3
Biotechniques for Air Pollution Control Proceedings of the 2nd International Congress on Biotechniques for Air Pollution Control, A Coruña, Spain, October 3-5, 2007
Christian Kennes and María C. Veiga (eds.)
A Coruña 2007
Universidade da Coruña Servizo de Publicacións
4
Biotechniques for Air Pollution Control Christian Kennes, María C. Veiga (eds.) A Coruña, 2007 Universidade da Coruña, Servizo de Publicacións Cursos_Congresos_Simposios, nº. 92 616 páxinas. 17 x 24 cm. Índice: páxinas 5-12 ISBN: 978-84-9749-258-4 Depósito legal: C 3157-2007 Materia: 504: Ciencias do medio ambiente. 66: Tecnoloxía química
Edición: Universidade da Coruña, Servizo de Publicacións, http://www.udc.es/publicaciones © Os autores © Universidade da Coruña Distribución: Galicia: CONSORCIO EDITORIAL GALEGO. Estrada da Estación 70-A, 36818, A Portela. Redondela (Pontevedra). Tel. 986 405 051. Fax: 986 404 935. Correo electrónico:
[email protected] España: BREOGÁN. C/ Lanuza, 11. 28022, Madrid. Tel. 91 725 90 72. Fax: 91 713 06 31. Correo electrónico:
[email protected]. Web: http://www.breogan.org Imprime: Lugami Artes Gráficas
Reservados todos os dereitos. Nin a totalidade nin parte deste libro pode reproducirse ou transmitirse por ningún procedemento electrónico ou mecánico, incluíndo fotocopia, gravación magnética ou calquera almacenamento de información e sistema de recuperación, sen o permiso previo e por escrito das persoas titulares do copyright.
5
Contents
ODOUR CONTROL IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE USING OLFACTORY GC-MS GAVIN PARCSI AND RICHARD M. STUETZ ...........................................................
15
A STATISTICAL PERSPECTIVE ON BIOFILTER PERFORMANCE IN RELATION TO THE MAIN PROCESS PARAMETERS AND CHARACTERISTICS OF UNTREATED FLOWS ANTON PHILIP VAN HARREVELD ........................................................................
29
REFURBISHMENT OF AN ODOUR COLLECTION AND BIOFILTER TREATMENT SYSTEM AT A MUNICIPAL SOLID WASTE COMPOSTING FACILITY IN PERTH, WESTERN AUSTRALIA TERRY J. SCHULZ AND STUART MCALL .............................................................
41
ODOUR AND H2S DEGRADATION IN A FULL SCALE BIOFILTER WITH A MINERAL BASED ORGANIC COATED FILTER MEDIA FRANZ-BERND FRECHEN, WOLFRAM FRANKE AND BJÖRN SCHOLL ......................
53
HIGH THROUGHPUT BIOFILTRATION FOR ODOUR CONTROL AT WATER PURIFICATION PLANT VITALY ZHUKOV, ANDREY VEPRITZKY, LEONID MITIN AND VLADIMIR POPOV ......
59
CONTINUOUS MONITORING OF ODOURS AT A BIOFILTER OUTLET SELENA SIRONI, LAURA CAPELLI, PAOLO CÉNTOLA, RENATO DEL ROSSO AND MASSIMILIANO IL GRANDE ................................................................................
67
REMOVAL OF ODOUR AND AMMONIA IN VENTILATION AIR FROM GROWING-FINISHING PIG UNITS USING VERTICAL BIOFILTERS ANDERS LEEGAARD RIIS ...................................................................................
77
MEMBRANE INLET MASS SPECTROMETRY (MIMS) AS A TOOL FOR EVALUATING BIOLOGICAL AIR FILTERS IN AGRICULTURE ANDERS FEILBERG ............................................................................................
85
6
MODELING AND TECHNOLOGICAL ASPECTS MODELING OF A FUNGAL BIOFILTER FOR THE ABATEMENT OF HYDROPHOBIC VOCs ALBERTO VERGARA-FERNÁNDEZ AND SERGIO REVAH .........................................
