (London), April, 66-70 (1975). Vermeer, D.J.; Lynn, S., and Vermeulen, T., Cation-exchange column behavior in desalination process with regenerant recovery, Ind. Eng. Chem. Process Des. Dev., 14(3), 290-297 (1975). 1976 Cooper, A.R., and Lynn, T.R., Coiled high-efficiency liquid chromatography columns, Sep. Sci., 11(1), 39-44 (1976). de Rosset, A.J.; Neuzil, R.W., and Korous, D.J., Liquid column chromatography as a predictive tool for continuous countercurrent adsorptive separations, Ind. Eng. Chem. Process Des. Dev., 15(2), 261-266 (1976). Ito, Y., and Bowman, R.L., Foam countercurrent chromatography, Sep. Sci., 11(3), 201-206 (1976). Kadlec, V., and Hubner, P., Ion exchange deionisation with recirculation of regenerant by heat, Chem. Ind. (London), 4 September, 744-746 (1976). Kataoka, T.; Nishiki, T., and Ueyama, K., Simultaneous mass transfer of acid and ions in a liquid anion exchanger, Chem. Eng. J., 12(3), 233-238 (1976). Mostafa, H.A., and Said, A.S., Theoretical-plate concept for fixed-bed adsorption and ion-exchange, Trans. IChemE, 54, T132-T134 (1976).

1019

Roland, L.D., Ion exchange: Operational advantages of continuous plants, Processing (Sutton, Engl.), January, 11-12 (1976). Slater, M.J., and Lucas, B.H., Flow patterns and mass transfer rates in fluidized-bed ion-exchange equipment, Can. J. Chem. Eng., 54, 264-270 (1976). Smirnov, N.N., Mathematical models of ion-exchange process, Int. Chem. Eng., 16(2), 234-240 (1976). Sussman, M.V., Continuous chromatography (review paper), Chemtech, April, 260-264 (1976). Talmon, Y., and Rubin, E., Chromatographic separation by foam, Sep. Sci., 11(6), 509-533 (1976). Weatherley, L.R., and Turner, J.C.R., Ion-exchange kinetics: Comparison between a macroporous and a gel resin, Trans. IChemE, 54, T89-T94 (1976). 1977 Holl, W., and Sontheimer, H., Ion exchange kinetics of the protonation of weak-acid ion-exchange resins, Chem. Eng. Sci., 32(7), 755-762 (1977). Pauls, R.E., and Rogers, L.B., Comparisons of methods for calculating retention and separation of chromatographic peaks, Sep. Sci., 12(4), 395-415 (1977). Pauls, R.E., et al., Experimental variables in recycle gas chromatography, Sep. Sci., 12(3), 289-306, (1977). Pusch, W., Ion-exchange membranes, Int. Chem. Eng., 17(1), 62-75 (1977). Shah, D.B., and Ruthven, D.M., Measurement of zeolitic diffusivities and equilibrium isotherms by chromatography, AIChE J., 23(6), 804-809 (1977). Umbreit, G.R., Chromatographic anomalies, Chemtech, February, 101-106 (1977). Various, Novel ion exchangers (topic issue), Chem. Ind. (London), 6 August, 634-652 (1977). 1978 Barker, P.E.; Ellison, F.J., and Hatt, B.W., Countercurrent chromatographic unit for continuous fractionation of dextran, Ind. Eng. Chem. Process Des. Dev., 17(3), 302-309 (1978). Chihara, K.; Suzuki, M., and Kawazoe, K., Adsorption rate on molecular sieving carbon by chromatography, AIChE J., 24(2), 237-246 (1978). Danesi, P.R., and Chiarizia, R., Mass transfer rate with liquid ion exchangers, J. Appl. Chem. Biotechnol., 28, 581-598(1978). De, A.K., and Sen, A.K., Synthetic inorganic ion-exchangers, Sep. Sci. Technol., 13(6), 517-540 (1978). Hubner, P., and Kadlec, V., Kinetic behavior of weak-base anion exchangers, AIChE J., 24(1), 149-154 (1978). Marra, R.A., and Cooney, D.O., Multicomponent sorption operations: Bed shrinking and swelling in an ionexclusion case, Chem. Eng. Sci., 33(12), 1597-1602 (1978). Smith, R.P., and Woodburn, E.T., Prediction of multicomponent ion exchange equilibria for ternary systems from binary systems data, AIChE J., 24(4), 577-587 (1978). Wiley, J.R., Decontamination of alkaline radioactive waste by ion exchange, Ind. Eng. Chem. Process Des. Dev., 17(1), 67-71 (1978). 1979 Abe, M., and Kasai, K., Synthetic inorganic ion-exchange materials, Sep. Sci. Technol., 14(10), 895-908 (1979). Agarwal, J.C., and Klumpar, I.V., Role of liquid ion exchange in processing of complex solutions, J. Chem. Technol. Biotechnol., 29, 730-740 (1979). Erickson, K.L., and Rase, H.F., Fixed-bed ion exchange with differing ionic mobilities and nonlinear equilibria, Ind. Eng. Chem. Fund., 18(4), 312-317 (1979). Goto, S.; Sato, N., and Teshima, H., Periodic operation for desalting water with thermally regenerable ionexchange resin, Sep. Sci. Technol., 14(3), 209-218 (1979). Gupta, A.R., Isotope effects in ion-exchange equilibria in aqueous and mixed solvent systems, Sep. Sci. Technol., 14(9), 843-858 (1979). Hadzismajlovic, D.E., et al., Mass transfer in liquid spout-fluid beds of ion exchange resin, Chem. Eng. J., 17(3), 227-236 (1979). Knaebel, K.S.; Cobb, D.D.; Shih, T.T., and Pigford, R.L., Ion-exchange rates in bifunctional resins, Ind. Eng. Chem. Fund., 18(2), 175-180 (1979). Various, Ion exchange in the water industry (topic issue), Chem. Ind. (London), 3 March, 142-165 (1979).

1980 Brown, J.M., and Wilson, D.J., Macroreticular resin columns, Sep. Sci. Technol., 15(8), 1533-1555 (1980). Burfield, D.R., and Smithers, R.H., Desiccant efficiency in solvent drying: Applications of cationic exchange resins, J. Chem. Technol. Biotechnol., 30, 491-496 (1980).

1020

Calmon, C., Explosion hazards of using nitric acid in ion-exchange equipment, Chem. Eng. (N.Y.), 17 November, 271-274 (1980). Curtis, M.A., et al., Liquid chromatographic fractionations of mixtures of polystyrene oligomers, Sep. Sci. Technol., 15(7), 1413-1428 (1980). Kennedy, D.C., Predict sorption of metals on ion-exchange resins, Chem. Eng. (N.Y.), 16 June, 106-118 (1980). MacLean, G.T., Effect of synthetic flocculant on ion-exchange resin, Sep. Sci. Technol., 15(8), 1555-1563 (1980). Omatete, O.O.; Clazie, R.N., and Vermeulen, T., Column dynamics of ternary ion exchange, Chem. Eng. J., 19(3), 229-250 (1980). Ruthven, D.M., and Kumar, R., An experimental study of single-component and binary adsorption equilibria by a chromatographic method, Ind. Eng. Chem. Fund., 19(1), 27-32 (1980). Soldatov, V.S., and Bichkova, V.A., Ternary ion-exchange equilibria, Sep. Sci. Technol., 15(2), 89-1 l0 (1980). Takahashi, T., and Gill, W.N., Hydrodynamic chromatography, Chem. Eng. Commun., 5(5), 367-380 (1980). Various, Advances in ion-exchange water treatment (topic issue), Chem. Ind. (London), 20 September, 712-743 (1980). Various, Chromatographic processes (symposium papers), Sep. Sci. Technol., 15(3), 587-696; 15(4), 697-798 (1980). 1981 Annino, R.., Chromatographs can run on air, Chemtech, August, 482-487 (1981). Barker, P.E., and Chuah, C.H., A sequential chromatographic process for the separation of glucose/fructose mixtures, Chem. Eng. (Rugby, Engl.), August, 389-393 (1981). Dyer, A.; Enamy, H., and Townsend, R.P., Plotting and interpretation of ion-exchange isotherms in zeolite systems, Sep. Sci. Technol., 16(2), 173-184 (1981). Gomez-Vaillard, R.; Kershenbaum, L.S., and Streat, M., Performance of continuous, cyclic ion-exchange reactors, Chem. Eng. Sci., 36(2), 307-326 (1981). Goto, S.; Goto, M., and Teshima, H., Simplified evaluations of mass-transfer resistances from batch-wise adsorption and ion-exchange data, Ind. Eng. Chem. Fund., 20(4), 368-375 (1981). Huang, J.C.; Forsythe, R., and Madey, R., Gas-solid chromatography of methane-helium mixtures: Transmission of step increase in concentration of methane through activated carbon adsorber bed at 25~ Sep. Sci. Technol., 16(5), 475-486 (1981). Matsuda, H.; Yamamoto, T.; Goto, S., and Teshima, H., Periodic operation for desalination with thermally regenerable ion-exchange resins (dynamic studies), Sep. Sci. Technol., 16(1), 31-42 (1981 ). Moharir, A.S.; Saraf, D.N., and Kunzru, D., Effect of crystal size distribution on chromatographic peaks in molecular sieve columns, Chem. Eng. Commun., 11(6), 377-386 (1981). Phillips, J.B.; Wright, N.A., and Burke, M.F., Probabilistic approach to digital simulation of chromatographic processes, Sep. Sci. Technol., 16(8), 861-884 (1981). Rahman, K., and Streat, M., Mass transfer in liquid fluidized beds of ion exchange particles, Chem. Eng. Sci., 36(2), 293-306 (1981). Raman, M.S., Polymer resins for water treatment, Chemtech, April, 252-255 (1981). Rice, R.G., and Foo, S.C., Continuous desalination using cyclic mass-transfer ion exchange with bifunctional resins, Ind. Eng. Chem. Fund., 20(2), 150-155 (1981). Said, A.S., Theory of nonlinear chromatography, Sep. Sci. Technol., 16(2), 113-134 (1981). Various, Pharmaceutical applications of ion exchange and solvent extraction (topic issue), Chem. Ind. (London), 3 October, 677-690 ( 1981). Various, Advances in chromatography (topic issue), Chem. Ind. (London), 17 October, 710-732 (1981). 1982 Clifford, D., Multicomponent ion-exchange calculations for selected ion separations, Ind. Eng. Chem. Fund., 21(2), 141-153 (1982). Graham, E.E., and Dranoff, J.S., Application of Stefan-Maxwell equations to diffusion in ion exchangers, Ind. Eng. Chem. Fund., 21(4), 360-369 (1982). Graham, E.E., and Fook, C.F., Rate of protein absorption and desorption on cellulosic ion exchangers, AIChE J., 28(2), 245-250 (1982). Huang, J.C., et al., Gas-solid chromatography of methane-helium mixtures: Moment analysis of breakthrough curves, Sep. Sci. Technol., 17(12), 1417-1424 (1982).

1021

Husain, S.W., et al., Synthesis and ion-exchange properties of lanthanum tungstate, Sep. Sci. Technol., 17(7), 935-944 (1982). Koff, F.W.; Sifniades, S., and Tunick, A.A., Ion-exchange process for recovery of hydroxylamine from Raschig synthesis mixtures, Ind. Eng. Chem. Process Des. Dev., 21 (2), 204-216 (1982). Kojima, T., et al., Fundamental study on recovery of copper with a cation-exchange membrane, Can. J. Chem. Eng., 60, 642-658 (1982). Novosad, J., and Myers, A.L., Thermodynamics of ion exchange as an adsorption process, Can. J. Chem. Eng., 60, 500-503 (1982). Pelosi, P., and McCarthy, J., Preventing fouling of ion-exchange resins, Chem. Eng. (N.Y.), 9 August, 75-78; 6 September, 125-128 (1982). Rao, M.G., and Gupta, A.K., Ion exchange processes accompanied by ionic reactions, Chem. Eng. J., 24(2), 181190(1982). Rousar, I., and Ditl, P., Kinetic characteristics of batch adsorber or ion exchange device operated under nonisothermal conditions, Chem. Eng. Commun., 18(5), 341-354 (1982). Schenk, H.J., et al., Development of sorbers for recovery of uranium from seawater, Sep. Sci. Technol., 17(11), 1293-1308 (1982). Slater, M.J., The relative sizes of fixed bed and continuous countercurrent flow ion exchange equipment, Trans. IChemE, 60, T54-T58 (1982). Various, Ion exchange in the petrochemical industry (topic issue), Chem. Ind. (London), 21 August, 561-573 (1982). 1983 Abe, M., and Hayashi, K., Synthetic inorganic ion-exchange materials, Solvent Extr. Ion Exch., 1(1), 97-112 (1983). Altshuller, D., Design equations and transient behaviour of the countercurrent moving-bed chromatographic reactor, Chem. Eng. Commun., 19(4), 363-376 (1983). Barba, D.; del Re, G., and Foscolo, P.U., Numerical simulation of multicomponent ion-exchange operations, Chem. Eng. J., 26(1), 33-40 (1983). Barker, P.E.; England, K., and Vlachogiannis, G., Mathematical model for the fractionation of dextran on a semicontinuous countercurrent simulated moving bed chromatograph, Chem. Eng. Res. Des., 61, 241-247 (1983). Begovich, J.M.; Byers, C.H., and Sisson, W.G., A high-capacity pressurized continuous chromatograph, Sep. Sci. Technol., 18(12), 1167-1192 (1983). Bobman, M.H.; Golden, T.C., and Jenkins, R.G., Ion exchange in selected low-rank coals: Equilibrium and kinetics, Solvent Extr. Ion Exch., 1(4), 791-826 (1983). Choppin, G.R., and Ohene-Aniapam, F., Equilibrium sorption of Am(IIl), Ce(III), and Eu(III), on Biorex-70 ionexchange resin, Solvent Extr. Ion Exch., 1(3), 585-596 (1983). Fujine, S.; Saito, K.; Shiba, K., and Itoi, T., Liquid mixing in a large-sized column for ion exchange, Solvent Extr. Ion Exch., 1(1), 113-126 (1983). Goto, M.; Hayashi, N., and Goto, S., Separation of electrolyte and nonelectrolyte by ion retardation resin, Sep. Sci. Technol., 18(5), 475-484 (1983). Shih, C.K., et al., Large-scale liquid chromatography separation system, Chem. Eng. Prog., 79(10), 53-57 (1983). Turner, J.C.R., and Murphy, T.K., A CSTR method for determining ion-exchange equilibria, Chem. Eng. Sci., 38(1), 147-154 (1983). Various, Uses of ion exchange in the food industry (topic issue), Chem. Ind. (London), 7 November, 804-824 (1983). 1984 Bailly, M., and Tondeur, D., Reversibility and performances in productive chromatography, Chem. Eng. Process., 18(6), 293-302 (1984). Bonmati R., et al., Industrial gas chromatography process applied to essential oils, Sep. Sci. Technol., 19(2), 113-156(1984). Costa, E.; Lucas, A., and Gonzalez, M.E., Ion exchange: Determination of interdiffusion coefficients, Ind. Eng. Chem. Fund., 23(4), 400-405 (1984). Jenkins, I.L., Ion exchange in the atomic energy industry with particular reference to actinide and fission product separation (review paper), Solvent Extr. Ion Exch., 2(1), 1-28 (1984). Jepson, B.E., and Shockey, G.C., Calcium hydroxide isotope effect in calcium isotope enrichment by ion exchange, Sep. Sci. Technol., 19(2), 173-182 (1984).

1022

Klein, G., Calculation of ideal or empirically modified mass-action equilibria in heterovalent multicomponent ion exchange, Comput. Chem. Eng., 8(3), 171-178 (1984). Miller, G.H., and Wankat, P.C., Moving port chromatography: A method of improving preparative chromatography, Chem. Eng. Commun., 31 ( 1), 21-44 (1984). Scott, F., Larger high-pressure liquid-chromatography systems, Process Eng. (London), February, 26-31 (1984). Tsuji, M., and Abe, M., Synthetic inorganic ion-exchange materials, Solvent Extr. Ion Exch., 2(2), 253-274 (1984). van der Meer, A.P.; Woerde, H.M., and Wesselingh, J.A., Mass transfer in countercurrent ion-exchange plate column, Ind. Eng. Chem. Process Des. Dev., 23(4), 660-664 (1984). Walton, H.F., Counter-ion effects in partition chromatography, Sep. Sci. Technol., 19(11), 849-856 (1984). 1985 Egawa, H.; Nonaka, T., and Maeda, H., Studies of selective adsorption resins, Sep. Sci. Technol., 20(9), 653-664 (1985). Hyun, S.H., and Danner, R.Po, Gas adsorption isotherms by use of perturbation chromatography, Ind. Eng. Chem. Fund., 24(1), 95-101 (1985). Hyun, S.H., and Danner, R.P., Adsorption equilibrium constants and intraparticle diffusivities in molecular sieves by tracer-pulse chromatography, AIChE J., 31 (7), 1077-1085 (1985). Kamiyanagi, K., and Furusaki, S., Analysis of chromatography by transfer functions, Int. Chem. Eng., 25(2), 301308 (1985). Law, H.H.; Wilson, W.L., and Gabriel, N.E., Separation of gold cyanide ion from anion-exchange resins, Ind. Eng. Chem. Process Des. Dev., 24(2), 236-238 (1985). Mathur, J.N., and Khopkar, P.K., Ion exchange behaviour of chelating resin Dowex A-1 with actinides and lanthanides, Solvent Extr. Ion Exch., 3(5), 753-762 (1985). Riveros, P.A., and Cooper, W.C., Extraction of silver from cyanide solutions with ion-exchange resins, Solvent Extr. Ion Exch., 3(3), 357-376 (1985). Sommer, C.C., et al., Recycle gas chromatography using coarse packings, Sep. Sci. Technol., 20(7), 523-540 (1985). Wildhagen, G.R.S.; Qassim, R.Y.; Rajagopal, K., and Rahman, K., Effective liquid-phase diffusivity in ion exchange, Ind. Eng. Chem. Fund., 24(4), 423-432 (1985). Yoshida, H.; Kataoka, T., and Ikeda, S., Intraparticle mass transfer in bidispersed porous ion exchanger, Can. J. Chem. Eng., 63(3), 422-435 (1985). 1986 Ecknig, W., and Polster, H.J., Supercritical chromatography of paraffins on a molecular sieve: Analytical and preparative scale, Sep. Sci. Technol., 21(2), 139-156 (1986). Frey, D.D., Prediction of liquid-phase mass-transfer coefficients in multicomponent ion exchange: Comparison of matrix, film-model, and effective-diffusivity methods, Chem. Eng. Commun., 47, 273-294 (1986). Geldart, R.W.; Yu, Q.; Wankat, P.C., and Wang, N., Improving elution and displacement ion-exchange chromatography by adjusting eluent and displacer affinities, Sep. Sci. Technol., 21(9), 873-886 (1986). Golden, L., Industrial use of ion exchange resins, Chem. Eng. (Rugby, Engl.), October, 31-34 (1986). Jackson, M.B., and Pilkington, N.H., Effect of the degree of crosslinking on the selectivity of ion-exchange resins, J. Chem. Yechnol. Biotechnol., 36(2), 88-94 (1986). Jun, S.H., and Ruckenstein, E., Separation of multicomponent mixture of proteins by potential barrier chromatography, Sep. Sci. Technol., 21(2), 111-138 (1986). Lefevre, L.J., Ion exchange: Problems and troubleshooting, Chem. Eng. (N.Y.), 7 July, 73-75 (1986). Strelow, F.W.E., Influence of resin loading on cation exchange distribution coefficients of some elements in hydrochloric acid, Solvent Extr. Ion Exch., 4(6), 1193-1208 (1986). Wilson, D.J., Modeling of ion-exchange column operation, Sep. Sci. Yechnol., 21 (8), 767-788; 21 (10), 991-1008 (1986). Yoshida, H., and Kataoka, T., Recovery of amine and ammonia by ion exchange method: Comparison of ligand sorption and ion exchange accompanied by neutralization reaction, Solvent Extr. Ion Exch., 4(6), 1171-1192 (1986). Yoshida, H.; Kataoka, Y., and Fujikawa, S., Kinetics in a chelate ion exchanger, Chem. Eng. Sci., 41(10), 25172530 (1986). Yu, Q., and Wang, N.H.L., Multicomponent interference phenomena in ion exchange columns, Sep. Purif. Methods, 15(2), 127-158 (1986).