95
GASOLINE BIOFILTRATION: AN ANALYTIC MODEL ANDREW M. GERRARD, MARTIN HALECKY AND JAN PACA .................................
107
MODELING OF BIOMASS ACCUMULATION AND FILTER BED STRUCTURE CHANGE IN BIOFILTERS FOR GASEOUS TOLUENE REMOVAL JINYING XI, HONG-YING HU AND CAN WANG ...................................................
115
MODELLING THE ADSORPTION OF STYRENE AND ACETONE ON ACTIVATED CARBON AND PERLITE BEDS ANDREW M. GERRARD, SEBASTIAN MOLLENCAMP, KEHINDE MAKINDE, JAN PACA AND ONDREJ MISIACZEK ...................................................................................
125
CO-TREATMENT OF BENZENE AND TOLUENE VAPOURS IN A BIOFILTER: A FACTORIAL DESIGN APPROACH ELDON R. RENE AND T. SWAMINATHAN .............................................................
133
MATHEMATICAL MODELING AND SIMULATION OF VOLATILE REDUCED SULFUR COMPOUNDS OXIDATION IN BIOTRICKLING FILTERS G. AROCA, M. CÁCERES, S. PRADO, C. SÁNCHEZ AND R. SAN MARTÍN .............
143
ARTIFICIAL NEURAL NETWORK BASED MODEL FOR EVALUATING PERFORMANCE OF IMMOBILIZED CELL BIOFILTER ELDON R. RENE, JUNG HOON KIM AND HUNG SUCK PARK ................................
153
BIODESULPHURISATION AND BIOGAS TREATMENT DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL PROCESSES TO DESULPHURISE INDUSTRIAL GASSES ALBERT J.H. JANSSEN, ROBIN VAN LEERDAM, PIM VAN DEN BOSCH, ERIK VAN ZESSEN, GIJS VAN HEERINGEN AND CEES BUISMAN ...........................................
167
7
STUDY OF A DESULFURIZATION PROCESS TO CONVERT DIBENZOTHIOPHENE TO 2-HYDROXYBIPHENYL BY RHODOCOCCUS RHODOCHROUS NRRL (B-2149) A.B. SOARES JÚNIOR, Y. K. P. GURGEL, B.M.E. CHAGAS, T. B. DOMINGOS, G.R. MACEDO AND E. S. SANTOS .....................................................................
185
CONTROL OF METHANE EMISSIONS ISSUING FROM LANDFILLS: THE CANADIAN CASE JOSIANE NIKIEMA AND MICHÈLE HEITZ .............................................................
193
DESULFURISATION OF BIOGAS BY BIOFILTRATION DIANA RAMÍREZ-SÁENZ AND INÉS GARCÍA PEÑA ...............................................
201
AN INNOVATIVE BIOTRICKLING FILTER FOR H2S REMOVAL FROM BIOGAS LAURA BAILÓN ALLEGUE .................................................................................
215
REMOVAL OF INORGANIC POLLUTANTS REMOVAL OF AMMONIA BY IMMOBILIZED NITROSOMONAS EUROPAEA IN A BIOTRICKLING FILTER PACKED WITH POLYURETHANE FOAM MARTÍN RAMÍREZ, JOSÉ MANUEL GÓMEZ AND DOMINGO CANTERO ...................
227
STUDY OF NH3 REMOVAL BY GAS-PHASE BIOFILTRATION: EFFECTS OF SHOCK LOADS AND WATERING RATE ON BIOFILTER PERFORMANCE GUILLERMO BAQUERIZO, JUAN PEDRO MAESTRE, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE ........................................................................
239
HIGH H2S CONCENTRATIONS ABATEMENT IN A BIOTRICKLING FILTER: START-UP AT CONTROLLED pH AND EFFECT OF THE EBRT AND O2/H2S SUPPLY RATIO MARC FORTUNY, MARC A. DESHUSSES, XAVIER GAMISANS, CARLES CASAS, DAVID GABRIEL AND JAVIER LAFUENTE .............................................................