1023

1987 Arve, B.H., and Liapis, A.I., Modeling and analysis of affinity chromatography in finite bath, AIChE J., 33(2), 179-193(1987). Higgins, I.R., and Denton, M.S., CSA continuous countercurrent ion exchange technology, Sep. Sci. Technol., 22(2), 997-1016 (1987). Huang, T.C.; Huang, Y.C., and Tsai, F.N., Intraparticle diffusion-controlled kinetics of phenol adsorption on ion exchange resins, Chem. Eng. Commun., 56, 77-86 (1987). Kataoka, T., and Yoshida, H., Dynamics in a thermally regenerable ion exchange column, Chem. Eng. J., 36(1), 41-50 (1987). Kataoka, T., et al., Liquid-side ion-exchange mass transfer in ternary system, AIChE J., 33(2), 202-210 (1987). Mikhail, E.M., and Misak, N.Z., Ion exchange characteristics of ceric tungstate: Kinetics of exchange, J. Chem. Technol. Biotechnol., 39(4), 219-230 (1987). Misak, N.Z., and Mikhail, E.M., Ion-exchange characteristics of a new manganese oxide, Solvent Extr. Ion Exch., 5(5), 939-976 (1987). Tavlarides, L.L.; Bae, J.H., and Lee, C.K., Solvent extraction, membranes and ion exchange in hydrometallurgical dilute metals separation, Sep. Sci. Technol., 22(2), 581-618 (1987). Various, Ultrapure water by ion exchange (topic issue), Chem. Ind. (London), 16 February, 104-118 (1987). Various, Preparative-scale chromatography (topic issue), Sep. Sci. Technol., 22(8), 1791-2110 (1987). Way, J.D., et al., Facilitated transport of carbon dioxide in ion exchange membranes, AIChE J., 33(3), 480-487 (1987). Yan, T.Y., and Shu, P., Regeneration of ion-exchange resin in nonaqueous media, Ind. Eng. Chem. Res., 26(4), 753-755(1987). 1988 Barker, P.E., and Ganetsos, G., Chemical and biochemical separations using preparative and large-scale batch and continuous chromatography, Sep. Purif. Methods, 17(1), 1-66 (1988). 13iscans, I3.; Riba, J.P., and Couderc, J.P., Continuous equipment for ion exchange in fluidized bed: Prospects and problems, Int. Chem. Eng., 28(2), 248-257 (1988). 13olden, W.13., and Groves, F.R., Batch sorption by ligand exchange: Determination of intraparticle diffusivity, Chem. Eng. Commun., 64, 125-136 (1988). Forsythe, R., et al., Gas-solid chromatography: Longitudinal and intraparticle diffusion of acetylene in activated carbon, Sep. Sci. Technol., 23(14), 2319-2328 (1988). Geckler, K.E.; Shkinev, V.M., and Spivakov, 13.Y., Liquid-phase polymer-based retention: A new method for selective ion separation, Sep. Purif. Methods, 17(2), 105-140 (1988). Haas, C.N., Existence of ternary interactions in ion exchange, AIChE J., 34(4), 702-703 (1988). Howard, A.J.; Carta, G., and 13yers, C.H., Separation of sugars by continuous annular chromatography, Ind. Eng. Chem. Res., 27(10), 1873-1882 (1988). Huang, T.C., and Cho, L.T., Adsorption of phenol on anion exchange resins in presence of p-nitrophenol, Chem. Eng. Commun., 74, 169-178 (1988). Hwang, Y.L.; Helfferich, F.G., and Leu, R.J., Multicomponent equilibrium theory for ion-exchange columns involving reactions, AIChE J., 34(10), 1615-1626 (1988). Kataoka, T., and Yoshida, H., Kinetics of ion exchange accompanied by neutralization reaction, AIChE J., 34(6), 1020-1026 (1988). Kawasaki, T., Specification of general theory of quasi-static linear gradient chromatography, Sep. Sci. Technol., 23(14), 2365-2378 (1988). Keum, D.K., and Lee, W.K., Simulation of moving feed-port chromatography by rate model with mass transfer effect, Sep. Sci. Technol., 23(14), 2349-2364 (1988). Miyai, Y.; Ooi, K., and Katoh, S., Recovery of lithium from seawater using a new type of ion-sieve adsorbent based on magnesium-manganese oxide, Sep. Sci. Technol., 23(1), 179-192 (1988). Mustafa, S.; Hussain, S.Y., and Ali, H., Ion exchange sorption of phosphate, Solvent Extr. Ion Exch., 6(4), 725738(1988). Riveros, P.A., and Cooper, W.C., Kinetic aspects of ion exchange extraction of gold, silver, and base-metal cyano complexes, Solvent Extr. Ion Exch., 6(3), 479-504 (1988). Sanders, S.J., et al., Modeling the separation of amino acids by ion-exchange chromatography, Chem. Eng. Prog., 84(8), 47-54 (1988). Sengupta, A.K., and Lim, L., Modeling chromate ion-exchange processes, AIChE J., 34(12), 2019-2029 (1988).

1024

Sisson, W.G.; Begovich, J.M.; Byers, C.H., and Scott, C.D., Continuous chromatography, Chemtech, August, 498-502 (1988). Solt, G.S., The basis of deionization plant design, Chem. Eng. (Rugby, Engl.), January, 14-15 (1988). Solt, G.S.; Nowosielski, A.W., and Feron, P., Predicting performance of ion-exchange columns, Chem. Eng. Res. Des., 66(6), 524-530 (1988). Staniewski, J.W.; Latto, B., and Hamielec, A.E., Sorption of water by poly-(sodium acrylate) resin from organic solutions and mixtures, Chem. Eng. Res. Des., 66(4), 371-377 (1988). Taffe, P., Compact water-deionizer unit, Processing (Sutton, Engl.), July, 23-26 (1988). Takeda, K., et al., Equilibrium principle of displacement chromatography, Sep. Sci. Technol., 23(14), 2329-2348 (1988). Thonchk, N.K., et al., Extraction of thiocyanate ions from coal gasification effluents by ion exchange, Chem. Eng. Res. Des., 66(6), 503-517 (1988). Various, Ion exchange and chromatographic separations (symposium papers), Sep. Sci. Technol., 23(12), 18531928 (1988). Wankat, P.C., and Koo, Y.M., Scaling rules for isocratic elution chromatography, AIChE J., 34(6), 1006-1019 (1988). Ward, K.J.; Kaliaguine, S.C.; Tanguy, P.A., and Jean, G., Numerical simulation of a chromatograph column: Linear case, Ind. Eng. Chem. Res., 27(8), 1474-1480 (1988). 1989 Agosto, M.; Wang, N.H.L., and Wankat, P.C., Moving-withdrawal liquid chromatography of amino acids, Ind. Eng. Chem. Res., 28(9), 1358-1364 (1989). Allen, R.M.; Addison, P.A., and Dechapunya, A.H., Characterization of binary and ternary ion exchange equilibria, Chem. Eng. J., 40(3), 151-158 (1989). Ball, M., and Harries, R.R., Resins for high-purity water production, J. Chem. Technol. Biotechnol., 45(2), 97108(1989). Barker, P.E.; Ganetsos, G., and England, K., Dextran fractionation using preparative-scale continuous chromatography, J. Chem. Technol. Biotechnol., 46(3), 209-218 (1989). Bolden, W.B.; White, T., and Groves, F.R., Continuous fixed-bed ligand exchange: The shrinking core model, AIChE J., 35(5), 849-852 (1989). Carta, G., et al., Separation of metals by continuous annular chromatography with step elution, Chem. Eng. Commun., 79, 207-228 (1989). Chitrakar, R., and Abe, M., Synthetic inorganic ion-exchange materials, Solvent Extr. Ion Exch., 7(4), 721-734 (1989). Ding, H.; Yang, M.C.; Schisla, D., and Cussler, E.L., Hollow-fiber liquid chromatography, AIChE J., 35(5), 814820 (1989). Fish, B.B., and Carr, R.W., Experimental study of countercurrent moving-bed chromatographic reactor, Chem. Eng. Sci., 44(9), 1773-1784 (1989). Gosling, I.S.; Cook, D., and Fry, M.D.M., Role of adsorption isotherms in design of chromatographic separations for downstream processing, Chem. Eng. Res. Des., 67(3), 232-242 (1989). Hartford, R.W.; Kojima, M., and O'Connor, C.T., Lanthanum ion exchange on H-ZSM5, Ind. Eng. Chem. Res., 28(12), 1748-1752 (1989). Hsu, T.B., and Pigford, R.L., Salt removal from water by continuous ion exchange using thermal regeneration, Ind. Eng. Chem. Res., 28(9), 1345-1352 (1989). Hudson, M.J., and Matejka, Z., Extraction of copper by selective ion exchangers with pendent ethyleneimine groups: Investigation of active sites, Sep. Sci. Technol., 24(15), 1417-1426 (1989). Jama, M.A., and Yucel, H., Equilibrium studies of sodium-ammonium, potassium-ammonium, and calciumammonium exchanges on clinoptilolite zeolite, Sep. Sci. Technol., 24(15), 1393-1416 (1989). Kawasaki, T., A fundamental structure of the general theory of overload quasi-static linear gradient chromatography, Sep. Sci. Yechnol., 24(14), 1109-1158 (1989). Khan, Z.H., and Hussain, K., Non-destructive analysis of crude oil by gel permeation chromatography, Fuel, 68(9), 1198-1202 (1989). Kocjan, R., and Przeszlakowski, S., Retention of heavy metals and their separation on silica gel modified with calconecarboxylic acid, Sep. Sci. Yechnol., 24(3), 291-302 (1989). Lamey, S.; Hesbach, P., and Childers, E., Separation of mild gasification liquid products using open-column chromatography, Energy Fuels, 3(5), 636-641 (1989).

1025

Larin, A.V., Criterion for the quantitative assessment of ideal conditions in chromatography, Adsorpt. Sci. Technol., 6(4), 212-218 (1989). Lee, W.C.; Huang, S.H., and Tsao, G.T., Design equations of chromatography by perturbation method, Chem. Eng. J., 40(3), 165-174 (1989). Lin, B., and Guiochon, G., Numerical simulation of chromatographic band profiles at large concentrations: Length of space increment and height equivalent to a theoretical plate, Sep. Sci. Technol., 24(1), 31-40 (1989). Lin, B.; Ma, Z., and Guiochon, G., Influence of axial dispersion on the bald profile in nonlinear chromatography using the Lax-Wendroff method, Sep. Sci. Technol., 24(11), 809-830 (1989). Liu, X.; Liu, J.C., and Cheng, J.K., New inorganic ion exchangers containing phosphorus, Sep. Sci. Technol., 24(1), 63-78 (1989). McCoy, B.J., Adsorption chromatography of a heterogeneous mixture, Chem. Eng. Sci., 44(4), 993-996 (1989). Mustafa, S., et al., Temperature effect on ion exchange sorption of phosphate, Solvent Extr. Ion Exch., 7(4), 705720 (1989). Oh, M.; Smith, J.M., and McCoy, B.J., Diffusion and adsorption in arrested-flow chromatography, AIChE J., 35(7), 1224-1226 (1989). Ostman, C.E., and Colmsjo, A.L., Separation of polycyclic aromatic compounds from complex oil samples using bonded phase backflush HPLC and GC-MS techniques, Fuel, 68(10), 1248-1250 (1989). Perrut, M., and Jusforgues, P., A new fractionation process: Preparative chromatography with a supercritical eluent, Int. Chem. Eng., 29(4), 646-653 (1989). Sagara, F., et al., Preparation and adsorption properties of manganese oxide-cellulose hybrid-type ion-exchanger for lithium ion: Application to enrichment of lithium ion from seawater, Sep. Sci. Technol., 24(14), 12271244(1989). Saunders, M.S.; Vierow, J.B., and Carta, G., Uptake of phenylalanine and tyrosine by a strong-acid cation exchanger, AIChE J., 35(1), 53-68 (1989). Schaeffer, S.T.; Zalkow, L.H., and Teja, A.S., Supercritical fluid isolation of monocrotaline from Crotalaria spectabilis using ion-exchange resins, Ind. Eng. Chem. Res., 28(7), 1017-1020 (1989). Sheth, A.C.; Prasad, J., and Butler, W.A., Desulfurization of alkali metal sulfates using anion-exchange resins, AIChE J., 35(3), 519-523 (1989). Strelow, F.W.E., Distribution coefficients and ion exchange behavior of some chloride complex forming elements with bio-rad AG5OW-X8 cation exchange resin in mixed nitric-hydrochloric acid solutions, Solvent Extr. Ion Exch., 7(4), 735-747 (1989). Venkateswarlu, K.S., et al., Use of strong-base organic anion exchangers for removal of suspended alumina particles in light water-heavy water systems, Sep. Sci. Technol., 24(5), 467-474 (1989). Yoshida, H., and Kataoka, T., Recovery of mercuric chloride using chloride-form ion exchanger, AIChE J., 35(2), 318-320 (1989). Yoshida, H., and Ruthven, D.M., Adsorption of gaseous ethylamine on H-form strong-acid ion exchangers, AIChE J., 35(11), 1869-1875 (1989). Yu, Q., and Wang, N.H.L., Computer simulations of dynamics of multicomponent ion exchange and adsorption in fixed beds: Gradient-directed moving finite element method, Comput. Chem. Eng., 13(8), 915-926 (1989). 1990 Allen, R.M., and Addison, P.A., Ion exchange equilibria for ternary systems from binary exchange data, Chem. Eng. J., 44(3), 113-118 (1990). Anon., Solid electrolyte separation offers pure gases, Chem. Eng. (Rugby, Engl.), March, 30 (1990). Bain, P.E., A model predicting equilibrium for plutonium sorption by anion exchange resin, Solvent Extr. Ion Exch., 8(2), 341-352 (1990). Bi, Y., and Wen-Bin, H., Two-barrel bile-acids-sensitive microelectrodes based on liquid ion exchanger, Biotechnol. Prog., 6(1), 62-66 (1990). Bolden, W.B., and Groves, F.R., Amine recovery by ligand exchange: Pore diffusion model, Ind. Eng. Chem. Res., 29(1), 116-121 (1990). Breeman, D.J., Backflush column removes GC sample water, Hydrocarbon Process., 69(3), 90-91 (1990). Byers, C.H., et al., Sugar separations on a pilot scale by continuous annular chromatography, Biotechnol. Prog., 6(1), 13-20 (1990). Carta, G., and Bauer, J.S., Analytic solution for chromatography with nonuniform sorbent particles, AIChE J., 36(1), 147-150 (1990).

1026

Chen, S.H.; Chao, K.J., and Lee, T.Y., Lanthanum-NaY zeolite ion exchange: Thermodynamics and thermochemistry, Ind. Eng. Chem. Res., 29(10), 2020-2024 (1990). Chen, T.L., and Hsu, J.T.A., Prediction of particle size effects on liquid chromatography performance by multilayer model, Chem. Eng. Commun., 98, 55-64 (1990). Ching, C.B.,; Chu, K.H., and Ruthven, D.M., A study of multicomponent adsorption equilibria by liquid chromatography, AIChE J., 36(2), 275-282 (1990). Costantino, U.; Marmottini, F., and Vivani, R., Ion exchange properties of zirconium phosphate phosphite with alkaline earth metal ions, Solvent Extr. Ion Exch., 8(4), 713-728 (1990). Cramer, S.M., and Subramanian, G., Recent advances in the theory and practice of displacement chromatography, Sep. Purif. Methods, 19(1), 31-92 (1990). Czok, M., and Guiochon, G., Comparison of the results obtained with different models for the simulation of preparative chromatography, Comput. Chem. Eng., 14(12), 1435-1444 (1990). De Carlo, E.H., Separation of lanthanide series elements in marine Fe-Mn crusts by ion-exchange chromatography and determination by IC/AES, Sep. Sci. Technol., 25(6), 781-798 (1990). DeCarli, J.P.; Carta, G., and Byers, C.H., Displacement separations by continuous annular chromatography, AIChE J., 36(8), 1220-1228 (1990). Ding, H., and Cussler, E.L., Overloaded hollow-fiber liquid chromatography, Biotechnol. Prog., 6(6), 472-478 (1990). Dye, S.R.; DeCarli, J.P., and Carta, G., Equilibrium sorption of amino acids by a cation-exchange resin, Ind. Eng. Chem. Res., 29(5), 849-857 (1990). Erkey, C., and Akgerman, A., Chromatography theory: Application to supercritical fluid extraction, AIChE J., 36(11), 1715-1721 (1990). Frey, D.D., Numerical simulation of multicomponent chromatography using spreadsheets, Chem. Eng. Educ., 24(4), 204-207 (1990). Gu, T., et al., Displacement effect in multicomponent chromatography, AIChE J., 36(8), 1156-1162 (1990). Gu, T.; Tsai, G-J., and Tsao, G.T., New approach to a general nonlinear multicomponent chromatography model, AIChE J., 36(5), 784-788 (1990). Inoue, K.; Yoshizuka, K., and Baba, Y., Adsorption of metal ions on a novel amine-type chelating resin, Solvent Extr. Ion Exch., 8(2), 309-324 (1990). Johansson, H.J.; Pettersson, T.N., and Berglof, J.H., Development of a chromatographic process for large-scale purification of Staphylococcal Enterotoxin B, J. Chem. Technol. Biotechnoi., 49(3), 233-242 (1990). Kawasaki, T., and Niikura, M., Overload quasi-static linear gradient chromatography: Theory versus hydroxyapatite high-performance liquid chromatography, Sep. Sci. Technol., 25(4), 397-436 (1990). Kelley, F.D., and Chimowitz, E.H., Near-critical phenomena and resolution in supercritical fluid chromatography, AIChE J., 36(8), 1163-1175 (1990). Kim, M.G.; Amos, L.W., and Barnes, E.E., Study of the reaction rates and structures of a phenol-formaldehyde resol resin by ]3C-NMR and gel permeation chromatography, Ind. Eng. Chem. Res., 29(10), 2032-2037 (1990). Komatsu, Y., et al., Adsorption behavior of cobalt(II) ions on layered dihydrogen tetratitanate hydrate fibers in aqueous solutions in the range from 298-523K, Solvent Extr. Ion Exch., 8(1), 173-186 (1990). Kubota, K., and Hayashi, S., An analysis of the elution curve in preparative chromatography with moving feed ports, Can. J. Chem. Eng., 68(3), 420-426 (1990). Kullberg, L.H., and Clearfield, A., Thermodynamics of alkali and alkaline earth metal ion-exchange on zirconium sulphophosphonates, Solvent Extr. Ion Exch., 8(1), 187-198 (1990). Lafferty, C., and Hobday, M., The use of low rank brown coal as an ion exchange material: Basic parameters and the ion exchange mechanism, Fuel, 69(1), 78-83 (1990). Lafferty, C., and Hobday, M., The use of low rank brown coal as an ion exchange material: Ionic selectivity and factors affecting utilization, Fuel, 69(1), 84-87 (1990). Lee, T.Y., et al., Lanthanum-NaY zeolite ion exchange: Kinetics, Ind. Eng. Chem. Res., 29(10), 2024-2028 (1990). Lin, Y.S., and Ma, Y.H., Analysis of liquid chromatography with nonuniform crystallite particles, AIChE J., 36(10), 1569-1576 (1990). Mardan, A.; Alstad, J., and Liljenzin, J.O., Development of a non-corroding radio-chromatographic system and measurement of its parameters, Solvent Extr. Ion Exch., 8(1), 117-136 (1990).