251
8
AMMONIA TRANSFORMATION IN A BIOTRICKLING AIR FILTER LARS PETER NIELSEN, MARIE LOUISE NIELSEN, MATHIAS ANDERSEN AND ANDERS M. NIELSEN .....................................................................................................
263
REMOVAL OF HYDROGEN SULFIDE USING UPFLOW AND DOWNFLOW BIOFILTERS WONGPUN LIMPASENI AND NATTAPOL RATTANAMUK ..........................................
273
PROPOSING A NEW BATCH METHOD FOR ASSESSMENT OF BIOLOGICAL ACTIVITY IN H2S DEGRADING BIOTRICKLING FILTERS L. OTEGI AND L. LARREA .................................................................................
283
EFFECT OF VARIOUS FACTORS TO AMMONIA BIODEGRADATION BY TWO STAGE BIOFILTRATION SYSTEM SILVIJA STRIKAUSKA, DZIDRA ZARI A, OLGA MUTERE, ULDIS VIESTURS AND ANDREJS .............................................................................................
293
MEMBRANE BIOREACTORS REMOVAL OF DIMETHYL SULFIDE IN A THERMOPHILIC MEMBRANE BIOREACTOR MUNKHTSETSEG LUVSANJAMBA, AMIT KUMAR AND HERMAN VAN LANGENHOVE ......................................................................
305
BIOLOGICAL WASTE GAS PURIFICATION USING MEMBRANES: OPPORTUNITIES AND CHALLENGES N.J.R. KRAAKMAN, N. VAN RAS, D. LLEWELLYN, D. STARMANS AND P. REBEYRE ......................................................................................................
313
TREATMENT OF THE CONFINED AIR OF A SPACECRAFT CABIN AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD AND VINCENT PLATEL ............
323
GAS-PHASE TOLUENE BIODEGRADATION BY BURKHOLDERIA VIETNAMIENSIS G4 IN A BIOFILM MEMBRANE REACTOR AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA AND HERMAN VAN LANGENHOVE ......................................................................
335
9
VOC REMOVAL IN CONVENTIONAL AND BIOTRICKLING FILTERS TREATMENT OF GAS PHASE STYRENE IN A BIOFILTER UNDER STEADY-STATE CONDITIONS ELDON R. RENE, MARÍA C. VEIGA AND CHRISTIAN KENNES ..............................
347
DEGRADATION OF SOLVENT MIXTURE VAPORS IN A BIOTRICKLING FILTER REACTOR: IMPACT OF HYDROPHILIC COMPONENTS LOADING AND LOADING RELEASE DYNAMIC JAN PACA, ONDREJ MISIACZEK, MARTIN HALECKY AND KIM JONES ...................
357
PERFORMANCE OF PEAT BIOFILTERS TREATING ETHYL ACETATE AND TOLUENE MIXTURES UNDER NON-STEADY-STATE CONDITIONS F.J. ÁLVAREZ-HORNOS, C. GABALDÓN, V. MARTÍNEZ-SORIA, P. MARZAL AND J.M. PENYA-ROJA ......................................................................................
367
CHARACTERIZATION OF A BIOTRICKLING FILTER TREATING METHANOL VAPOURS A. ÁVALOS RAMÍREZ, J.P. JONES AND M. HEITZ ...............................................
379
PERFORMANCE EVALUATION OF FUNGAL BIOFILTERS PACKED WITH PALL RING, LAVA ROCK, AND PERLITE FOR α-PINENE REMOVAL YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES ....................................
385
STYRENE DEGRADATION IN PERLITE BIOFILTER: THE OVERALL PERFORMANCE CHARACTERISTICS AND DYNAMIC RESPONSE M. HALECKY, J. PACA, A. M. GERRARD AND C.R. SOCCOL ..............................
397
BIODEGRADATION OF METHYL ETHYL KETONE AND METHYL ISOPROPYL KETONE IN A COMPOSITE BEAD BIOFILTER WU-CHUNG CHAN AND KANG-HONG PENG .......................................................
405
REMOVAL OF A MIXTURE OF OXYGENATED VOCs IN A BIOTRICKLING FILTER F. J. ÁLVAREZ-HORNOS, C. GABALDÓN, V. MARTÍNEZ-SORIA, P. MARZAL, J.M. PENYA-ROJA AND F. SEMPERE ............................................................................
415
10
EMERGING BIOREACTOR TECHNOLOGIES SOLID-LIQUID TWO-PHASE PARTITIONING BIOREACTORS FOR THE TREATMENT OF GAS-PHASE VOCs ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS AND NEAL G. BOUDREAU .......
427
MICROBIAL SOLVENT REGENERATION IN BIOTREATMENT OF AIR CONTAMINATED BY STYRENE ERIC DUMONT AND YVES ANDRÈS .....................................................................
439
LONG-TERM STABILITY OF PSEUDOMONAS PUTIDA CULTURES DURING THE OFF-GAS TREATMENT OF TOLUENE RAÚL MUÑOZ, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL AND SANTIAGO VILLAVERDE ..............................................................................
447
DEVELOPMENT OF A NOVEL BIOSCRUBBING PROCESS FOR COMPLETE TREATMENT OF NOX FROM FLUE GASES SANJEEV S.R. ARJUNAGI AND LIGY PHILIP .........................................................
457
DICHLOROMETHANE REMOVAL USING MIXED CULTURES IN A BIOFILTER AND A MODIFIED ROTATING BIOLOGICAL CONTACTOR– START UP STUDIES R. RAVI, LIGY PHILIP AND T. SWAMINATHAN .....................................................
469
BEHAVIOUR AND OPTIMIZATION OF A NOVEL MONOLITH BIOREACTOR FOR WASTE GAS TREATMENT YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES ....................................
477
BIODEGRADATION OF BTXS AND SUBSTRATE INTERACTIONS IN A BIOACTIVE FOAM REACTOR JIHYEON SONG AND SHONG-GYU SHIN .............................................................
487
CHARACTERIZATION AND PERFORMANCE EVALUATION OF A TWOPHASE PARTITIONING BIOREACTOR FOR VOLATILES ORGANIC COMPOUNDS TREATMENT IN OFF-GAS JEAN-MARC ALDRIC AND PHILIPPE THONART .....................................................
495
11
REMOVAL OF DICHLOROMETHANE FROM WASTE GASES USING A FIXED-BED BIOTRICKLING FILTER AND A CONTINUOUS STIRRED TANK BIOREACTOR LAURA BAILÓN, YOLANDA DOPICO, MARCELL NIKOLAUSZ, MATTHIAS KÄSTNER, MARÍA C. VEIGA AND CHRISTIAN KENNES .........................................................
505
MICROBIAL ECOLOGY AND BIOFILMS DEVELOPMENT OF A RELIABLE EXTRACTION METHOD FOR THE RECOVERY OF TOTAL GENOMIC DNA FROM WOODCHIP COLONIZING BIOFILM INVOLVED IN GAS BIOFILTRATION LÉA CABROL, LUC MALHAUTIER, JANICK ROCHER, FRANCK POLY, XAVIER LE ROUX, MARC JOVIC, ANNE-SOPHIE LEPEUPLE AND JEAN-LOUIS FANLO ..............
517
FT-IR CHARACTERIZATION OF BIOFILMS FORMED ON ENGINEERED BIOFILTRATION MEDIA TREATING VOLATILE ORGANIC EMISSIONS FOR THE FOREST PRODUCTS INDUSTRY KIM JONES, MILI KHILNANI, ANAND KARRE, SERGIO SANTOS AND JAN PACA ....
529
MONITORING AND CHARACTERIZATION OF BACTERIAL POPULATIONS OF TWO BIOLOGICAL AIR FILTERS DURING THE START UP PHASE M. JOVIC, L. CABROL, F. DUCRAY, R. GAGNEUX AND A. S. LEPEUPLE ..............