1027

Mardan, A., and Liljenzin, J.O., Sulfonation and performance of surface-sulfonated (5 and 10 mm diameter) highly crosslinked porous and (15 micron diameter) 10% crosslinked gel-type resins, Solvent Extr. Ion Exch., 8(1), 137-150 (1990). Mardan, A.; Alstad, J., and Liljenzin, J.O., Elution behavior comparison of dyno-resins for rare-earths/2-hydroxy isobutyric acid system, Solvent Extr. Ion Exch., 8(1), 151-172 (1990). Mustafa, S.; Hussain, S.Y., and Ahmad, R., Phosphate/hydroxide exchange studies on Amberlite IRA-400, Solvent Extr. Ion Exch., 8(2), 325-340 (1990). Ray, A., et al., The simulated countercurrent moving bed chromatographic reactor, Chem. Eng. Sci., 45(8), 24312438(1990). Rice, R.G., and Heft, B.K., Radial flow chromatography in compressed pancake-shaped beds, Chem. Eng. Commun., 98, 231-240 (1990). Rovere, C.E., et al., Chemical class separation of shale oils by low-pressure liquid chromatography on thermallymodified adsorbants, Fuel, 69(9), 1099-1104 (1990). Sun, Y.D.; Grevillot, G., and Tondeur, D., Modelling and optimization of the cyclic regime of an ion-exchange process for sugar juice softening, Biochem. Eng. J., 43(2), B53-B66 (1990). Takahashi, Y., and Goto, S., Adsorption isotherms of amino acids and kinetic analysis of ion-exchange chromatographs by the moment method, Sep. Sci. Technol., 25(11), 1131-1140 (1990). Trobajo, C., et al., Lamellar inorganic ion exchangers: Li+, Na +, H+ ion exchange in gamma-titanium phosphate, Solvent Extr. Ion Exch., 8(4), 729-740 (1990). Usuda, S., et al., Desorption behavior of plutonium from anion-exchange resin with HNO3-HI mixed acid solution, Sep. Sci. Technol., 25(11), 1225-1238 (1990). Vasheghani-Farahani, E., et al., Swelling of ionic gels in electrolyte solutions, Ind. Eng. Chem. Res., 29(4), 554560 (1990). Yoshida, H., and Kataoka, T., Recovery of mercury from a mercury(II)-form chelate resin by electrolytic desorption, Ind. Eng. Chem. Res., 29(10), 2152-2154 (1990). Zuyi, T., and Jinlong, N., Shell-progressive model with changing bulk concentration and exchanger volume in ion exchange, Solvent Extr. Ion Exch., 8(1), 99-116 (1990). 1991 Ackley, M.W., and Yang, R.T., Diffusion in ion-exchanged clinoptilolites, AIChE J., 37(11), 1645-1656 (1991). Adams, R.J.W., and Hudson, M.J., Reversible extraction of ionic species using electrochemically assisted ion exchange: Cobalt(II) using alpha-zirconium hydrogen phosphate, Solvent Extr. Ion Exch., 9(3), 497-514 (1991). Akintoye, A.; Ganetsos, G., and Barker, P.E., The inversion of sucrose on a semicontinuous countercurrent chromatographic bioreactor-separator, Food Bioprod. Process., 69(C 1), 35-44 (1991 ). Alen, R.; Sjostrom, E., and Suominen, S., Application of ion-exclusion chromatography to alkaline pulping liquors: Separation of hydroxy carboxylic acids from inorganic solids, J. Chem. Technol. Biotechnol., 51(2), 225-234 (1991). Baksh, M.S.A., and Yang, R.T., Chromatographic separations by pillared clay, Sep. Sci. Technol., 26(10), 13771394(1991). Barker, P.E., and Bridges, S., Continuous annular chromatography for the separation of beet molasses, J. Chem. Technol. Biotechnol., 51(3), 347-360 (1991). Barker, P.E., and Joshi, K., The recovery of fructose from inverted sugar beet molasses using continuous chromatography, J. Chem. Technol. Biotechnol., 52(1), 93-108 (1991 ). Bloomingburg, G.F., et al., Continuous separation of proteins by annular chromatography, Ind. Eng. Chem. Res., 30(5), 1061-1067 (1991). Bridger, N.J.; Jones, C.P., and Neville, M.D., Electrochemical ion exchange, J. Chem. Technol. Biotechnol., 50(4), 469-482 (1991). Bruening, M.L., et al., A novel, highly selective anion-exchange column prepared by binding Pd 2+ to an immobilized ligand, Sep. Sci. Yechnol., 26(6), 761-772 (1991). Campbell, D., and Foundos, A., Chromatographs meet environmental needs, Hydrocarbon Process., 70(2), 63-65 (1991). Cerny, J.; Sebor, G., and Mitera, J., Comparison of the selectivity of extrographic and chromatographic fractionations, Fuel, 70(7), 857-860 (1991). Das, N.R., and Lahiri, S., Liquid ion exchangers and their uses in the separation of zirconium, niobium, molybdenum, hafnium, tantalum and tungsten, Solvent Extr. Ion Exch., 9(2), 337-350 (1991).

1028

Ganetsos, G., and Barker, P.E., Large-scale chromatography in industrial processing, J. Chem. Technol. Biotechnol., 50(1), 101-108 (1991). Goetz, V., and Graves, D.J., Axial dispersion in a magnetically stabilized fluidized-bed liquid chromatography column, Powder Technol., 64(1), 81-92 (1991). Gu, T.; Tsai, G.J., and Tsao, G.T., Multicomponent adsorption and chromatography with uneven saturation capacities, AIChE J., 37(9), 1333-1340 (1991). Gu, T.; Tsai, G.J., and Tsao, G.T., A theoretical study of multicomponent radial flow chromatography, Chem. Eng. Sci., 46(5), 1279-1288 (1991). Harkins, D.A., and Schweitzer, G.K., Preparation of site-selective ion-exchange resins, Sep. Sci. Technol., 26(3), 345-354 (1991). Harries, R.R., Ion exchange kinetics in ultra-pure water systems, J. Chem. Technol. Biotechnol., 51(4), 437-448 (1991). Helfferich, F.G., The h- and w-transformations in multicomponent fixed-bed ion exchange and adsorption: Equivalent mathematics, different scope, Chem. Eng. Sci., 46(12), 3320-3323 (1991). Hsu, T.B., and Pigford, R.L., Mass transfer in a thermally regenerable ion-exchange resin by continuous cycling, Ind. Eng. Chem. Res., 30(5), 1067-1075 (1991). Huang, H., et al., The sorption behavior of boric acid on weak-base anion exchange resin, Solvent Extr. Ion Exch., 9(2), 319-336 (1991 ). Hufton, J.R., and Danner, R.P., Gas-solid diffusion and equilibrium parameters by tracer pulse chromatography, Chem. Eng. Sci., 46(8), 2079-2092 (1991). Jacobson, S.; Golshan-Shirazi, S., and Guiochon, G., Isotherm selection for band profile simulations in preparative chromatography, AIChE J., 37(6), 836-844 (1991). Jaroniec, M.; Madey, R., and Lu, X., Application of gas-solid adsorption chromatography for characterizing adsorbent heterogeneity, Sep. Sci. Technol., 26(2), 269-278 (1991). Jeng, C.Y., and Langer, S.H., Rate process analysis in the liquid chromatographic reactor: An application of the first statistical moment, Ind. Eng. Chem. Res., 30(7), 1489-1499 (1991). Kawakita, T., and Matsuishi, T., Elution kinetics of lysine from a strong cation-exchange resin with ammonia water, Sep. Sci. Technol., 26(7), 991-1004 (1991). Kawakita, T., et al., Breakthrough curve of lysine on a column of a strong cation-exchange resin of the ammonium form, Sep. Sci. Technol., 26(5), 619-636 (1991). Kawakita, T.; Matsuishi, T., and Koga, Y., Optimization of lysine adsorption process using strong cationexchange resin, Sep. Sci. Technol., 26(6), 869-884 (1991). Landau, I.; Belfer, A.J., and Locke, D.C., Measurement of limiting activity coefficients using non-steady-state gas chromatography, Ind. Eng. Chem. Res., 30(8), 1900-1906 (1991). Lobarzewski, J.; Kowalska-Pylka, H., and Cybulski, W., A simple affinity chromatography method for the separation of gastric proteases from mucous substances, J. Chem. Technol. Biotechnol., 52(3), 359-368 (1991). Lukac, M., and Perina, Z., A dynamic model of physical processes in chromatographic glucose-fructose separation, Chem. Eng. Sci., 46(4), 959-966 (1991). Ma, Z., and Guiochon, G., Application of orthogonal collocation on finite elements in the simulation of nonlinear chromatography, Comput. Chem. Eng., 15(6), 415-426 (1991). Mannhardt, K., and Novosad, J.J., Chromatographic movement of surfactant mixtures in porous media, Chem. Eng. Sci., 46(1), 75-84 (1991). Maranon, E., and Sastre, H., Ion exchange equilibria of heavy metals onto chemically modified apple residues, Solvent Extr. Ion Exch., 9(3), 515-532 (1991). Masoliver, J., and Weiss, G.H., Transport equations in chromatography with a finite speed of signal propagation, Sep. Sci. Technol., 26(2), 279-290 (1991). Moon, J.K., and Lee, W.K., Adsorption characteristics of cresols with eluent composition in adsorption chromatography, Sep. Sci. Technol., 26(5), 675-688 (1991). Oi, T., et al., Fractionation of lithium isotopes in cation-exchange chromatography, Sep. Sci. Technol., 26(10), 1353-1376(1991). Orford, C.D.; Adlard, M.W., and Perry, D., Isolation of gamma-(L-alpha-aminoadipyl)-L-cysteinyl-D-valine from culture broths by covalent chromatography, J. Chem. Technol. Biotechnol., 50(4), 523-534 (1991). Rice, R.G., and Heft, B.K., Separations via radial flow chromatography in compacted particle beds, AIChE J., 37(4), 629-632 (1991).

1029

Row, K.H., and Raw, J.I., Parameter estimation of cyclobutane pyrimidine dimers and monomers of uracil and thymine in reversed-phase high-performance liquid chromatography, Sep. Sci. Technol., 26(1), 15-24 (1991). Surakitbanharn, Y.; Muralidharan, S., and Freiser, H., Separation of palladium(II) from platinum(II), iridium(Ill), and rhodium(IIl) using centrifugal partition chromatography, Solvent Extr. Ion Exch., 9(1), 45-60 (1991 ). Suwondo, E., et al., Simulation via orthogonal collocation on finite element of a chromatographic column with nonlinear isotherm, Chem. Eng. Commun., 102, 161-188 (1991 ). Takahashi, Y., and Goto, S., Continuous separations of amino acids by using an annular chromatograph with rotating inlet and outlet, Sep. Sci. Technol., 26(1), 1-14 (1991). Takase, H., and Yoshimura, Y., Mass transfer from a slurry adsorbent to an ion-exchange resin, Int. Chem. Eng., 31(2), 351-358 (1991). Ysuji, M., and Komarneni, S., An evaluation method of chromatographic parameters from the ion-exchange isotherm of A13+-substituted tobermorite cation exchanger, Sep. Sci. Technol., 26(5), 647-660 (1991). Whitley, R.D., et al., Effects of protein aggregation in isocratic nonlinear chromatography, AIChE J., 37(4), 555568(1991). Wong, J.W.; Albright, R.L., and Wang, N.H.L., Immobilized metal ion affinity chromatography (IMAC): Chemistry and bioseparation applications, Sep. Purif. Methods, 20(1), 49-106 (1991). Worthy, W., New perfusion-chromatography separation method, Chem. Eng. News, 18 November, 25-26 (1991). Yang, B.L., and Goto, S., Complete separation of albumin and hemoglobin by metal chelate affinity chromatography, Sep. Sci. Technol., 26(5), 637-646 (1991 ). Yasuda, S., and Kawazu, K., Separation of germanium from ethylene glycol distillates by N-methylglucamine resin, Sep. Sci. Technol., 26(9), 1273-1278 (1991). Yu, Q., and Do, D.D., Reversed displacement chromatography of adsorptions with unfavourable equilibrium isotherms, Biochem. Eng. J., 46(3), B93-B98 (1991). Zecchini, E.J., and Foutch, G.L., Mixed-bed ion-exchange modeling with amine-form cation resins, Ind. Eng. Chem. Res., 30(8), 1886-1892 (1991). 1992 Alexandratos, S.D., and Kaiser, P.T., Reaction kinetics of polystyrene-based phosphinic acid ion exchange/redox resins with metal ions, Solvent Extr. Ion Exch., 10(3), 539-550 (1992). Anon., Advances in ion exchange, Chem. Eng. (N.Y.), September, 63-71 (1992). Anon., Ion exchange for esterification, Chem. Eng. (Rugby, Engl.), 10 December, 14-15 (1992). Barker, P.E., et al., Bioreaction-separation on continuous chromatographic systems, Biochem. Eng. J., 50(2), BZ3-BZ8 (1992). Bauza, R., et al., Separation of mono-, di-, and tri-stearin from an industrial mixture of glycerides by normal- and reverse-phase HPLC, Sep. Sci. Technol., 27(5), 645-662 (1992). Bhandari, V.M.; Juvekar, V.A., and Patwardhan, S.R., Sorption studies on ion exchange resins: Sorption of strong acids on weak base resins, Ind. Eng. Chem. Res., 31 (4), 1060-1073 (1992). Bhandari, V.M.; Juvekar, V.A., and Patwardhan, S.R., Sorption studies on ion exchange resins: Sorption of weak acids on weak base resins, Ind. Eng. Chem. Res., 31(4), 1073-1080 (1992). Bhandari, V.M.; Juvekar, V.A., and Patwardhan, S.R., Modified shrinking core model for reversible sorption on ion-exchange resins, Sep. Sci. Yechnol., 27(8), 1043-1064 (1992). Bhattacharyya, D.K., and Dutta, N.C., Role of hydrous titanium oxide on the uptake of several tracer cations, and separation of carrier-free 125mTefrom I25Sb and 13Zlfrom 132Te,Sep. Sci. Technol., 27(3), 399-408 (1992). Binous, H., and McCoy, B.J., Chromatographic reactions of three components: Application to separations, Chem. Eng. Sci., 47(17), 4333-4344 (1992). Bricio, O.; Coca, J., and Sastre, H., A comparative study of kinetic models for ion-exchange using macroporous resins and concentrated solutions, Solvent Extr. Ion Exch., 10(2), 381-400 (1992). Bridges, S., and Barker, P.E., Modelling continuous chromatographic separations, Chem. Eng. Sci., 47(5), 12991306 (1992). Brooks, C.A., and Cramer, S.M., Steric mass-action ion exchange: Displacement profiles and induced salt gradients, AIChE J., 38(12), 1969-1978 (1992). Calvarin, L.; Roche, B., and Renon, H., Anion exchange and aggregation of dicyanocobalamin with quaternary ammonium salts in apolar environment, Ind. Eng. Chem. Res., 31(7), 1705-1709 (1992). Carta, G., et al., Chromatography of reversibly reacting mixtures: Mutarotation effects in sugar separations, Chem. Eng. Sci., 47(7), 1645-1658 (1992).

1030

Cavender, M.R.; Chiang, H.L., and Myers, K., Optimize ion exchange resins replacement, Chem. Eng. Prog., 88(9), 56-59 (1992). Chase, H.A., and Draeger, N.M., Expanded-bed adsorption of proteins using ion-exchangers, Sep. Sci. Technol., 27(14), 2021-2040 (1992). Chiarizia, R.; Horwitz, E.P., and Dietz, M.L., Acid dependency of the extraction of selected metal ions by a strontium-selective extraction chromatographic resin: Calculated vs. experimental curves, Solvent Extr. Ion Exch., 10(2), 337-362 (1992). de Bokx, P.K.; Baarslag, P.C., and Urbach, H.P., Calculation and experimental verification of solute retention in liquid chromatography using binary eluents, Sep. Sci. Technol., 27(7), 875-900 (1992). De Lucas, A.; Zarca, J., and Canizares, P., Ion-exchange equilibrium of Ca2+, Mg 2+, K +, Na +, and H+ ions on Amberlite IR-120: Experimental determination and theoretical prediction of the ternary and quaternary equilibrium data, Sep. Sci. Technol., 27(6), 823-842 (1992). Durao, M.I.G.; Costa, C.A.V., and Rodrigues, A.E., Saturation and regeneration of ion exchangers with volume changes, Ind. Eng. Chem. Res., 31 (11), 2564-2572 (1992). Eccles, H., and Greenwood, H., Chelate ion-exchangers: The past and future applications, a user's view, Solvent Extr. Ion Exch., 10(4), 713-728 (1992). Economopoulos, N., et al., A plant kinetic study of alcoholic fermentation using reversed-flow gas chromatography, Sep. Sci. Technol., 27(15), 2055-2070 (1992). EI-Naggar, I.M., and Aly, H.F., Kinetics of Cs +, Sc3+, and Eu 3+ exchange on crystalline atimonic acid, Solvent Extr. Ion Exch., 10(1), 145-158 (1992). Gu, T., et al., Modeling of gradient elution in multicomponent non-linear chromatography, Chem. Eng. Sci., 47(1), 253-262 (1992). Heininger, M.W., and Meloan, C.E., A resin with selectivity for the removal and recovery of chromate from contaminated water, Solvent Extr. Ion Exch., 10(1), 159-172 (1992). Horwitz, E.P.; Chiarizia, R., and Dietz, M.L., A novel strontium-selective extraction chromatographic resin, Solvent Extr. Ion Exch., 10(2), 313-336 (1992). Hossain, M.M., and Do, D.D., The effects of denaturation in the displacement chromatographic behaviour of proteins, Biochem. Eng. J., 49(3), B29-B39 (1992). Huang, S.Y., and Jin, J.D., Operation strategy for displacement chromatography: Selection of optimum mobile phase for separation of weak adsorptive nucleotides, Chem. Eng. Sci., 47(1), 21-30 (1992). Kaur, P., et al., Studies on the sorption behaviour of some amino acids on silica gel pretreated with alkalis in relation to chromatography, Adsorpt. Sci. Technol., 8(3), 157-173 (1992). Kim, S.U., et al., Peak compression in stepwise pH elution with flow reversal in ion exchange chromatography, Ind. Eng. Chem. Res., 31(7), 1717-1730 (1992). Larson, K.A., and Wiencek, J.M., Liquid ion exchange for mercury removal from water over a wide pH range, Ind. Eng. Chem. Res., 31(12), 2714-2722 (1992). Lee, K.N., and Lee, W.K., A theoretical model for the separation of glucose and fructose mixtures by using a semicontinuous chromatographic refiner, Sep. Sci. Technol., 27(3), 295-312 (1992). Leung, B.K.O., and Hudson, M.J., A novel weak-base anion-exchange resin which is highly selective for the precious metals over base metals, Solvent Extr. Ion Exch., 10( 1), 173-190 (1992). Levy, D., et al., Immobilization of quaternary ammonium anion exchangers in sol-gel glasses, Sep. Sci. Technol., 27(5), 589-598 (1992). Lewandowski, R., and Lameloise, M.L., Study of exclusion equilibrium between a sucrose-NaCl solution and an ion exchange resin, Chem. Eng. Process., 31(4), 207-212 (1992). Mijangos, F., and Diaz, M., Metal-proton equilibrium relations in a chelating iminodiacetic resin, Ind. Eng. Chem. Res., 31(11), 2524-2532 (1992). Miyabe, K., and Suzuki, M., Chromatography of liquid-phase adsorption on octadecylsilyl-silica gel, AIChE J., 38(6), 901-910 (1992). Mohammad, A.W.; Stevenson, D.G., and Wankat, P.C., Pressure drop correlations and scale-up of size exclusion chromatography with compressible packings, Ind. Eng. Chem. Res., 31(2), 549-561 (1992). Oi, T., et al., Fractionation of strontium isotopes in cation-exchange chromatography, Sep. Sci. Technol., 27(5), 631-644 (1992). Olson, K.C., and Gehant, R.L., Applications of ultrafast HPLC to process development of recombinant DNAderived proteins, Biotechnol. Prog., 8(6), 562-566 (1992).

1031

Panneman, H.J., and Beenackers, A.A.C.M., Solvent effects on the hydration of cyclohexene catalyzed by a strong acid ion exchange resin: Solubility of cyclohexene in aqueous sulfolane mixtures, Ind. Eng. Chem. Res., 31(4), 1227-1231 (1992). Park, C.M., and Meyer, W., Separation of ~37Cs, 9~ and 232Th in aqueous solution by using a multistage countercurrent batch contactor ion-exchange system, Sep. Sci. Technol., 27(2), 223-238 (1992). Rodrigues, A.E., et al., Influence of adsorption-desorption kinetics on the performance of chromatographic processes using large-pore supports, Chem. Eng. Sci., 47(17), 4405-4414 (1992). Samanta, S.K.; Ramaswamy, M., and Misra, B.M., Studies on cesium uptake by phenolic resins, Sep. Sci. Yechnol., 27(2), 255-268 (1992). Savkovic-Stevanovic, J., et al., Reaction distillation with ion exchangers, Sep. Sci. Technol., 27(5), 613-630 (1992). Sengupta, A.K., and Zhu, Y., Metals sorption by chelating polymers: A unique role of ionic strength, AIChE J., 38(1), 153-157 (1992). Soldatov, V.S., Mathematical modelling of ion exchange equilibria, J. Chem. Technol. Biotechnol., 55(3), 298300 (1992). Tsuji, M., and Komarneni, S., An extended method for analytical evaluation of distribution coefficients on selective inorganic ion exchangers, Sep. Sci. Technol., 27(6), 813-822 (1992). Velayudhan, A., and Ladisch, M.R., Effect of modulator sorption in gradient elution chromatography: Gradient deformation, Chem. Eng. Sci., 47(1), 233-240 (1992). Viard, V., and Lameloise, M.L., Modelling glucose-fructose separation by adsorption chromatography on ion exchange resins, J. Food Eng., 17(1), 29-48 (1992). Yang, B.L., and Goto, S., Separation and concentration of adenosine triphosphate and adenosine monophosphate by using two chromatographic columns, Sep. Sci. Technol., 27(4), 547-556 (1992). Yoshida, H.; Shimizu, K., and Kataoka, T., Recovery of amine and paints from electrodeposition wastewater by an H-form ion exchanger: Desorption process, Ind. Eng. Chem. Res., 31(3), 934-941 (1992). 1993 Ashrafizadeh, S.N.; Weber, M.E., and Vera, J.H., Cation exchange with reverse micelles, Ind. Eng. Chem. Res., 32(1), 125-132 (1993). Besirli, N., and Baysal, B.M., Ion-exchange studies with some complex ions on ion-exchange resil~s, Solvent Extr. Ion Exch., 11(3), 541-554 (1993). Bhandari, V.M.; Juvekar, V.A., and Patwardhan, S.R., Sorption of dibasic acids on weak base resins, Ind. Eng. Chem. Res., 32( 1), 200-206 (1993). Blazy, P., et al., Selective recovery of rhenium from gas-scrubbing solutions of molybdenite roasting using direct precipitation and separation on resins, Sep. Sci. Technol., 28(11), 2073-2096 (1993). Carta, G., and Rodrigues, A.E., Diffusion and convection in chromatographic processes using permeable supports with a bidisperse pore structure, Chem. Eng. Sci., 48(23), 3927-3935 (1993). Chiarizia, R., et al., Uptake of metal ions by a new chelating ion-exchange resin: Acid dependencies of transition and post-transition metal ions, Solvent Extr. Ion Exch., 11(5), 967-986 (1993). Choudhary, V.R., and Mayadevi, S., Adsorption of methane, ethane, ethylene, and carbon dioxide on high-silica pentasil zeolites and zeolite-like materials using gas chromatography pulse technique, Sep. Sci. Technol., 28(13), 2197-2210 (1993). Egawa, H., et al., Recovery of uranium from seawater: Long-term stability tests for high-performance chelating resins containing amidoxime groups and evaluation of elution process, Ind. Eng. Chem. Res., 32(3), 540-547 (1993). Egawa, H., et al., Recovery of uranium from seawater: System arangements for the recovery of uranium from seawater by spherical amidoxime chelating resins utilizing natural seawater motions, Ind. Eng. Chem. Res., 32(4), 709-715 (1993). EI-Naggar, I.M., et al., Ion-exchange equilibrium of the CuZ+/H+, ZnZ+/H+ and pb2+/H§ ions on hydrated ferric oxide, Solvent Extr. Ion Exch., 11(4), 683-692 (1993). Felinger, A., and Guiochon, G., The change of pressure drop during large-scale chromatography of viscous samples, Biotechnol. Prog., 9(5), 450-455 (1993). Fernandez, A.; Suarez, C., and Diaz, M., Kinetics of metal ion exchange in iminodiacetic resins at low concentrations, J. Chem. Technol. Biotechnol., 58(3), 255-260 (1993). Fish, B.B.; Carr, R.W., and Aris, R., Optimization of the countercurrent moving-bed chromatographic separator, AIChE J., 39(10), 1621-1627 (1993).

1032

Fries, W., and Chew, D., Ion exchange to remove heavy metals, Chemtech, 23(2), 32-35 (1993). Gonzalez-Patino, F.; Catalan, J., and Galan, M.A., Affinity chromatography: Effect of particle size on adsorption equilibrium and mass transfer kinetics, Chem. Eng. Sci., 48(9), 1567-1574 (1993). Hashim, M.A.; Chu, K.H., and Tsan, P.S., Ion-exchange equilibria of conalbumin and myoglobin, Food Bioprod. Process., 71(C4), 273-278 (1993). Hayashita, T., et al., Effect of ring-size variation within dibenzocrown ether resins upon ion-pair sorption of alkali-metal cations from aqueous and aqueous methanol solutions, Sep. Sci. Technol., 28(17), 2607-2620 (1993). Hejtmanek, V., and Schneider, P., Axial dispersion under liquid-chromatography conditions, Chem. Eng. Sci., 48(6), 1163-1168 (1993). Horwitz, E.P., et al., Uptake of metal ions by a new chelating ion-exchange resin: Acid dependencies of actinide ions, Solvent Extr. Ion Exch., 11(5), 943-966 (1993). Huhn, G.F., et al., Purification of nucleoside-5'-diphosphates: A new ion-exchange method, Sep. Sci. Technol., 28(11), 1959-1970 (1993). Jones, I.L., and Carta, G., Ion exchange of amino acids and dipeptides on cation resins with varying degree of cross-linking: Equilibrium, Ind. Eng. Chem. Res., 32( 1), 107-116 (1993). Jones, I.L., and Carta, G., Ion exchange of amino acids and dipeptides on cation resins with varying degree of cross-linking: Intraparticle transport, Ind. Eng. Chem. Res., 32(1), 117-125 (1993). Kaneko, H.; Tsuji, M., and Tamaura, Y., Thermodynamic study of M3+/H+ exchange systems on titanium antimonate cation exchanger, Solvent Extr. Ion Exch., 11(4), 693-712 (1993). Komatsu, Y.; Fujiki, Y., and Sasaki, T., Ion-exchange separation of sodium and potassium ions on dihydrogen tetratitanate hydrate fibers at various temperatures, Solvent Extr. Ion Exch., 11(1), 159-169 (1993). Konishi, Y., et al., Recovery of zinc, cadmium, and lanthanum by biopolymer gel particles of alginic acid, Sep. Sci. Technol., 28(9), 1691-1702 (1993). Koutake, M., et al., Osmotic pressure model of membrane fouling applied to the ultrafiltration of whey, J. Food Eng., 18(4), 313-334 (1993). Kraaijeveld, G., and Wesselingh, J.A., The kinetics of film-diffusion-limited ion exchange, Chem. Eng. Sci., 48(3), 467-474 (1993). Lee, W.C., Analysis of preparative chromatography by local equilibrium model, Chem. Eng. Commun., 122, 6984(1993). Lee, W.C.; Tsai, G.J., and Tsao, G.T., Analysis of chromatography by plate theory, Sep. Technol., 3(4), 178-197 (1993). Locke, B.R., and Arce, P., Modeling electrophoretic transport of polyelectrolytes in beds of nonporous spheres, Sep. Technol., 3(2), 111-120 (1993). Lumetta, G.J., et al., Preliminary evaluation of chromatographic techniques for the separation of radionuclides from high-level radioactive waste, Solvent Extr. Ion Exch., 11(4), 663-682 (1993). Luo, R.G., and Hsu, J.T., Intraparticle protein diffusion effect on gradient elution chromatography, Sep. Technol., 3(4), 221-229 (1993). Mak, A.N.S., et al., Continuous ion exchange in a pulsed packed column containing structured packing: Masstransfer-controlled kinetics, Chem. Eng. Sci., 48(4), 701-714 (1993). Metwally, M.S., and Samy, T.M., Selectivity of M+-H+ ion-exchange absorption on sulfonic resins in ternary solutions, Sep. Sci. Technol., 28(13), 2273-2278 (1993). Ming, F., and Howell, J.A., Parameter estimation for a column adsorption model incorporating axial dispersion: Application to a novel monolithic ion-exchange column, Food Bioprod. Process., 71(C4), 267-272 (1993). Oi, T., et al., Fractionation of calcium isotopes in cation-exchange chromatography, Sep. Sci. Technol., 28(11), 1971-1984(1993). Park, W.K., and Michaels, E.D., Displacement band chromatography of hydrogen sulfites for enrichment of sulfur isotopes, Sep. Sci. Yechnol., 28(1), 477-486 (1993). Perona, J.J., Model for Sr-Cs-Ca-Mg-Na ion-exchange equilibria on chabazite, AIChE J., 39(10), 1716-1720 (1993). Quinta Ferreira, R.M., and Rodrigues, A.E., Diffusion and catalytic zero-order reaction in a macroreticular ion exchange resin, Chem. Eng. Sci., 48(16), 2927-2950 (1993). Robinson-Piergiovanni, P.S.; Crane, L.J., and Nau, D.R., Solid-phase extraction columns: A tool for teaching biochromatography, Chem. Eng. Educ., 27(1), 34-37 (1993).

1033

Romdhane, I.H., and Danner, R.P., Polymer-solvent diffusion and equilibrium parameters by inverse gas-liquid chromatography, AIChE J., 39(4), 625-635 (1993). Rudge, S.R.; Basak, S.K., and Ladisch, M.R., Solute retention in electrochromatography by electrically induced sorption, AIChE J., 39(5), 797-808 (1993). Sarmidi, M.R., and Barker, P.E., Saccharification of modified starch to maltose in a continuous rotating annular chromatograph, J. Chem. Technol. Biotechnol., 57(3), 229-236 (1993). Sarmidi, M.R., and Barker, P.E., Simultaneous biochemical reaction and separation in a rotating annular chromatograph, Chem. Eng. Sci., 48(14), 2615-2624 (1993). Sato, K., et al., Temperature gradient method for continuous countercurrent gas-liquid chromatography, Sep. Sci. Technol., 28(7), 1409-1420 (1993). Schisla, D.K., et al., Polydisperse tube diameters compromise multiple open tubular chromatography, AIChE J., 39(6), 946-953 (1993). Seidel-Morgenstern, A., and Guiochon, G., Modelling of the competitive isotherms and the chromatographic separation of two enantiomers, Chem. Eng. Sci., 48(15), 2787-2798 (1993). Seidel-Morgenstern, A., and Guiochon, G., Theoretical study of recycling in preparative chromatography, AIChE J., 39(5), 809-819 (1993). Stenger, H.G.; Hu, K., and Simpson, D.R., Chromatographic separation and concentration of sulfur dioxide in flue gases, Ind. Eng. Chem. Res., 32(11), 2736-2739 (1993). Suwondo, E., et al., Optimization of a liquid chromatographic separation process, Comput. Chem. Eng., 17(supplement), S135-S140 (1993). Tao, Z., and Wang, C., Determination of ion exchange equilibrium constants for weakly dissociating ion exchange resins, Solvent Extr. Ion Exch., 11(4), 713-728 (1993). Tsuji, M.; Komarneni, S., and Abe, M., Ion-exchange selectivity for alkali metal ions on a cryptomelane-type hydrous manganese dioxide, Solvent Extr. Ion Exch., 11(1), 143-158 (1993). Webb, S.W., Multicomponent inverse gas chromatography for analyses of sorption in polymers, AIChE J., 39(4), 701-706 (1993). Whitley, R.G.; Van Cott, K.E., and Wang, N.Iq.L., Analysis of nonequilibrium adsorption/desorption kinetics and implications for analytical and preparative chromatography, Ind. Eng. Chem. Res., 32(1), 149-159 (1993). Xue, T., and Osseo-Asare, K., Behavior of silver-thiourea complexes in nation resin, Sep. Sci. Yechnol., 28(4), 1077-1084 (1993). Yamamoto, S.; Suehisa, T., and Sano, Y.J., Preparative separation of proteins by gradient-elution and stepwiseelution chromatography: Zone-sharpening effect, Chem. Eng. Commun., 119, 221-230 (1993). Yonemoto, T., et al., A novel continuous rotating annular liquid chromatograph with a multichannel peristaltic pump for variable eluent withdrawal, Sep. Sci. Technol., 28(17), 2587-2606 (1993). Zhong, G.M., and Meunier, F., Linear perturbation chromatography theory: Moment solution for two-component nonequilibrium adsorption, Chem. Eng. Sci., 48(7), 1309-1316 (1993). Zhong, G.M., and Meunier, F., Interference theory: Moment solution for two-component nonequilibrium adsorption chromatography, Chem. Eng. Sci., 48(24), 4105-4108 (1993). Zhong, G.M., and Meunier, F., Interference theory: Moment solution for three-component nonequilibrium adsorption chromatography, Chem. Eng. Sci., 48(24), 4109-4114 (1993). Zhu, J.; Ma, Z., and Guiochon, G., The thickness of shock layers in liquid chromatography, Biotechnol. Prog., 9(4), 421-428 (1993). Zuyi, T., and Changshou, W., Determination of ion exchange equilibrium constants for weakly dissociating ion exchange resins, Solvent Extr. Ion Exch., 11(2), 171-186 (1993). 1994 Alan, D.J., and Franses, E.I., Ion adsorption and ion exchange in ultrathin films of fatty acids, AIChE J., 40(6), 1046-1054 (1994). Ashley, K.R., et al., Sorption behavior of pertechnetate on Reillex-HPQ anion exchange resin from nitric acid solution, Solvent Extr. Ion Exch., 12(2), 239-260 (1994). Bhagat, R.D., and Turel, Z.R., Radiochemical separation of thallium(I) using cerium(IV) molybdate as an ionexchanger, Sep. Sci. Yechnol., 29(5), 663-670 (1994). Bloomingburg, G.F., and Carta, G., Separation of protein mixtures by continuous annular chromatography with step elution, Biochem. Eng. J., 55(1), B19-B28 (1994). Carlsson, F.; Axelsson, A., and Zacchi, G., Mathematical modelling and parametric studies of affinity chromatography, Comput. Chem. Eng., 18(supplement), $657-$662 (1994).

1034

Carrere, H., et al., Whey proteins extraction by fluidized ion exchange chromatography: Isotherms determination and process modelling, Food Bioprod. Process., 72(C4), 216-226 (1994). Chang, R.M., and Lee, W.C., An affinity adsorbent derived from aminopropyl silica for serine protease chromatography, J. Chem. Technol. Biotechnol., 59(2), 133-140 (1994). Chaudhary, A.J., et al., Heavy metals in the environment: Anion exchange properties of poly-4-vinyl pyridine from acid chloride solutions, J. Chem. Technol. Biotechnol., 60(4), 353-358 (1994). Chiarizia, R., and Horwitz, E.P., Uptake of metal ions by a new chelating ion-exchange resin: Calculations on the effect of complexing anions on actinides, Solvent Extr. Ion Exch., 12(4), 847-872 (1994). Chiarizia, R.; Horwitz, E.P., and Alexandratos, S.D., Uptake of metal ions by a new chelating ion-exchange resin: Kinetics, Solvent Extr. Ion Exch., 12(1), 211-237 (1994). Ching, C.B.; Chu, K.H., and Hidajat, K., Multicomponent separation using a column-switching chromatographic method, AIChE J., 40(11), 1843-1849 (1994). Cortina, J.L., et al., Solvent impregnated resins containing di-(2-ethylhexyl) phosphoric acid: Preparation and study of the retention and distribution of the extractant on the resin, Solvent Extr. Ion Exch., 12(2), 349-370 (1994). Cortina, J.L., et al., Solvent impregnated resins containing di-(2-ethylhexyl) phosphoric acid: Study of the distribution equilibria of Zn(II), Cu(II) and Cd(II), Solvent Extr. Ion Exch., 12(2), 371-392 (1994). Davies, V.R., Troubleshoot ion-exchange equipment, Chem. Eng. Prog., 90(1), 63-71 (1994). de Lucas Martinez, A.; Zarca Diaz, J., and Canizares, P.C., Ion-exchange equilibrium in a binary mixture: Models for its characterization, Int. Chem. Eng., 34(4), 486-497 (1994). DeSilva, F., Ion exchanger design, Chem. Eng. (N.Y.), July, 86-88 (1994). El-Naggar, I.M." Abdel Hamid, M.M., and Aly, H.F., Kinetics and mechanism of isotopic exchange for Co2+/*Co2+ in tin(IV) antimonate, Solvent Extr. Ion Exch., 12(3), 651-665 (1994). Felinger, A., and Guiochon, G., Optimizing experimental conditions for minimum production cost in preparative chromatography, AIChE J., 40(4), 594-605 (1994). Fernandez, A.; Rendueles, M., and Diaz, M., Co-ion behavior at high concentration cationic ion exchange, Ind. Eng. Chem. Res., 33(11), 2789-2794 (1994). Fernandez, A.; Rodrigues, A.E., and Diaz, M., Modelling of K-Na exchange in fixed beds with highly concentrated feed, Chem. Eng. J., 54(1), 17-22 (1994). Giona, M., et al., Simplified analysis of chromatographic-column dynamics, Chem. Eng. Sci., 49(4), 541-548 (1994). Gorry, M.; Amin, P., and Richardson, D.W., Design of demineralizers, Chem. Eng. (N.Y.), March, 112-118 (1994). Grzywnowicz, K., and Lobarzewski, J., A purification method for specific serine proteases using one-step affinity chromatography, J. Chem. Technol. Biotechnol., 60(2), 153-160 (1994). Guria, C., and Chanda, M., Shell-core models for ion-exchanger loading in finite bath: Sorption of aqueous sulphur dioxide on cross-linked poly(4-vinyl pyridine), Chem. Eng. Res. Des., 72(4), 503-512 (1994). Gusler, G.M., and Cohen, Y., Equilibrium swelling of highly cross-linked polymeric resins, Ind. Eng. Chem. Res., 33(10), 2345-2357 (1994). Hairston, D., Markets for ion exchange resins, Chem. Eng. (N.Y.), June, 57-59 (1994). Haupt, R.A., and Sellers, T., Characterizations of phenol-formaldehyde resol resins, Ind. Eng. Chem. Res., 33(3), 693-697 (1994). Horwitz, E.P.; Chiarizia, R., and Alexandratos, S.D., Uptake of metal ions by a new chelating ion-exchange resin: The effect of solution matrix on actinides, Solvent Extr. Ion Exch., 12(4), 831-846 (1994). Ihsanullah, H., Optimization of various factors for the separation of technetium using anion-exchange resins, Sep. Sci. Yechnol., 29(2), 239-248 (1994). Kabay, N., Use of weak-acid cation-exchange resins Purolite C105(H +) and Purolite C106(H § for the adsorption of UO22+, Sep. Sci. Yechnol., 29(5), 679-683 (1994). Kallrath, J., et al., Simulation of chromatographic reactors, Comput. Chem. Eng., 18(supplement), $331-$336 (1994). Kitakawa, A.; Yonemoto, T., and Tadaki, T., A mathematical model for the separation of amino acids using ion exchange chromatography, Food Bioprod. Process., 72(C4), 201-208 (1994). McCoy, B.J. and Goto, M., Continuous-mixture model of chromatographic separations, Chem. Eng. Sci., 49(14), 2351-2358 (1994).