539
BIOFILTER RESPONSE TO BIOMASS REACTIVATION FOR VOC TREATMENT A. ELÍAS, A. BARONA, G. GALLASTEGI, M. LARRAÑAGA AND M. FERNÁNDEZ ...
551
A COMPARATIVE STUDY OF THE CHARACTERISTICS AND PHYSICAL BEHAVIOUR OF DIFFERENT PACKING MATERIALS COMMONLY USED IN BIOFILTRATION ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE ......................................................................................
561
SUITABILITY OF DUST AND BIOAEROSOLS FROM A PIG STABLE AS INOCULUM FOR BIOLOGICAL AIR FILTERS ANJA KRISTIANSEN, PER HALKJÆR NIELSEN AND JEPPE LUND NIELSEN ..............
573
12
PILOT-SCALE AND INDUSTRIAL APPLICATIONS BIOFILTRATION OF BITUMEN VAPOURS–OPERATIONAL ASPECTS MATTHIEU GIRARD, JEAN-LOUIS FANLO, NICOLAS TURGEON, GERARDO BUELNA AND PAUL LESSARD ..........................................................................................
587
COMPARISON OF THREE PILOT PLANTS FILLED WITH ORGANIC MATERIALS FOR THE TREATMENT OF AIR POLLUTANTS FROM A COMPOSTING PLANT SÉBASTIEN BASSIVIÈRE, FLORENCE DUCRAY AND CHRISTOPHE RENNER ..............
597
BIOFILTRATION SYSTEMS FOR THE TREATMENT OF WASTE GAS FROM INDUSTRIAL PLANTS IAN PHILLIPS ....................................................................................................
605
13
ODOUR CONTROL
This page intentionally left blank
15
Improving odour management and abatement performance using Olfactory GC-MS GAVIN PARCSI AND RICHARD M STUETZ Centre for Water and Waste Technology, School of Civil and Environmental Engineering, University of New South Wales, Sydney, NSW, 2033, Australia
ABSTRACT The measurement of odorous emissions is usually assessed either as odour concentrations (OC) by dilution olfactometry or by the chemical analysis of the odorous compounds such as hydrogen sulphide or the separation of complex gas mixture using analytical instrumentation such as gas chromatography. These techniques either provide information on the perceived effect of the emission (olfactory) or characterise the odours in terms of their chemical composition (analytical) but provide limited information on the relationship between odour impact and the chemical composition. The integration of chemical and olfactory techniques using olfactory-gas chromatography allows for the correlation of chemical and sensory measurements via the coupling of an olfactory port to a GC. The incorporation of mass spectrometry (GC-MS-O) enables individual odorants to be separated, identified and characterised according to their intensity and character. GC-MS-O analysis of emissions from poultry sheds has shown that samples vary in terms of their chemical compositions (i.e. different odorants profiles) as well as the different intensities measured and demonstrates the potential benefits that GC-MS-O analysis can offer in identifying key chemical markers for odour management in terms of odorant removal (i.e. receptor impact) and abatement loading due to chemical saturation.