1035

Mehablia, M.A.; Shallcross, D.C., and Stevens, G.W., Prediction of multicomponent ion exchange equilibria, Chem. Eng. Sci., 49(14), 2277-2286 (1994). Mijangos, F., and Diaz, M., Kinetic of copper ion exchange onto iminodiacetic resin, Can. J. Chem. Eng., 72(6), 1028-1035 (1994). Miyabe, K., and Suzuki, M., Mass-transfer phenomena on the surface of adsorbents in reversed-phase chromatography, Ind. Eng. Chem. Res., 33(7), 1792-1802 (1994). Murty, B.N.; Yadav, R.B., and Syamsundar, S., Preparation of high-purity zirconia from zircon: An anionexchange purification process, Sep. Sci. Yechnol., 29(2), 249-260 (1994). Nash, K.L., et al., Uptake of metal ions by a new chelating ion exchange resin: Protonation constants via potentiometric titration and solid state 3~p-NMR spectroscopy, Solvent Extr. Ion Exch., 12(1), 193-210 (1994). Parkinson, G., Industrial scale innovations in chromatography, Chem. Eng. (N.Y.), August, 30-33 (1994). Pehlivan, E., and Yildiz, S., Column chromatography and kinetics of nucleosides and nucleic acid bases on immobilized nickel- and cobalt-CDAE-sporopollenin, Sep. Sci. Technol., 29(7), 887-896 (1994). Pehlivan, E., et al., Sorption of heavy metal ions on new metal-ligand complexes chemically derived from Lycopodium clavatum, Sep. Sci. Technol., 29(13), 1757-1768 (1994). Plante, L.D.; Romano, P.M., and Fernandez, E.J., Viscous fingering in chromatography visualized via magnetic resonance imaging, Chem. Eng. Sci., 49(14), 2229-2242 (1994). Ray, A.K.; Carr, R.W., and Aris, R., The simulated countercurrent moving bed chromatographic reactor: A novel reactor-separator, Chem. Eng. Sci., 49(4), 469-480 (1994). Robinson, S.M.; Arnold, W.D., and Byers, C.H., Mass-transfer mechanisms for zeolite ion exchange in wastewater treatment, AIChE J., 40(12), 2045-2054 (1994). Sabharwal, K.N.; Vasudeva Rao, P.R., and Srinivasan, M., Extraction of actinides by bifunctional phosphinic acid resin, Solvent Extr. Ion Exch., 12(5), 1085-1102 (1994). Suchorebraya, S.A., et al., On the studies of molybdenum(Vl) sorption on titanium phosphate's ion exchangers, Solvent Extr. Ion Exch., 12(1), 173-192 (1994). Sun, D., et al., Separation and recovery of nickel and molybdenum using continuous rotating annular chromatography, Sep. Sci. Technol., 29(7), 831-844 (1994). Takahashi, Y., and Goto, S., Continuous separation of fructo-oligosaccharides using an annular chromatograph, Sep. Sci. Technol., 29(10), 1311-1318 (1994). Tan, H.K.S., and Spinner, I.H., MuIticomponent ion exchange column dynamics, Can. J. Chem. Eng., 72(2), 330341 (1994). Trochimczuk, A.W.; Horwitz, E.P., and Alexandratos, S.D., Complexing properties of diphonix, a new chelating resin with diphosphonate ligands, toward Ga(III) and In(IlI), Sep. Sci. Technol., 29(4), 543-550 (1994). Tyc, I., and Green, B.R., Phenol-formaldehyde based weak-base resins for the recovery of gold, Solvent Extr. Ion Exch., 12(4), 817-830 (1994). Vamos, R.J., and Haas, C.N., Reduction of ion-exchange equilibria data using an error in variables approach, AIChE J., 40(3), 556-569 (1994). Whitley, R.D.; Zhang, X., and Wang, N.H.L., Protein denaturation in nonlinear isocratic and gradient elution chromatography, AIChE J., 40(6), 1067-1081 (1994). 1995 Ahrnad, J., The use of impregnated silica gel layers and modified celluloses in thin-layer chromatographic analysis in inorganic mixtures (review paper), Sep. Sci. Technol., 30(12), 2429-2454 (1995). Alexandratos, S.D., and Hussain, L.A., Bifunctionality as a means of enhancing complexation kinetics in selective ion exchange resins, Ind. Eng. Chem. Res., 34(1), 251-254 (1995). Antia, F.D.; Fellegvari, I., and Horvath, C., Displacement of proteins in hydrophobic interaction chromatography, Ind. Eng. Chem. Res., 34(8), 2796-2810 (1995). Ashley, K.R., et al., Sorption behavior of perrhenate and oxochromium(VI) on Reillex-HPQ anion exchange resin from nitric acid solution, Solvent Extr. Ion Exch., 13(2), 353-368 (1995). Ayar, A., Yildiz, S., and Pehlivan, E., Ligand-exchange chromatography of some amino acids on Co(II)-loaded CMDAE-sporopollenin resin, Sep. Sci. Yechnol., 30(15), 3081-3086 (1995). Bilewicz, A., and Narbutt, J., Specific and nonspecific interactions of amine complexes of silver, zinc, and cadmium with ion exchangers, Solvent Extr. Ion Exch., 13(6), 1083-1096 (1995). Bosch, P., et al., Co 2+ ion exchange in zeolite NaA, Sep. Sci. Technol., 30(17), 3399-3403 (1995).

1036

Bunzl, K., Calculation of the Helfferich number to identify the rate-controlling step of ion exchange for a batch process, Ind. Eng. Chem. Res., 34(8), 2584-2587 (1995). Cannell, K.P., and Vincent, J.B., Metal ion concentration, time, and pH dependence of metal ion binding to a transferrin metalloprotein affinity chromatography matrix, J. Chem. Technol. Biotechnol., 64(1), 96-100 (1995). Chiarizia, R., et al., Uptake of metal ions by a new chelating ion-exchange resin: Alkaline earth cations, Solvent Extr. Ion Exch., 13(6), 1063-1082 (1995). Chowdiah, V., and Foutch, G.L., A kinetic model for cationic-exchange-resin regeneration, Ind. Eng. Chem. Res., 34(11 ), 4040-4048 (1995). Clearfield, A., Inorganic ion exchangers: A technology ripe for development, Ind. Eng. Chem. Res., 34(8), 28652872 (1995). Cloete, F.L.D., and Marais, A.P., Recovery of very dilute acetic acid using ion exchange, Ind. Eng. Chem. Res., 34(7), 2464-2467 (1995). De Kock, F.P., and Van Deventer, J.S.J., Statistical thermodynamic model for competitive ion exchange, Chem. Eng. Commun., 135, 21-46 (1995). de Lucas, A., et al., Ion-exchange equilibria of DL-lysine monohydrochloride on Amberlite IRA-420, Solvent Extr. Ion Exch., 13(6), 1123-1144 (1995). de Lucas, A.; Canizares, P., and Rodriguez, J.F., Removal of alkaline catalysts from polyols by ion exchange: Selection of an ion-exchange resin, Sep. Sci. Technol., 30(1), 125-140 (1995). de Lucas, A.; Canizares, P., and Rodriguez, J.F., Removal of alkaline catalysts from polyols by ion exchange: Regeneration process optimization, Sep. Sci. Technol., 30(6), 949-962 (1995). Ersoz, M.; Vural, U.S., and Yildiz, S., Sorption of aromatic amines on a new ligand exchanger of sporopolleninbound Co 2+ ion, Sep. Sci. Technol., 30(18), 3555-3566 (1995). Farschman, C.A.; Manos, J.B., and Pinto, N.G., Short-fiber chromatography columns: Potential for process-scale bioseparations, Sep. Sci. Technol., 30(7), 1325-1350 (1995). Fernandez, A.; Diaz, M., and Rodrigues, A., Kinetic mechanisms in ion exchange processes, Chem. Eng. J., 57(1), 17-26 (1995). Franzreb, M.; Holl, W.H., and Eberle, S.H., Liquid-phase mass transfer in multicomponent ion exchange: Systems with irreversible chemical reactions in the film, Ind. Eng. Chem. Res., 34(8), 2670-2675 (!995). Frey, D.D." Barnes, A., and Strong, J., Numerical studies of multicomponent chromatography using pH gradients, AIChE J., 41(5), 1171-1183 (1995). Gadam, S.D.; Gallant, S.R., and Cramer, S.M., Transient profiles in ion-exchange displacement chromatography, AIChE J., 41(7), 1676-1686 (1995). Gerstner, J.A., et al., Take another look at displacement chromatography, Chemtech, 25(11), 27-32 (1995). Hashim, M.A.; Chu, K.H., and Tsan, P.S., Effects of ionic strength and pH on the adsorption equilibria of lysozyme on ion exchangers, J. Chem. Technol. Biotechnol., 62(3), 253-260 (1995). Hayashi, H., et al., Thermodynamics of alkali and alkaline earth metal ion-exchange on cerium(IV) hydrogen phosphate, Solvent Extr. Ion Exch., 13(6), 1145-1158 (1995). Heuer, C.; Seidel-Morgenstern, A., and Hugo, P., Experimental investigation and modelling of closed-loop recycling in preparative chromatography, Chem. Eng. Sci., 50(7), 1115-1128 (1995). Hwang, S.J., and Lu, W.J., Ion exchange in a semifluidized bed, Ind. Eng. Chem. Res., 34(4), 1434-1439 (1995). Inukai, Y., et al., Separation of fullerenes by chromatography on coal, Fuel, 74(1), 83-87 (1995). Jarvis, N.V.; Wagener, J.M., and Hancock, R.D., Evaluation of a diethanolamine chelating resin using two-phase potentiometry, Solvent Extr. Ion Exch., 13(3), 591-611 (1995). Jen, S.C., and Pinto, N.G., Nonlinear chromatography of beta-lactoglobulins A and B, Ind. Eng. Chem. Res., 34(8), 2685-2691 (1995). Kalinitchev, A., Effects of nonlinearity in the theory of multicomponent chromatography, Ind. Eng. Chem. Res., 34(8), 2625-2633 (1995). Kataoka, T., et al., Mass transfer of dilute soluble silica in a fixed bed of an OH--type strong anion exchange resin, Chem. Eng. Res. Des., 73(1), 40-44 (1995). Kitakawa, A., et al., Modeling and simulation of continuous rotating annular ion-exchange chromatography for separation of amino acids, Sep. Sci. Yechnol., 30(16), 3089-3110 (1995). Koh, J.H.; Wang, N.H.L., and Wankat, P.C., Ion exchange of phenylalanine in fluidized/expanded beds, Ind. Eng. Chem. Res., 34(8), 2700-2711 (1995).

1037

Krishnamurthy, R., et al., Conditions promoting metal-catalyzed oxidations during immobilized Cu-iminodiacetic acid metal affinity chromatography, Biotechnol. Prog., 11(6), 643-650 (1995). Ladisch, M.; Westgate, P., and Brewer, M., Framework for correlating composition dependent equilibrium conversion in methyl tert-butyl ether formation by ion-exchange catalysts, Ind. Eng. Chem. Res., 34(8), 28112816(1995). Law, R.V.; Sherrington, D.C., and Korosu, H., Solid state 13C-MAS/NMR studies of anion exchange resins and their precursors, Ind. Eng. Chem. Res., 34(8), 2740-2749 (1995). Lee, K., and Hong, J., Separation and recovery of lead by cation exchange process combined with precipitation, AIChE J., 41(12), 2653-2657 (1995). Lee, W.C.; Hsiao, C.C., and Ruaan, R.C., Affinity chromatography of glucose-specific lectin using silica-based support, J. Chem. Yechnol. Biotechnol., 64(1), 66-72 (1995). Lira, B.G., et al., Recovery of praziquantel from racemic mixtures by continuous chromatography and crystallisation, Chem. Eng. Sci., 50(14), 2289-2298 (1995). Limoni-Relis, B., and Schmuckler, G., Interseparation of platinum metals in concentrated solution by gel permeation chromatography, Sep. Sci. Technol., 30(3), 337-346 (1995). Lorenz, L.F., and Christiansen, A.W., Interactions of phenolic resin alkalinity, moisture content, and care behavior, Ind. Eng. Chem. Res., 34(12), 4520-4523 (1995). Maeda, H.; Egawa, H., and Jyo, A., Preparation of chelating resins selective to boric acid by functionalization of macroporous poly(glycidyl methacrylate) with 2-amino-2-hydroxymethyl-l,3-propanediol, Sep. Sci. Yechnol., 30(18), 3545-3554 (1995). Martinez, M.; Carrancio, A., and Aracil, J., Evaluation of kinetic and thermodynamic parameters of amino acids on modified divinylbenzene-polystyrene resins using a liquid chromatography technique, Ind. Eng. Chem. Res., 34(12), 4486-4493 (1995). Matsunaga, H.; Sun, W., and Chow, A., Extraction characteristics of niobium(V) and tantalum(V) on trioctylmethyl ammonium chloride-impregnated polyurethane foam, Sep. Sci. Technol., 30(16), 3229-3236 (1995). Meier, L.; Terrell, K., and Vincent, J.B., Use of thermolysin metalloprotein affinity metal chromatography in the decontamination of actinide-bearing solutions, J. Chem. Technol. Biotechnol., 64(2), 149-152 (1995). Melis, S.; Cao, G., and Morbidelli, M., A new model for the simulation of ion exchange equilibria, Ind. Eng. Chem. Res., 34(11), 3916-3924 (1995). Menendez, F., et al., Separation of cobalt(II) and nickel(II) ions from acid aqueous solutions by Co2+-2H+ ionexchange into gamma-titanium phosphate, Solvent Extr. Ion Exch., 13(1), 179-187 (1995). Miyabe, K., and Suzuki, M., Solvent effect on adsorption phenomena in reversed-phase liquid chromatography, AIChE J., 41(3), 536-547 (1995). Miyabe, K., and Suzuki, M., Chromatographic study on liquid-phase adsorption on octadecylsilyl-silica gel, AIChE J., 41(3), 548-558 (1995). Mohammad, A., and Tiwari, S., Thirty-five years of thin-layer chromatography in the analysis of inorganic anions (review paper), Sep. Sci. Yechnol., 30(19), 3577-3614 (1995). Nesbitt, A.B., and Petersen, F.W., Feasibility of recovering high-valency metal cyanide complexes with a fluidized bed of resin, Sep. Sci. Technol., 30(15), 2979-2988 (1995). Ooi, K., et al., Lithium isotope fractionations on inorganic ion-exchangers with different ion-sieve properties, Sep. Sci. Technol., 30(20), 3761-3770 (1995). Parker, L.M.; Milestone, N.B., and Newman, R.H., The use of hydrotalcite as an anion absorbent, Ind. Eng. Chem. Res., 34(4), 1196-1202 (1995). Petruzzelli, D.; Kalinitchev, A., and Tiravanti, G., Chloride/sulfate ion exchange kinetics on fibrous resins: Two independent models for film diffusion control, Ind. Eng. Chem. Res., 34(8), 2618-2624 (1995). Petruzzelli, D.; Passino, R., and Tiravanti, G., Ion exchange process for chromium removal and recovery from tannery wastes, Ind. Eng. Chem. Res., 34(8), 2612-2617 (1995). Przeszlakowski, S., Sorption of salicylic and 5-sulfosalicylic acids on silica gel impregnated with Aliquat-336, Sep. Sci. Technol., 30(15), 3073-3080 (1995). Pyrzynska, K., Solid phase extraction for preconcentration and separation of selenium species, Solvent Extr. Ion Exch., 13(2), 369-389 (1995). Rao, M.G., Binary and ternary ion-exchange equilibria in zeolitized tuff: Cs-Sr-Co system, Sep. Sci. Technol., 30(7), 1385-1396 (1995).

1038

Ray, A.K., and Carr, R.W., Experimental study of a laboratory-scale simulated countercurrent moving bed chromatographic reactor, Chem. Eng. Sci., 50(14), 2195-2202 (1995). Ray, A.K., and Carr, R.W., Numerical simulation of a simulated countercurrent moving bed chromatographic reactor, Chem. Eng. Sci., 50(19), 3033-3042 (1995). Romdhane, I.H.; Danner, R.P., and Duda, J.L., Influence of the glass transition on solute diffusion in polymers by inverse gas chromatography, Ind. Eng. Chem. Res., 34(8), 2833-2840 (1995). Row, K.H., and Choi, D.K., Asymmetry of peak shapes in linear gas-liquid chromatography, Sep. Sci. Technol., 30(19), 3615-3628 (1995). Sabharwal, K.N.; Vasudeva Rao, P.R., and Srinivasan, M., Extraction of uranium by macroporous bifunctional phosphinic acid resin, Solvent Extr. Ion Exch., 13(3), 561-574 (1995). Samanta, S.K., and Misra, B.M., Ion exchange selectivity of a resorcinol-formaldehyde polycondensate resin for cesium in relation to other alkali metal ions, Solvent Extr. Ion Exch., 13(3), 575-590 (1995). Schisla, D.K.; Carr, P.W., and Cussler, E.L., Hollow-fiber array affinity chromatography, Biotechnol. Prog., 11(6), 651-658 (1995). Schulte, L.D., et al., Cleanup of hydrochloric acid waste streams from actinide processes using extraction chromatography, Sep. Sci. Yechnol., 30(7), 1833-1848 (1995). Soldatov, V.S., Quantitative presentation of poteniometric titration curves of ion exchangers, Ind. Eng. Chem. Res., 34(8), 2605-2611 (1995). Spires, K., and Vincent, J.B., Transferrin metalloprotein affinity metal chromatography, J. Chem. Technol. Biotechnol., 62(4), 373-379 (1995). Streat, M., "The waters were made sweet": Advances in ion exchange technology, Ind. Eng. Chem. Res., 34(8), 2841-2848 (1995). Suwondo, E.; Floquet, P., and Domenech, S., Optimal policies of a liquid chromatography separation, Chem. Eng. Commun., 134, 73-86 (1995). Swami, N., and Dreisinger, D.B., Kinetics of zinc removal from cobalt electrolytes by ion exchange, Solvent Extr. Ion Exch., 13(6), 1037-1062 (1995). Tondeur, D., Paradigms and paradoxes in modeling adsorption and chromatographic separations, Ind. Eng. Chem. Res., 34(8), 2782-2788 (1995). van Deemter, J.J.; Zuiderweg, F.J., and Klinkenberg, A., Longitudinal diffusion and resistance to mass transfer as causes ofnonideality in chromatography, Chem. Eng. Sci., 50(24), 3867-3882 (1995). Velayudhan, A., and Horvath, C., Characterizing, and compensating for, enrichment effects in displacement chromatography with multicomponent carriers, Ind. Eng. Chem. Res., 34(8), 2789-2795 (1995). Velayudhan, A.; Hendrickson, R.L., and Ladisch, M.R., Simultaneous concentration and purification through gradient deformation chromatography, AIChE J., 41(5), 1184-1193 (1995). Vunnum, S., et al., Immobilized metal affinity chromatography: Modeling of nonlinear multicomponent equilibrium, Chem. Eng. Sci., 50(11), 1785-1804 (1995). Walton, H.F., Ligand-exchange chromatography: A brief review, Ind. Eng. Chem. Res., 34(8), 2553-2554 (1995). Watson, J.S., Simplified predictions of breakthrough fronts for constant-pattern adsorption and ion exchange, Sep. Sci. Technol., 30(7), 1351-1372 (1995). Wesselingh, J.A.; Vonk, P., and Kraaijeveld, G., Exploring the Maxwell-Stefan description of ion exchange, Chem. Eng. J., 57(2), 75-90 (1995). Yamazaki, H.; Inoue, Y., and Hirano, T., Studies of hydrous silicon-titanium(IV) oxide ion exchanger: The rate of isotopic exchange of Na + and Cs + ions between the exchanger in the relevant ionic form and aqueous solutions, Solvent Extr. Ion Exch., 13(1), 157-170 (1995). Yoshida, H., and Kishimoto, N., Adsorption of glutamic acid on weakly basic ion exchanger: Equilibria, Chem. Eng. Sci., 50(14), 2203-2210 (1995). Ysambertt, F., et al., Isolation and characterization of metalloporphyrins from a heavy crude oil by Soxhlet adsorption chromatography and HPLC-SEC, Sep. Sci. Technol., 30(12), 2539-2550 (1995). Zhao, D.; Sengupta, A.K., and Zhu, T., Trace contaminant sorption through polymeric ligand exchange, Ind. Eng. Chem. Res., 34(8), 2675-2684 (1995). Zheng, Z., et al., Estimation of cesium ion exchange distribution coefficients for concentrated electrolytic solutions when using crystalline silicotitanates, Ind. Eng. Chem. Res., 34(6), 2142-2147 (1995). Zhu, J., and Guiochon, G., Production rate of an isotachic train in displacement chromatography, AIChE J., 41(1), 45-57 (1995).

1039

1996 Aly, H.M., Kinetic studies on Na + and K+ exchange on cerium phosphate, Solvent Extr. Ion Exch., 14(1), 171177 (1996). Aly, H.M., The Al-13-phosphatoantimonic acid synthesis and ion exchange properties, Solvent. Extr. Ion Exch., 14(5), 94%954 (1996). Ashley, K.R., et al., Sorption behavior of perrhenate ion on Reillex-HP anion exchange resin from nitric acid and sodium nitrate/hydroxide solutions, Solvent Extr. Ion Exch., 14(2), 263-284 (1996). Asplund, S.E., and Edvinsson, R.K., A method for the rapid simulation of preparative liquid chromatography, Comput. Chem. Eng., 20(5), 507-516 (1996). Banakh, O.S., et al., Modified zeolites in gas chromatography for the analysis of air pollutants, Adsorpt. Sci. Technol., 14(4), 209-216 (1996). Bart, H.J., et al., Continuous chromagraphic separation of fructose, mannitol and sorbitol, Chem. Eng. Process., 35(6), 459-472 (1996). Basiuk, V.A., and Gromovoy, T.Y., Free energies of adsorption of amino acids, short linear peptides and 2,5piperazinediones on silica from water as estimated from high-performance liquid-chromatographic retention data, Adsorption, 2(2), 145-152 (1996). Bee-Gim, L., and Chi-Bun, C., Characterization of chiral adsorbents on the chromatographic separation of praziquantel enantiomers, Ind. Eng. Chem. Res., 35(1), 169-175 (1996). Bortun, A.I.; Bortun, L.N., and Clearfield, A., Ion exchange properties of a cesium ion selective titanosilicate, Solvent Extr. Ion Exch., 14(2), 341-354 (1996). Bowen, W.R., and Moran, E., Separation of amino acids at a synthetic ion exchange resin: Thermodynamics and energetics, Ind. Eng. Chem. Res., 35(2), 573-585 (1996). Brooks, C.A., and Cramer, S.M., Solute affinity in ion-exchange displacement chromatography, Chem. Eng. Sci., 51(15), 3847-3860 (1996). Buragohain, P.V.; Gill, W.N., and Cramer, S.M., Novel resin-based ultrapurification system for reprocessing IPA in the semiconductor industry, Ind. Eng. Chem. Res., 35(9), 3149-3154 (1996). Byers, C.H., and Williams, D.F., Efficient recovery of lanthanides by continuous ion exchange, Ind. Eng. Chem. Res., 35(4), 993-998 (1996). Caceres, A., et al., Analysis of photostabilizer in high density polyethylene by reverse- and normal-phase HPLC, Sep. Sci. Technol., 31 (16), 2287-2298 (1996). Carrere, H., et al., Whey proteins extraction by fluidized ion exchange chromatography: Simplified modeling and economical optimization, Chem. Eng. J., 64(3), 307-318 (1996). Chiarizia, R., et al., Uptake of metal ions by a new chelating ion exchange resin: Simultaneous uptake of cationic and anionic species, Solvent Extr. Ion Exch., 14(3), 519-542 (1996). Chiarizia, R., et al., Uptake of metal ions by a new chelating ion exchange resin: Silica grafted diphosphonic acid, Solvent Extr. Ion Exch., 14(6), 1077-1100 (1996). Colby, C.B., et al., Simulation of compression effects during scaleup of a commercial ion-exchange process, Biotechnol. Prog., 12(5), 662-681 (1996). Condo, P.D., et al., Partition coefficients and polymer-solute interaction parameters by inverse supercritical fluid chromatography, Ind. Eng. Chem. Res., 35(4), 1115-1123 (1996). de Lucas, A., et al., Ion exchange equilibrium of potassium on strong acid resins in polyol media, Solvent Extr. Ion Exch., 14(1), 141-160 (1996). de Lucas, A., et al., Ion exchange kinetics of DL-lysine monohydrochloride on Amberlite IRA-420, Solvent Extr. Ion Exch., 14(6), 1115-1136 (1996). DePaoli, S.M., and Perona, J.J., Model for Sr-Cs-Ca-Mg-Na ion-exchange uptake kinetics on Chabazite, AIChE J., 42(12), 3434-3441 (1996). Dumont, N., et al., Extraction of cesium from an alkaline leaching solution of spent catalysts using an ionexchange column, Sep. Sci. Technol., 31 (7), 1001-1010 (1996). Dunn, C.D., and Ghowsi, K., Hydrostatic flow injection and diffusional injection in reverse-direction micellar electrokinetic capillary chromatography: Theory and application, Sep. Sci. Technol., 31 (7), 993-1000 (1996). Felinger, A., and Guiochon, G., Optimizing experimental conditions in overloaded gradient elution chromatography, Biotechnol. Prog., 12(5), 638-644 (1996). Fernandez, E.J., et al., A column design for reducing viscous fingering in size exclusion chromatography, Biotechnol. Prog., 12(4), 480-487 (1996).

1040

Frey, D.D., Local-equilibrium behavior of retained pH and ionic strength gradients in preparative chromatography, Biotechnol. Prog., 12(1), 65-72 (1996). Gallant, S.R.; Vunnum, S., and Cramer, S.M., Modeling gradient elution of proteins in ion-exchange chromatography, AIChE J., 42(9), 2511-2520 (1996). Gandhi, N.N., et al., Immobilization ofMucor miehei lipase on ion exchange resins, Chem. Eng. J., 61(2), 149156 (1996). Hamaker, K.H., and Ladisch, M.R., Intraparticle flow and plate height effects in liquid chromatography stationary phases, Sep. Purif. Methods, 25(1), 47-83 (1996). Hamaker, K.H., et al., Chromatography for rapid buffer exchange and refolding of secretory leukocyte protease inhibitor, Biotechnol. Prog., 12(2), 184-189 (1996). Hameed, M.S., and Abid, M.F., Unsteady-state model for continuous moving-bed chromatography, Sep. Sci. Yechnol., 31 (15), 2065-2088 (1996). Hasnat, A., and Juvekar, V.A., Ion-exchange kinetics: Heterogeneous resin-phase model, AIChE J., 42(1), 161175 (1996). Hasnat, A., and Juvekar, V.A., Kinetics of ion exchange accompanied by neutralization reaction, AIChE J., 42(8), 2374-2376 (1996). Inoue, Y., and Abe, M., Synthetic inorganic ion exchange materials: Ion exchange mechanism of lithium ions on cubic tantalic acid, Solvent Extr. Ion Exch., 14(3), 507-518 (1996). Jansen, M.L., et al., Rigorous model for ion exchange equilibria of strong and weak electrolytes, AIChE J., 42(7), 1911-1924(1996). Jansen, M.L., et al., Effect of pH and concentration on column dynamics of weak electrolyte ion exchange, AIChE J., 42(7), 1925-1937 (1996). Juang, R.S., and Chou, T.C., Sorption of citric acid from aqueous solutions by macroporous resins containing a tertiary amine equilibria, Sep. Sci. Technol., 31 (10) 1409-1426 (1996). Juang, R.S., and Lee, S.H., Column sorption of divalent metals from sulfate solutions by extractant-impregnated macroporous resins, J. Chem. Technol. Biotechnol., 66(2), 153-159 (1996). Kaczmarski, K., Orthogonal collocation on finite elements with moving boundaries for simulation of nonlinear multicomponent chromatography: Influence of fluid velocity variation on retention time in LC and HPLC, Comput. Chem. Eng., 20(1), 49-64 (1996). Kaczmarski, K., and Antos, D., Fast finite difference method for solving multicomponent adsorptionchromatography mode Is, Comput. Chem. Eng., 20( 11 ), 1271 - 1276 (1996). Kasuga, F.; Yamazaki, H., and Inoue, Y., Studies of the hydrous niobium(V) oxide ion exchanger: Rate of the isotopic exchange of cesium ions between the exchanger in the Cs + form and aqueous solution, Solvent Extr. Ion Exch., 14(1), 161 - 170 (1996). Khamizov, R.K., et al., Recovery of pure magnesium compounds from seawater by the use of the effect of isothermal supersaturation in the ion-exchange process, Sep. Sci. Technol., 31(1), 1-20 (1996). Khan, S.H., and Rahman, K., Inversion of sucrose solution by ion exchange: Evaluation of reaction rate and diffusivity, Chem. Eng. J., 61 (1), 7-12 (1996). Kondoh, A.; Oi, T., and Hosoe, M., Fractionation of barium isotopes in cation-exchange chromatography, Sep. Sci. Technol., 31(1), 39-48 (1996). Kruglov, A.V.; Andreev, B.M., and Pojidaev, Y.E., Continuous isotope separation in systems with solid phase: Separation of nitrogen isotopes with use of ion-exchange resin, Sep. Sci. Technol., 31(4), 471-490 (1996). Kruglov, A.V.; Bjorklund, M.C., and Cart, R.W., Optimization of the simulated countercurrent moving-bed chromatographic reactor for the oxidative coupling of methane, Chem. Eng. Sci., 51(11), 2945-2950 (1996). Lim, B.G., and Ching, C.B., Modelling studies on the transient and steady-state behaviour of a simulated countercurrent chromatographic system, Sep. Yechnol., 6(1), 29-42 (1996). Lin, S.H., and Wu, C.L., Ammonia removal from aqueous solution by ion exchange, Ind. Eng. Chem. Res., 35(2), 553-558 (1996). Lin, W., and Hsieh, Y.L., Kinetics of metal ion absorption on ion-exchange and chelating fibers, Ind. Eng. Chem. Res., 35(10), 3817-3821 (1996). Ma, Z.; Whitley, R.D., and Wang, N.H.L., Pore and surface diffusion in multicomponent adsorption and liquid chromatography systems, AIChE J., 42(5), 1244-1262 (1996). Mardan, A., and Liljenzin, J.O., Unusual cation exchange equilibrium ofNa+/H + for a 65% crosslinked, strongly acidic cation exchange resin, Solvent Extr. Ion Exch., 14(4), 739-753 (1996).

1041

Mehablia, M.A.; Shallcross, D.C., and Stevens, G.W., Ternary and quaternary ion exchange equilibria, Solvent Extr. Ion Exch., 14(2), 309-322 (1996). Melis, S., et al., Ion-exchange equilibria of amino acids on a strong acid resin, Ind. Eng. Chem. Res., 35(6), 1912-1920(1996). Melis, S., et al., Separation between amino acids and inorganic ions through ion exchange: Development of a lumped model, Ind. Eng. Chem. Res., 35(10), 3629-3636 (1996). Norton, T.T., and Fernandez, E.J., Viscous fingering in size exclusion chromatography: Insights from numerical simulation, Ind. Eng. Chem. Res., 35(7), 2460-2468 (1996). Ooi, K., and Abe, M., Ion-exchange equilibria of alkaline earth metal ions/hydrogen ions on tin(IV) antimonate, Solvent Extr. Ion Exch., 14(6), 1137-1148 (1996). Pehlivan, E., et al., Ligand-exchange chromatography of aromatic amines on resin-bound cobalt ion, Sep. Sci. Technol., 31 (11 ) 1643-1648 (1996). Rendueles de la Vega, M.; Loureiro, J.M., and Rodrigues, A.E., Equivalence between Nernst-Planck and "corrected" Fick's law in modeling fixed-bed ion exchange processes, Chem. Eng. J., 61(2), 123-132 (1996). Ristic, M.D., and Rajakovic, L.V., Boron removal by anion exchangers impregnated with citric and tartaric acids, Sep. Sci. Technol., 31 (20), 2805-2814 (1996). Rodrigues, A.E.; Chenou, C., and de la Vega, M.R., Protein separation by liquid chromatography using permeable POROS Q/M particles, Chem. Eng. J., 61(3), 191-202 (1996). Rogers, R.D., et al., New technologies for metal ion separations: Aqueous biphasic extraction chromatography for uptake ofpertechnetate, Solvent. Extr. Ion Exch., 14(5), 919-946 (1996). Sabharwal, K.N., et al., Recovery of uranium from acid media by macroporous bifunctional phosphinic acid resin, Solvent Extr. Ion Exch., 14(6), 1101-1114 (1996). Schmidt-Traub, H., and Strube, J., Dynamic simulation of simulated-moving-bed chromatographic processes, Comput. Chem. Eng., 20 (Suppl. A), $641-$646 (1996). Simms, C.C.; Arumugam, B.K., and Wankat, P.C., Modified displacement chromatography cycles for gas systems, Chem. Eng. Sci., 51(5), 701-712 (1996). Singh, I.J., and Misra, B.M., Studies on sorption of radiocesium on copper-hexacyanoferrate-loaded resins, Sep. Sci. Technol., 31(12), 1695-1706 (1996). Soroush, M., and Kravaris, C., MPC formulation of GLC, AIChE J., 42(8), 2377-2381 (1996). Takeda, K., and Morita, K., Enrichment factor, height of separation unit, and separation efficiency by ion exchange with chemical reaction, Sep. Sci. Technol., 31(19), 2655-2670 (1996). Tan, H.K.S., Acid-salt separation by selective adsorption with ion-exchange resins, Sep. Sci. Technol., 31(16), 2209-2218 (1996). Tanaka, Y., Ion-exchange properties for Na + and K+ on a series of alpha-manganese dioxide ion exchangers, Solvent Extr. Ion Exch., 14(2), 323-340 (1996). Totura, G., Innovative uses of specialty ion exchange resins provide new cost-effective options for metals removal, Environ. Prog., 15(3), 208-212 (1996). Vailaya, A., and Horvath, C., Retention thermodynamics in hydrophobic interaction chromatography, Ind. Eng. Chem. Res., 35(9), 2964-2981 (1996). Van Lishout, Y.M.M., and Leighton, D.T., Absorption-induced separations in oscillatory liquid chromatography, AIChE J., 42(4), 940-952 (1996). Varadharaj, A., et al., Synthesis and performance characteristics of methylmethacrylate-divinylbenzene copolymer-based chelating resin for gallium metal recovery, J. Chem. Technol. Biotechnol., 67(2), 149-152 (1996). Velayudhan, A., and Horvath, C., Isotherm measurement by frontal chromatography in the presence of an adsorbing mobile phase modulator, Ind. Eng. Chem. Res., 35(4), 1173-1179 (1996). Vunnum, S.; Gallant, S., and Cramer, S., Immobilized metal affinity chromatography: Displacer characteristics of traditional mobile phase modifiers, Biotechnol. Prog., 12(1), 84-91 (1996). Wu, Y.Y.J., et al., Basic physical and chemical properties of Reillex-HPQ anion exchange resin and its sorption behavior ofhalides in aqueous nitric acid solution, Solvent Extr. Ion Exch., 14(2), 285-308 (1996). Zeng, X., and Murray, G.M., Synthesis and characterization of site-selective ion-exchange resins templated for lead(II) ion, Sep. Sci. Technol., 31 (17), 2403-2418 (1996). Zheng, Z., et al., Ion exchange of Group I metals by hydrous crystalline silicotitanates, Ind. Eng. Chem. Res., 35(11), 4246-4256 (1996).