1 INTRODUCTION Complaints due to odour annoyance have become a major issue for intensive livestock, waste management and wastewater treatment operators as the repeated release of unpleasant odours from these facilities can constitute a nuisance to a local population (Gostelow et al., 2003). This impact has become more significant with the expansion
15
16
GAVIN PARCSI AND RICHARD M. STUETZ
of suburbia and the associated rural encroachment, resulting in residential and commercial properties becoming in closer proximity to these facilities than in the past. Traditionally, odour management has been maintained by the use of buffer distances between industry and receptors or by the installation of odour abatement systems that either collect and dispersion the emission or treat the emission to acceptable level to limit receptor impact. Conventional odour abatement systems include chemical scrubbers, biofilters, bioscrubbers and biotrickling filters. Often these systems do not deliver the expected reduction in odour emissions and / or meet their original design specifications in terms of removal efficiency, resulting in the emission of odorous compounds to local receptors leading to odour complaints. The cause of these process failures is often due to inadequate characterisation of the emission source in terms of odour composition and mass loading. A secondary effect of inadequate odour composition information is the ineffective evaluation of odour control systems performance during its operation. The design and optimisation of odour management and abatement systems is based on an understanding of the emissions present in the facilities with background environmental conditions. Typical odours emitted from intensive livestock, waste management and wastewater treatment facilities usually consist of a wide range of odorants; the essential components being hydrogen sulphide (H2S), methanethiol, dimethyl sulfide, aldehydes and some ketones. Most odour abatement designs are based on the use of one or two key odorants such as H2S, reduced sulphur compounds and / or VOC to determine the loading capacity for the system. This approach often doesn’t adequately account for the actual composition and individual concentrations that vary over time and rank the emission differences in terms of odorant removal (i.e. receptor impact) and abatement loading due to chemical saturation. The measurement of odours can either be assessed as odour concentration units (OU) by dilution olfactometry (using the CEN or equivalent national standard for dilution olfactometry) or analytical techniques such as the use of surrogates chemical markers (like H2S) or the chemical analysis of odorous mixtures by chromatographic techniques such as gas chromatography coupled with mass spectrometry (GC-MS) for quantification of individual compounds (Gostelow et al., 2001). Sensory measurements employ human panels (Figure 1) to characterise the odours in terms of their perceived effect but give no information regarding composition, whereas analytical measurements characterise odours in terms of their chemical composition but give little information as to their sensory impact. Current chemical methods for odour monitoring can include field sampling and laboratory analysis (Figure 2) of gaseous emissions such as H2S, volatile organic carbon (VOC), and ammonia measurements and the continuous in-situ monitoring of H2S, volatile organic carbon (VOC), and ammonia.
IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE
Figure 1. Olfactory analysis of odour samples.
Figure 2. Field and continuous monitoring of H2S.
17
18
GAVIN PARCSI AND RICHARD M. STUETZ
More recently the integration of chemical and olfactory techniques has been applied to odour analysis to allow the correlation of chemical and sensory measurements via the coupling of an olfactory port to gas chromatograph-mass spectroscopy (GCMS-O). GC-MS-O (Figure 3) allows individual odorants to be separated and identified individually as well as allowing the odour contribution for each compound to be characterised. The olfactory detection port (ODP) consists of a nose cone where panellists perceive the separated odorous compounds by continuously sniffing the GC column effluent and characterises it in terms of intensity and an odour description. The end of GC column is split into two streams via a column splitter (Figure 4) that directs column effluent to the MS and ODP via heated transfer lines.
Figure 3. Olfactory-GC-MS showing odour detection port (ODP) on right.
Olfactory-GC and Olfactory-GC-MS is well established in other science fields such as food aroma’s and taste and odours in drinking water but has limited application to environmental odour analysis until recently. In drinking water taste and odours (or off-flavours) monitoring GC-MS-O analysis has been successfully applied to the characterisation of common off-flavours such as geosmin and MIB (Hochereau and Bruchet, 2004) and has been used to produce odour wheels (Figure 5), which relate the odour descriptors to the chemical composition of odorants (Suffet et al., 1999).
IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE
Figure 4. Column splitter directing column effluent to the MS and ODP.
Figure 5. Example of odour wheels for off-flavours in drinking water (Suffet et al., 1999).
19
20
GAVIN PARCSI AND RICHARD M. STUETZ
GC-MS-O applications for the assessment of environmental odours has mainly focused on characterising changes in composition of odorous emissions from various agricultural and waste management operations such as swine finishing and poultry sheds and dairy facilities. Studies (Kai and Schäfer, 2004; Wright et al., 2005; Parcsi et al., 2007) have shown that emissions from different intensive livestock operations comprise different chemicals and odorants and that some species that gave an olfactometry response did not always correspond to a response from any other detector, conversely some compounds with large detector responses gave little or no olfactometry response. Additionally speculation is often made as to the identity of the compound based upon it odour characteristic and associated compounds within the matrix. This paper will describe the application of using olfactory-GC-MS for the characterisation of non-methane volatile organic compounds (NMVOC) emissions from tunnel ventilated broiler sheds in Australia and discuss how this technique can be more broadly applied to improve the design and optimisation of odour abatement performance through improved understanding of variations in the composition of odorous emissions in terms of receptor impact (i.e. different odorant profiles) and chemical loading on odour abatement systems.