1042

Zhong, G., and Guiochon, G., Analytical solution for the linear ideal model of simulated moving bed chromatography, Chem. Eng. Sci., 51 (18), 4307-4320 (1996). Zurer, P.S., Chromatography and mass spectrometry, Chem. Eng. News, 18 March, 38-46 (1996). 1997 Achuthan, P.V.; Janardanan, C., and Ramanujam, A., Water sorption isotherms and ionic hydration of uranium and thorium forms ofDowex 50W resins, Solvent Extr. Ion Exch., 15(4), 631-646 (1997). Anklam, M.R.; Prudhomme, R.K., and Finlayson, B.A., Ion exchange chromatography laboratory: Experimentation and numerical modeling, Chem. Eng. Educ., 31 (1), 26-31 (1997). Bartos, B.; Bilewicz, A., and Delmas, R., Synthesis and ion exchange properties of various forms of manganese dioxide for cations of the I and II groups, Solvent Extr. Ion Exch., 15(3), 533-546 (1997). Bhandari, V.M.; Juvekar, V.A., and Patwardhan, S.R., Ion-exchange studies in the removal of polybasic acids: Anomalous sorption behavior of phosphoric acid on weak base resins, Sep. Sci. Technol., 32(15), 2481-2496 (1997). Bortun, A.I.; Bortun, L.N., and Clearfield, A., A novel layered zirconium phosphate Zr203(HPO4): Synthesis and characterization of properties, Solvent Extr. Ion Exch., 15(2), 305-328 (1997). Bortun, A.I.; Bortun, L.N., and Clearfield, A., Synthesis and characterization of ion exchange properties of spherically granulated titanium phosphate, Solvent Extr. Ion Exch., 15(3), 515-532 (1997). Bortun, A.I.; Bortun, L.N., and Clearfield, A., Evaluation of synthetic inorganic ion exchangers for cesium and strontium removal from contaminated groundwater and wastewater, Solvent Extr. Ion Exch., 15(5), 909-929 (1997). Bortun, A.I.; Bortun, L.N., and Khainakov, S.A., Modified titanium phosphates as cesium selective ion exchangers, Solvent Extr. Ion Exch., 15(5), 895-907 (1997). Bricio, O.; Coca, J., and Sastre, H., Effect of the heterogeneity of macroporous styrene-DVB resins on ionexchange equilibria, Solvent Extr. Ion Exch., 15(4), 647-664 (1997). Chiarizia, R.; Horwitz, E.P., and Alexandratos, S.D., Diphonix(R) resin: A review of its properties and applications, Sep. Sci. Technol., 32(1), 1-35 (1997). Clearfield, A., et al., Synthesis and characterization of a novel layered sodium titanium silicate Na2TiSi2OT.2H_~O, Solvent Extr. Ion Exch., 15(2), 285-304 (1997). Cortina, J.L., and Miralles, N., Kinetic studies on heavy metal ions removal by impregnated resins containing di(2,4,4-trimethylpentyl) phosphinic acid, Solvent Extr. Ion Exch., 15(6), 1067-1083 (1997). Cumming, I.W.; Tai, H., and Beier, M., A model to predict the performance of an electrochemical ion exchange cell, Chem. Eng. Res. Des., 75(1 ) 9-13 (1997). Dave, S.M.; Patil, S.S., and Suresh, A.K., Ion exchange for product recovery in lactic acid fermentation, Sep. Sci. Yechnol., 32(7), 1273-1294 (1997). Defilippi, I.; Yates, S., and Sedath, R., Scale-up and testing of a novel ion exchanger for strontium, Sep. Sci. Technol., 32(1 ), 93-113 (1997). Delucas, A.; Canizares, P., and Rodriguez, J.F., Ion-exchange kinetics for the removal of potassium from crude polyols on strong acid resins, Sep. Sci. Technol., 32(11), 1805-1820 (1997). Dickson, M.L.; Norton, T.T., and Fernandez, E.J., Chemical imaging of multicomponent viscous fingering in chromatography, AIChE J., 43(2), 409-418 (1997). Ernest, M.V.; Bibler, J.P.; Whitley, R.D., and Wang, N.H.L., Development of a carousel ion-exchange process for removal of cesium-137 from alkaline nuclear waste, Ind. Eng. Chem. Res., 36(7), 2775-2788 (1997). Ernest, M.V., et al., Effects of mass action equilibria on fixed-bed multicomponent ion-exchange dynamics, Ind. Eng. Chem. Res., 36(1), 212-226 (1997). Farkas, T.; Sepaniak, M.J., and Guiochon, G., Radial distribution of the flow velocity, efficiency and concentration in a wide HPLC column, AIChE J., 43(8), 1964-1974 (1997). Farnan, D.; Frey, D.D., and Horvath, C., Intraparticle mass transfer in high-speed chromatography of proteins, Biotechnol. Prog., 13(4), 429-439 (1997). Frey, D.D., Mechanism for glutamic acid adsorption on a weak-base ion exchanger, Chem. Eng. Sci., 52(7), 1227-1231 (1997). Greenstein, E.M., Filters for resins, Chem. Eng. (N.Y.), June, 155 (1997). Gu, D.; Nguyen, L., and Philip, C.V., Cs+-ion exchange kinetics in complex electrolyte solutions using hydrous crystalline silicotitanates, Ind. Eng. Chem. Res., 36(12), 5377-5383 (1997). Habbaba, M.M., and Ulgen, K.O., Analysis of protein adsorption to ion exchangers in a finite bath, J. Chem. Technol. Biotechnol., 69(4), 405-414 (1997).

1043

Hasnat, A., and Juvekar, V.A., Dynamics of ion-exchange involving multivalent cations, Chem. Eng. Sci., 52(14), 2439-2442 (1997). Hasnat, A., and Juvekar, V.A., Ion exchange in weak acid resin: Diffusion in shrinking core, AIChE J., 43(10), 2605-2608 (1997). Jansen, M.L., et al., Effect of dissociation equilibria on ion-exchange processes of weak electrolytes, AIChE J., 43(1 ), 73-82 (1997). Juang, R.S., and Chen, M.L., Comparative equilibrium studies on the sorption of metal ions with macroporous resins containing a liquid ion-exchanger, Sep. Sci. Technol., 32(5), 1017-1035 (1997). Juang, R.S., and Ju, C.Y., Equilibrium sorption of copper(II)-ethylenediaminetetraacetic acid chelates onto crosslinked, polyaminated chitosan beads, Ind. Eng. Chem. Res., 36(12), 5403-5409 (1997). Jyo, A.; Yamabe, K., and Egawa, H., Metal ion selectivity of a macroreticular styrene-divinylbenzene copolymerbased methylenephosphonic acid resin, Sep. Sci. Technol., 32(6), 1099-1106 (1997). Kawamura, Y.; Yoshida, H., and Asai, S., Effects of chitosan concentration and precipitation bath concentration on the material properties of porous crosslinked chitosan beads, Sep. Sci. Technol., 32(12), 1959-1974 (1997). Kitakawa, A.; Yamanishi, Y., and Yonemoto, T., Complete separation of amino acids using continuous rotating annular ion exchange chromatography with partial recycle of effluent, Ind. Eng. Chem. Res., 36(9), 38093814(1997). Kuhr, J.H., et al., Ion exchange properties of a Western Kentucky low-rank coal, Energy Fuels, 11(2), 323-326 (1997). Kulikov, N.S., Molecular modelling in chromatostructural analysis: A new approach to the GC/MS study of isomers, Adsorpt. Sci. Yechnol., 15(2), 115-124 (1997). Lee, D.D.; Walker, J.F., and Taylor, P.A., Cesium-removal flow studies using ion-exchange, Environ. Prog., 16(4), 251-262 (1997). Lee, J.G.; Lee, W.C., and Wang, F.S., Simulation of pH elution in high-performance affinity chromatography using non-porous adsorbents, Chem. Eng. J., 65(3), 175-186 (1997). Lilga, M.A,; Orth, R.J., and Sukamto, J.P.H., Metal ion separations using electrically switched ion exchange, Sep. Purif. Yechnol., 11(3), 147-158 (1997). Lucas, A.D., et al., Potassium removal from water-methanol-polyol mixtures by ion exchange on Amberlite 252, Chem. Eng. J., 66(2), 137-148 (1997). Luo, R.G., and Hsu, J.T., Rate parameters and gradient correlations for gradient-elution chromatography, AIChE J., 43(2), 464-474 (1997). Luo, R.G., and Hsu, J.T., Optimization of gradient profiles in ion-exchange chromatography for protein purification, Ind. Eng. Chem. Res., 36(2), 444-450 (1997). Ma, Z., and Wang, N.H.L., Standing wave analysis of SMB chromatography: Linear systems, AIChE J., 43(10), 2488-2508 (1997). Mardan, A., Enrichment of boron-10 by inverse-frontal chromatography using quaternized 4-vinylpyridinedivinylbenzene and anion-exchange resin, Sep. Sci. Technol., 32(13), 2115-2125 (1997). Marquez, N.; Subero, N., and Anton, R.E., Effect of alkylate isomerism upon surfactant retention in an HPLC column and partitioning between water and oil, Sep. Sci. Technol., 32(6), 1087-1098 (1997). Matijasevic, L.; Vasic-Racki, D., and Pavlovic, N., Separation of glucose/fructose mixtures: Analysis of elution of profiles, Chem. Eng. J., 65(3), 209-212 (1997). McNulty, J.Y., The many faces of ion-exchange resins, Chem. Eng. (N.Y.), June, 94-100 (1997). Milan, Z., et al., Ammonia removal from anaerobically treated piggery manure by ion exchange in columns packed with homoionic zeolite, Chem. Eng. J., 66(1), 65-72 (1997). Miyabe, K., and Takeuchi, S., Surface diffusion phenomena in reversed-phase liquid chromatography with methanol/water and acetonitrile/water mixtures, Ind. Eng. Chem. Res., 36(10), 4335-4341 (1997). Miyoshi, H., Diffusion coefficients of ions through ion-exchange membranes for Donnan dialysis using ions of the same valence, Chem. Eng. Sci., 52(7), 1087-1096 (1997). Muralidharan, P.K., and Ching, C.B., Determination of multicomponent adsorption equilibria by liquid chromatography, Ind. Eng. Chem. Res., 36(2), 407-413 (1997). Nakayama, M., and Egawa, H., Recovery of gallium(III) from strongly alkaline media using a Kelex-100-loaded ion-exchange resin, Ind. Eng. Chem. Res., 36(10), 4365-4368 (1997). Nikolaev, N.P.; Muraviev, D.N., and Muhammed, M., Dual-temperature ion-exchange separation of copper and zinc by different techniques, Sep. Sci. Technol., 32(1), 849-866 (1997).

1044

Noriega, J.A.; Tejeda, A., and Magana, I. Modeling column regeneration effects on dye-ligand affinity chromatography, Biotechnol. Prog., 13(3), 296-300 (1997). Oi, T.; Shimazaki, H., and Ishii, R., Boron isotope fractionation in liquid chromatography with boron-specific resins as column packing material, Sep. Sci. Technol., 32(11), 1821-1834 (1997). Pais, L.S.; Loureiro, J.M., and Rodrigues, A.E., Separation of 1,1'-bi-2-naphthol enantiomers by continuous chromatography in simulated moving bed, Chem. Eng. Sci., 52(2), 245-258 (1997). Porter, C.E.; Riley, F.D., and Vandergrift, R.D., Fermium purification using Teva resin extraction chromatography, Sep. Sci. Technol., 32(1), 83-92 (1997). Prazeres, D.M.F., A theoretical analogy between multistage ultrafiltration and size-exclusion chromatography, Chem. Eng. Sci., 52(6), 953-960 (1997). Ramirez-Vick, J.E., and Garcia, A.A., Recent developments in the use of group-specific ligands for affinity bioseparations, Sep. Purif. Methods, 25(2), 85-130 (1997). Rastogi, R.K.; Mahajan, M.A., and Chaudhuri, N.K., Separation of thorium from uranium product at the tail end of thorium fuel reprocessing using macroporous cation-exchange resin, Sep. Sci. Technol., 32(10), 1711-1723 (1997). Rendueles, M.; Fernandez, A., and Diaz, M., Coupling of ion exchange with industrial processes: Application in fertilizer production and modeling of the key elution step, Solvent Extr. Ion Exch., 15(1), 143-168 (1997). Rendueles, M.; Fernandez, A., and Diaz, M., Sorption of counter and co-ions at high concentration in ion exchangers, Solvent Extr. Ion Exch., 15(4), 665-688 (1997). Rincon, J., et al., Selection of a cation exchange resin to produce lactic acid solutions from whey fermentation broths, Solvent Extr. Ion Exch., 15(2), 329-346 (1997). Robichaud, M.J.; Sathyagal, A.N., and Can', P.W., An improved oil emulsion synthesis method for large, porous zirconia particles for packed- or fluidized-bed protein chromatography, Sep. Sci. Technol., 32(15), 25472559 (1997). Roddick, F.A., and Britz, M.L., Production of hexanoic acid by free and immobilised cells of Megasphaera elsdenii: Influence of in-situ product removal using ion exchange resin, J. Chem. Technol. Biotechnol., 69(3), 383-391 (1997). Rogers, R.D.; Griffin, S.T., and Horwitz, E.P., Aqueous biphasic extraction chromatography (ABEC): Uptake of pertechnetate from simulated Hanford tank wastes, Solvent Extr. Ion Exch., 15(4), 547-562 (1997). Shalliker, R.A., et al., Examination of various pore size zirconias for potential chromatographic applications, Powder Yechnol., 91 ( 1), 17-24 (1997). Shelley, S., Ion exchange curb water and chemical use, Chem. Eng. (N.Y.), December, 117-118 (1997). Simon, G., et al., Preparative-scale separation of amino acids by using thermal ion exchange parametric pumping, Chem. Eng. Sci., 52(4), 467-480 (1997). Sujatha, V.; Sarma, C.B., and Raju, G.J.V.J., Studies on ionic mass transfer with coaxially placed helical tapes on a rod in forced convection flow, Chem. Eng. Process., 36(1), 67-74 (1997). Tanaka, Y., and Ysuji, M., Thermodynamic study of alkali metal ions/proton exchanges on an alpha-type manganese dioxide, Solvent Extr. Ion Exch., 15(4), 709-729 (1997). Tsaur, Y., and Shallcross, D.C., Comparison of simulated performance of fixed ion exchange beds in linear and radial flow, Solvent Extr. Ion Exch., 15(4), 689-708 (1997). Tsaur, Y., and Shallcross, D.C., Modeling of ion exchange performance in a fixed radial flow annular bed, Ind. Eng. Chem. Res., 36(6), 2359-2367 (1997). van Buel, M.J.; van der Wielen, L.A.M., and Luyben, C.C.A.M., Effluent concentration profiles in centrifugal partition chromatography, AIChE J., 43(3), 693-702 (1997). Various, New adsorbents and ion exchange materials (topic issue), Adsorption, 3(1), 5-105 (1997). Varotsis, N., and Pasadakis, N., Rapid quantitative determination of aromatic groups in lubricant oils using gel permeation chromatography, Ind. Eng. Chem. Res., 36(12), 5516-5519 (1997). Vlasenko, E.V.; Gavrilova, T.B., and Daidakova, I.V., Intermolecular interactions in gas chromatography on carbon black coated with monolayers of hydrocarbons with different electronic structures, Adsorpt. Sci. Yechnol., 15(2), 79-90 (1997). Warshawsky, A., et al., Solvent-impregnated resins via acid-base interaction of poly(4-vinylpyridine) resin and di(2-ethylhexyl) dithio-phosphoric acid, Solvent Extr. Ion Exch., 15(2), 259-284 (1997). Williams, C.J., and Edyvean, R.G.J., Ion exchange in nickel biosorption by seaweed materials, Biotechnol. Prog., 13(4), 424-428 (1997).

1045

Wolfgang, J.; Prior, A., and Byers, C.H., Continuous separation of carbohydrates by ion-exchange chromatography, Sep. Sci. Technol., 32(1 ), 71-82 (1997). Wu, D.J., et al., Recovery and purification of paclitaxel using low-pressure liquid chromatography, AIChE J., 43(1), 232-242 (1997). Yun, T.; Zhong, G.M., and Guiochon, G., Experimental study of the influence of the flow rates in SMB chromatography, AIChE J., 43(11), 2970-2983 (1997). Zagorodni, A.A.; Muiraviev, D.N., and Muhammed, M., The separation of Zn and Cu using chelating ion exchangers and temperature variations, Sep. Sci. Technol., 32(1), 413-429 (1997). Zheng, L.G., Some practical properties of sec-alkyl (C11-13) hydrogen styrylphosphonate, Solvent Extr. Ion Exch., 15(6), 1043-1049 (1997). Zheng, Z.; Anthony, R.G., and Miller, J.E., Modeling multicomponent ion exchange utilizing hydrous crystalline silicotitanates by a multiple interactive ion exchange site model, Ind. Eng. Chem. Res., 36(6), 2427-2434 (1997). Zhong, G., and Guiochon, G., Simulated moving bed chromatography: Effects of axial dispersion and mass transfer under linear conditions, Chem. Eng. Sci., 52(18), 3117-3132 (1997). Zhong, G., and Guiochon, G., Simulated moving bed chromatography: Comparison between the behaviors under linear and nonlinear conditions, Chem. Eng. Sci., 52(23), 4403-4418 (1997). Zhong, G.M.; Smith, M.S., and Guiochon, G., Effect of the flow rates in linear, ideal, simulated moving-bed chromatography, AIChE J., 43(11), 2960-2969 (1997). Zurer, P.S., Chromatography and mass spectrometry, Chem. Eng. News, 31 March, 42-47 (1997).

LIST OF J O U R N A L S S U R V E Y E D Abbreviation

Adsorption Journal Adsorption Science and Technology Advances in Chemical Engineering American Institute of Chemical Engineers Journal Ammonia Plant Safety Biotechnology Progress Canadian Journal of Chemical Engineering Catalysis Reviews in Science and Engineering Chemical Engineering (McGraw-Hill, New York) The Chemical Engineer (IChemE, UK) Chemical Engineering in Australia Chemical Engineering Communications Chemical Engineering Education Chemical Engineering Journal (including Biochemical Engineering Journal) Chemical and Engineering News Chemical Engineering and Processing Chemical Engineering Progress Chemical Engineering Research and Design Chemical Engineering Science Chemistry and Industry Chemtech Computers in Chemical Engineering Developments in Chemical Engineering and Mineral Proceesing Energy and Fuels Energy World Environmental Progress Food and Bioproducts Processing Fuel

Adsorption Adsorpt. Sci. Technol. Adv. Chem. Eng. AIChE J. Ammonia Plant Safety Biotechnol. Prog. Can. J. Chem. Eng. Catal. Rev. Sci. Eng. Chem. Eng. (N.Y.) Chem. Eng. (Rugby, Engl.) Chem. Eng. Aust. Chem. Eng. Commun. Chem. Eng. Educ. Chem. Eng. J. (Biochem. Eng. J.) Chem. Eng. News Chem. Eng. Process. Chem. Eng. Prog. Chem. Eng. Res. Des. Chem. Eng. Sci. Chem. Ind. (London) Chemtech Comput. Chem. Eng. Dev. Chem. Eng. Mineral Process. Energy Fuels Energy World Environ. Prog. Food Bioprod. Process. Fuel

1046

Gas Separation and Purification Hydrocarbon Processing Industrial and Engineering Chemistry Process Design and Development Industrial and Engineering Chemistry Research International Journal of Heat and Mass Transfer Journal of Applied Chemistry Journal of Chemical Technology and Biotechnology Journal of Food Engineering Journal of Institute of Energy Journal of Institute of Fuel Journal of Loss Prevention in the Process Industries Powder Handling and Processing Powder Technology Plant/Operations Progress Process Engineering Processing Process Safety and Environmental Protection Process Safety Progress Separation and Purification Methods Separation Science Separation Science and Technology Separation Technology Solvent Extraction and Ion Exchange Transactions of IChemE

Gas Sep. Purif. Hydrocarbon Process. Ind. Eng. Chem. Process Des. Dev. Ind. Eng. Chem. Res. Int. J. Heat Mass Transfer J. Appl. Chem. J. Chem. Technol. Biotechnol. J. Food Eng. J. Inst. Energy J. Inst. Fuel J. Loss Prev. Process Ind. Powder Handling Process. Powder Technol. Plant/Operations Prog. Process Eng. (London) Processing (Sutton, Engl.) Process Safety Environ. Prot. Process Safety Prog. Sep. Purif. Methods Sep. Sci. Sep. Sci. Technol. Sep. Technol. Solvent Extr. Ion Exch. Trans. IChemE

1047

Subject Index A Activated carbon for enrichment and separation of toxic metals, 771 modification for the analytical purposes, 772-783 Activated Carbon Fibres (ACF), 740 Adhesion soot particles to polymeric surface, 194-198 work, 178 Adsorbent activated carbons, 639,640, 808-810,817 activated carbons for ELCD, 821 et seq. BWC test, 822,824 GWC test, 822,824 SHED test, 822,824 production, 825-830 adsorption - active based on natural sorbents, 700-715 adsorption of NOx, 435 boehmite, 394 carbon materials for vapour pollutants, 397 et seq. activated carbons from lignocelulosic origin, 398,406-414 glassy carbons, 398, 407-415 carbon nanotubes, 639,640 dessicant materials, 839 elutrilithe, 382,383 for tobacco smoke, 852 GAC, 763 gibbsite, 394 natural, their structure and properties, 659-670

new composite, 381 et seq. selective and reversible for adsorption of NOx CuO - based, 444-449 mixed metal oxides, 442-444 sulfated metal oxides, 449-481 superconducting, 442 slag media, 533 zeolites, 245 et seq., 639,640 Adsorption and space craft applications carbon dioxide removal (Skylab), 460-463 future directions, 466,467 trace contaminant control, 458-460 water recovery, 463-466 CFCs in faujasite zeolite, 275-280 in polarographic/voltammetric environmental analysis, 111 et seq. from very diluted atmospheres by carbons differential heats, 419,422,423 under dynamic conditions, 417-430 in life support systems, 455 et seq. inside zeolites - molecular modelling, 246,259-265,273-275 in space environmental control, 455 et seq. microbial, 843 of cations on carbonates, 367-369 of cations on clay minerals, 362-367 of corrosion inhibitors, 867-89determining surface excess, 872-877

1048 differential capacity curves of EDL, 867-872 of exchangeable of ions on the metal oxides, 355-362 of metals to living or dead cells, 904 of microorganisms to surfaces, 842 of phosphorus by slag media, 533 et seq. of the toxic ions by activated carbons, 783- 797 of trace metal cations in soils, see also Sorption, 319 et seq. of vapour polutants by carbons, 397 et seq. nitric oxides, 435 et seq. from combustion gases, 435 et seq. on ion - exchanged zeolites, 437-441 on electrode, 112,113 on ideal activated carbon for ELCD test, 824 on layer deposited along optical fibre, 935-937 on powdered activated carbon, 726,727 on solid aerosol surfaces, 571 et seq. of organic compounds on inorganic oxides, 580-606 of organic pollutants on heterogeneous surfaces, 577-580 phenomena in electroanalysis, 111,112 sampling and sample preparation, 8-14 surfactants application in carbody experimental techniques, 178-180 on diesel engine soot, 177,178, 190-194 washing, 177 et seq. supercritical gases by nanospaces, 635 et seq.

Adsorption excess, 179 of surfactants at interface, 203 Adsorption heat pumps, 949 general idea, 949 operating principles, 950,951 prototypes and tests, 971-975 refrigerants and adsorbents, 951-953 thermodynamic cycles, 953-963 thermodynamic performances, 963-970 Adsorption isotherm DR, 643 for phenol compounds, 386-389 Frumkin, 867,877 Gibbs, 179, 873 on unwashed diesel engine soot, 188,189 virial, 867,884 Adsorption potential Lennard-Jones, 640,643 Steel, 640 Adsorption, see also Sorption phenomena in environmental systems, 285 Adsorptive Stripping Potentiometry (AdSP),112 Adsorptive Stripping Tensammetry (AdST),122 Adsorptive Stripping Voltammetry (AdSV),ll2 Air Purification (AP), 213 Air Revitalization System (ARS), 457 Airborne microbes, 834-836 Amino-Methyl Phosphonic Acid (AMPA), 724 Analysis of surfactants in the aquatic environment, 135 et seq. solved and unresolved questions, 135-168 in the aquatic environment initial steps, 137,138 ionic surfactants, 140-150 non-ionic surfactants, 150-162

1049 separation of surfactants and their metabolites, 138-140 Anodic Stripping Voltammetry (ASV),II3 Apollo, 463 Aquatic environment, 135 concentrations of surfactants, 149 Artificial kidney, 77 Artificial liver, 77 Atomic Absorption Spectroscopy (AAS), 798 Atomic Emission Spectroscopy (AES), 8OO

B Boyd-Adamson-Myers (BAM) theory, 746 Bed Capacity Factor (BCF), 216 Benzoic Acid(BA),82 Benzoylperoxide(BPO),82 Bibliography, on adsorption and adsorptive type separations for environmental protection, 979 et seq. Biological fluids, 77-85 Biomineralization, 904,905 Biosorption and detoxification of heavy metals, 904 and metal-bacteria interactions, 904-906 marine environmental, 904 Bismuth Active Substances (BIAS) method, 152 Blood Urea Nitrogen(BUN),83 Breakthrough curves, 545 for different beds, 848,853 for GAC beds, 729-733 for slag media, 545-547 Breakthrough moment, 733 Butane Working Capacity (BWC) test, 822

C Capillary Electrochromatography (CEC), 31 Capillary Electrophoresis (CE), 31, 83 Capillary Gel Electrophoresis(CGE),31 Capillary Isoelectric Foeusing(CIEF),31 Capillary Isotachophoresis(CITP),31 Capillary Zone Electrophoresis (CZE),31, 84 Carbody surfaces contact angle measurements, 201-204 wetting by adsorption surfactants, 198-203 Carbody washing adsorption of surfactants as condition, 208-210 Cation Exchange Capacity (CEC), 323 Cell surface layer (S-layer), 903 CFCs, 246 Chemical transducer, 926,940 Chemisorption and surface reactions on metaloxides, 606-626 on solid aerosol surfaces, 571 et seq. Chlorofluorocarbons (CFCs), 214, 949 Chornobyl nuclear power plant, 698 natural sorbents as decontamination agents, 698-700 Chromatography definitions and applications, 14-31 Clapeyron diagram, 954 Cobalt Thiocynate Active Substances (CTAS) method, 154 Coefficient Of Performance (COP), 963 Colloid science in the soil systems, 351 et seq. Combustion gas, 435,436 Complexation behaviour of organophosphorus polymer-supported reagents, 475-492 selective of metal ions, 473 et seq.

1050 Computational studies on the design of zeolite catalysts and novel adsorbents, 245 for pollution control, 246 et seq. Computer simulation methods, 246 Molecular Dynamics (MD) method, 246 Monte Carlo (MC) method, 246 Condensation Nucleus Counter (CNC), 846 Contaminant pesticides in the surface water, 723,724 Control of air pollution by adsorption, 807 adsorbents, 807 adsorption isotherms, 811,812 fixed bed probes, 812-816 of environment in space station, 455-458 Corrosion mechanism, 863 Corrosion inhibitors, 863-867 Critical Micelle Concentration (CMC), 181 Crystal growth the influence of bacteria in, 903 Crystal promotion, 906,907 Crystallization environmental applications, 903-905 physiological role, 905

D Detergency carbody surface, 204-207 Diesel- Exhaust Particles (DEP), 178 Diesel engine soot characterization, 185-189 Diffusion inside zeolites - molecular modelling, 246,259-265 Diffusion coefficient, 399

Dimethyl Methylphosphonate (DMMP), 213 Double Electrical Layer, see also Electrical Double Layer basis for electrochemical analysis, 111 Dual Mechanism Bifunctional Polymers (DMBPs), 481-484

E Electric Double Layer (EDL), see also Double Electrical Layer, 287, 288, 867 charge density, 886 diffuse charge density, 353 Gouy- Chapman (GC) model, 353 Gouy-Chapman-Stern-Grahame (GCSG) model, 288 Inner Helmholtz Plane (IHP), 355 MUSIC model, 354 soil/soil solution interface, 322 Stern layer model, 322 zeta potential, 354 Electroanalytical methods polarography/voltammetry, 111-113 Adsorptive Stripping Potentiometry (AdSP),112 Adsorptive Stripping Tensammetry (AdST), 122 Adsorptive Stripping Voltammetry (AdSV),112,113 Anodic Stripping Voltammetry (ASV),III-113 Electrochemical Detector (ECD),78 Electrode charge, 873 potential, 873 Elutrilithe, 381 chemical composition, 383 Energy Dispersive X-Ray Fluorescence (EDXRF), 794 Engine diesel, 435 gasoline, 435

1051 catalysts, 435,436 lean- burn, 436 reach- burn, 442 Enrichment (E), 216 Enthalpy of adsorption on glassy carbons, 428 Environmental analysis, 770,771 evaluation of the environmental profiles, 772-774 Environmental analysis, 78 adsorption phenomena, 3 et seq. adsorption in polarographic/ voltammetric, 111 et seq. Adsorption Stripping Voltammetry (AdSV) methods,112-125 chromatographic methods, 14-33 miscellaneous electroanalytical methods, 125-131 electrocapillary, 129-131 other methods, 131 tensammetry, 127-129 micellaneous methods, 31-33 sampling methods 4,8-14 steps, 3,14 Environmental Control and Life Support System (ECLSS), 456 design of, 456-458 Environmental media, 3 Environmental pollution diesel engines and DEP, 178 general problem, 571-574 Environmental Protection Agency (EPA), 463 Environmental Tobacco Smoke (ETS), 838 Equation Dubinin - Isotova (DI), 413 Dubinin - Radushkevich (DR), 401 Freundlich, 725 Kelvin, 195 Koryta, 113 Young, 203

Equivalent Background Concentration (EBC) model, 725 Evaporative Loss Control Devices (ELCD) filtres, 821 Extraction biovailability of metals, 341-344 metal leachability from the soil to phosphorous - based complexants, 474-492 the ground water, 344,345 selective sequential from soils, 331-335 solvent with soluble complexants, 473

F Fibre Optical Chemical Sensor (FOCS), 925 adsorption-based optical transduction in, 925 et seq. applications, 940-946 construction, 925,926 working principles, 930-934 Flame Atomic Absorption Spectroscopy

(FANS) FTIR, 780,783

G G immunoglobulins (IgG), 904 Gas Chromatography (GC),21-26 Gas-Liquid Chromatography(GLC),78 Gasoline Working Capacity (GWC), 822 equipment, 823 Gemini, 463 Global environmental problems, 214 emission control of greenhouse gases, 214 emission control of ozone depletion gases, 214 recovery of CFCs and VOC, 214 Glycosaminoglycans (GAGs), 905 Granular Activated Carbon (GAC), 724

1052 Graphite Furnace Atomic Absorption Spectrometer (GFAAS), 786 Greenhouse effect, 949 Groundwater treatment, 766

H Halobacteria, 903 Hamaker constant, 197, 375 Hanging Mercury Drop Electrode (HMD E), 155 Heterogeneous atmospheric organic chemistry, 572,627 Heterogeneous catalysts environment- friendly, 246 Friedel-Crafts, 246 hazardous, 246 molecular shape selectivity, 259 zeolites, 245 et seq. High Performance Liquid Chromatography(HPLC),15-21 Homogeneous Surface Diffusion Model (HSDM), 728 HPLC - Mass Spectrometry(MS),82 HPLC,78

Ideal Adsorbed Solution Theory (IAST), 725 In Situ Resource Utilization (ISRU), 466 Indirect Tensammetric Method (ITM),156 Indirect Tensammetric Technique (ITT),156 Indoor air pollutants, 833-836 Inductively Coupled Plasma Atomic Emission Spectrometer (ICPAES), 538 Inductively Coupled Plasma-Atomic Emission Spectroscopy(ICP-AES), 123 Inductively Coupled Plasma-Mass Spectroscopy(I CP-MS), 124 Infrared Spectroscopy (IRS), 442

International Space Station (ISS), 456-458 Inverse Gas Solid Chromatography (IGSC), 399 Ion exchange, 449-451 and sewages, 504-521 in soil materials, 320-328 new trends, 521-523 removal of metallic ions from water kinetics model in natural water system, 745 selecticity, 498-504 water recovery, 504-521 with crosslinked polymer beads, 473-492 Ion exchangers active carbon, 772 selective (tabulated data), 489-503 Ionic surfactants anionic surfactants, 140-146 cationic surfactants, 146-150

K KINEQL program, 285,294 Kinetics gas adsorption on activated carbons, 813 ionic solute adsorption, 745 main steps, 745 mathematical model, 748-755 of chemisorption and surface reactions, 610-612 of ion sorption processes, 299-306 factors influencing, 299,300 models, 300 KINEQL, 300-306 of phosphorus adsorption, 535-548 of physisorption on carbon materials, 399,400,407 of zinc ion reduction, 890

L Lambert - Beer law, 180, 930

1053 Langmuir-Blodgett films, 940, 941 Life Cycle Assessment (LCA), 763 backgrounds, 764 Linear Alkylobenzene Sulphonate (LAS),I39 Linear Driving Force (LDF), 730 Liquid-Liquid Extraction(LLE), 37, 84 Liquid-Solid Extraction(LSE),37 Local environmental problems, 214 removal SOx and NOx from flue gas, 214 solvent vapor fractionation, 214 SVR, 214 Low Molecular Weight Organic Acid (LMWOA), 342

Molecular modelling, see also Computation Studies adsorption and diffusion in zeolites, 245 in relevance to environmental multitechnique methods, 246, 247-266 protection, 245 et seq. Molecular shape selectivity, 259 Monitoring, 8-14 general problem, 3 monitored substances, 2-7 sampling methods, 8-14, 37-39 M-Toluidine (MT), 867 MUltiSIte Complexation (MUSIC) model, 354

M M immunoglobulins (IgM), 904 Man-made hydrocarbon emissions, 822 Mars atmosphere, 466 Marshall Space Flight Center (MSFC), 462 Mass Transfer Zone (MTZ), 812 Mercury, 463 Methyl Methacrylate(MMA),82 4,4' Methylenedianiline(MDA), 78 Methylene Blue Active Substance (MBAS) method, 140 Methylenediisocyanite(MDI),78 Micellar Electrokinetic Capillary Chromatography(MECC), 31 Micellar Electrokinetic Chromatography (MEKC),83 Micellar Electrokinetic Gas Chromatography (MEGC),84 Micropore filling mechanism, 644 MINEQL program, 291 Mixed adsorption layers formed by corrosion inhibitors, 863 et seq. Modelling of metal ion sorption phenomena, 285 et seq.

N N,N-Dimethyl P-Toluidine(DMPT),82 Nanoparticles or nanomaterials, 638-640 Nanopore Molecular Engineering, 654 Nanospace system, see also Nanoparticles, 635,638-640 adsorption in, 640-642 National Aeronautics and Space Administration (NASA), 457 Natural Organic Matter (NOM), 724 n-butanol (BU), 866 Non-ionic surfactants, 150-169 Nucleation and sodium chloride crystal growth, 903 et seq.

O Opthode adsorption based, 934,935,946 Optical fibre description, 926-930 Ozone depletion, 949 Ozone/UV, 763

1054

P

R

Parson's function, 873 Percolation HCH/chlorobenzenes contaminated groundwater, 763 water from landfill, 763 Pollutants anionic organic and inorganic, 381 applications of the adsorption phenomena for their analyses, 3 et seq. cationic organic and inorganic, 381 environmental, 5-7 anthropogenic, 5 natural, 5 secondary, 5 neutral organic and inorganic, 381 phenol compounds, 382,383 vapour, 400 et seq. Polycyclic Aromatic Nitrogen Heterocyclic (PANHs) compounds, 13 Polyethylene Glycol (PEG), 139, 904 Polymethyl Methacylate(PMMA),82 Polyurethane surface, 178 Polyurethane (PU),78 Pore Diffusion Model (PDM), 537,538 Potential zero charge (pzc), 867,872 Potentially Toxic Elements (PTE), 352 Pressure Swing Adsorption (PSA), 213 application for the environment, 213 et seq. Air Purification (AP), 213, 216, 219, 230-232 defence applications, 213 fundamental of environmental PSA processes, 214,215-218 mathematical modelling, 218-241, PSA-SVR, 219,220-230,232-241 Production of shape selective of alkylaromatics, 259-265 Pseudosurfactant, 144,145 P-Toluidine (PT), 867

Rapid Small Scale Column Test (RSSCT), 728 Raw material almond shells, 398 lignocellulase, 398 olive stones, 398 organic copolymers, 398 Recovery of CFCs and VOC, 214 of gasoline vapour, 821 of gold and platinum metals, 504-508 of humidity condensate, 455 of metal ions from aqueous solutions, 473 et seq. of organics from chemical processes and storage-tanks, 214 of proteins, 903 of silver, 508-509 of water, 463-466 Relative Standard Deviation (RSD), 801 Removal nitrogen oxides, 269 by selective catalytic reduction, 269-275 of contaminants in defense applications, 214 of disperse impurities from water, 670-682 of hazardous compounds in biological fluids, 77 et seq. of inorganic cations from water, 691-700 of microorganisms and particulates from indoor air, 833 adsorbent based removal systems, 839,840 control strategies, 838,839 fundamentals and mechanisms, 841-844 laboratory test, 844-861

1055 of CFCs, 246 of liquid soils from a surface, 177 of metallic ions from water and pollutants from wastewater, 381 et seq. of nitric oxides, 436 et seq. of oil and petroleum pollutants from water, 710-715 of organic molecules and ions form water, 682-691 of pesticides from the surface water, 733-737 of phosphorus by slag, 523 et seq. characteristics of the adsorbents, 538-540 from water and wastewater, 533 kinetic models, 535-548 Pure Diffusion Model (PDM), 536, 548,549 using soil and slag media, 534 of SOx and NOx from flue gases, 214 of trace gas-phase contaminations, 455,460-463 sewages, 504-521 chromium, 516 copper, 514 lead, 513 mercury, 510-513 nickel and vanadium, 515 tin and cobalt, 509 zinc, 519-521 Respirable Suspended Particulates (RSP), 833

Selective Catalytic Reduction (SCR), 435 nitric oxide with hydrocarbons, 435,436 Separation of surfactants and their metabolites, 138-140 gas stripping technique, 139 ion - exchange, 139

liquid/liquid extraction, 138 solid/phase extraction, 139 Sequential Quadratic Programming (SOP), 294 Sewage Treatment Plant (STP), 533 SHED emission test, 821 Skylab, 460,461 Slag media, 533 Soil adsorbing and complexing properties, 369-373 environmental general problems, 351-353 geochemical phases, 329-331 phase distribution of metals, 336-341 solids, 319-323 solution, 321,322 trace metal cation sorption in, 319 et seq. types, 320 Solid aerosol composition, 573-577 industrial, 571-573 Solid Phase Extraction (SPE), 11, 37 application to environmental analyses, 37,65-73 impact of various factors, 51-64 overview, 37-39 basic steps of SPE, 45-47 chemical characteristic of the sorbent, 41,42 physical characteristic of the sorbent, 39,40 sorbent selection, 43,44 SPE format, 44,45 Solvent Vapor Recovery (SVR), 213 Sorption of trace metal cations in soils, 319 et seq. phenomena in environmental systems, 285 et seq. Sorption equilibria of metal ions, 286-299

1056 factors influencing, 286-288 models, 288 SCFM, 288 SPE procedure for blood MDA,82 SPE,77 Super Critical Fluid Chromatography (SCFC),30,31 Super Fluid Critical Extraction (SFCE),77 Supercritical gas, 637 control with solid nanospaces, 635 et seq. practical importance, 635,654 physical properties, 637 Surface Active Substance (SAS), 111, 682 Surface charge, 771 Surface Complex Formation Model (SCFM), 285,288Surface Diffusion Model (SDM), 535 Surface pressure, 873 Surface tension critical of a carbody surface, 198,199 surfactant solutions, 179,182-185 Surfactant applications, 136,137 main types and mixtures, 136 Surfactant solution aquatic, 135 characterization, 180 Surfactants, 180,181 Synthesis of organophosphorus polymersupported reagents, 475-492 Synthesis of acyclic enones by solid acid catalysts, 246-253

T Temperature Programmed Desorption (TPD), 442 Temperature Swing Adsorption (TSA), 220 Tetrahydrofurane (THF), 399

Thin Layer Chromatography(TLC), 26-30 Thiourea (TU), 867 Toxic compounds to human health selective retention, removal and elution for analysis of,85-106 Trace analysis by atomic spectroscopy, 797-803 slurry sampling technique, 802 Transport of metal ions in natural systems, 285, 306-312 HYDROGEOCHEM model, 285, 308-312

U Ultraviolet Detector (UVD),78

V Vapor Compression Distillation (VCD), 463 Volatile Organic Compounds (VOC), 213, 416

W Wastewater treatment, 381 et seq., 728-742 advanced adsorption techniques, 763 advanced oxidation by means of ozone/UV, 763,7655,769 environmental analysis, 770,771 environmental profiles, 772-774 GAC on reactivation basis, 763,765,767 by adsorptive slag media, 533 by filtration, 737 by ion exchange, 473 in space craft, 463-466 ion exchange, 497, 745 natural sorbents, 659-700 adsorption- active based on, 700-715

1057 new composite adsorbents, 382-394 Water purification, 659 et seq. treatment, 745 Water production by activated carbon filtration, 723 et seq. GAC filtration, 737-742 general problems, 725-727 HSDM, 728,729 membrane/GAC filtration, 738-740 ozone-GAC - filtration, 741 Water Recovery System (WRS), 457

Zeolites model of cages and windows in zeolite Y, 253,254 pore dimensions and architecture, 249,250