2 MATERIALS AND METHODS The results that are presented here focus on odorous samples from two tunnel ventilated broiler sheds in Queensland and Victoria, Australia. Samples were collected on sorbent tubes containing either a Tenax TA sorbent (for n-C7 to n-C30 compounds) or a Carbotrap 300 sorbent (a blend of Carbopack C, Carbopack B and Carbosieve SIII for ethane to n-C20) (Markes International, UK), using calibrated sampling pumps. The sample volumes were recorded for each tube to allow for relative quantification. The use of different sorbents ensures that the compounds identified in subsequent analysis accurately represent the suite of compounds that are being emitted from the poultry sheds. The analytes were thermally desorbed from the sorbents and refocused within the cold trap of the thermal desorber (Markes Unity, Markes International, UK). Sample analysis was performed using a GC-MS (Agilent 6890N GC, 5973NMSD, Agilent Technologies) coupled to an Olfactory Detection Port (ODP2 Gerstel GmbH & Co., Germany) (Figure 3). The compounds were identified using gas chromatographic separation and mass selective detection with a HP-5MS capillary column (30m x 0.25mm x 0.25μm Film Thickness, Agilent Technologies). The flow rate of the gas chromatograph was maintained at a constant pressure using helium as the carrier gas. The oven was temperature programmed for a total run time of 44.00min, (50°C for 2 min, 5.00°C/min to 250°C hold for 2 min) this provided adequate separation of the eluting compounds. The mass selective detector was operating in continuous
IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE
21
scan mode (50 – 550 m/z) for GC-MS only analysis. The mass spectra were recorded using the Agilent ChemStation software and analysed offline using the Enhanced Data Analysis package (Agilent Technologies). The identification of the volatile organic compounds relied upon the matching of the acquired mass spectra with the ChemStation data bases (NIST02 and Wiley275). Identification of the compounds present within the matrix yielded a large number of different classes of compounds including aromatics, sulphur containing organic species, nitrogen containing species, aldehydes, ketones, alcohols, terpines and other general hydrocarbons. GC-MS-O analysis involved splitting the gas-chromatograph effluent between the mass selective detector and an Olfactory Detection Port. The scan range of the mass selective detector was increased at this stage to provide a more reliable match to the spectral databases (35 – 550 m/z). The mass spectra were recorded using the Agilent ChemStation software and the odour chromatograms were recorded using the Gerstel ODP Recorder software. Analysis was performed offline using the Agilent ChemStation Data Analysis software. To optimise the use of the panellist as an odour detector the split between the MSD and ODP was initially set at 1:1, before being refined to 2:3 (MSD:ODP), these split ratios were calculated using the Gerstal Column Calculator (Gerstel GmbH & Co., Germany.) These calculations were based on a column flow of 1.6mL.min-1 for the carrier gas Helium with an initial temperature of 50°C with the flow programmed to be constant flow as the temperature increases. In addition to the collection and analysis of NMVOCs, odour bags were collected onsite and analysed at local laboratories (as determined by dynamic dilution olfactometry as per CEN standards), this allows for the comparison to be drawn between the NMVOC emissions and the odour concentrations.
3 RESULTS A range of odour samples were collected during four sampling programs from two tunnel ventilated broiler sheds in Queensland and Victoria, Australia in order to characterisation of NMVOC emissions over the chicken growing out cycle (typically 9 weeks). 3.1 GC-MS ANALYSIS GC-MS analysis revealed that there was a marked variation in not only the abundance of species that were present during the grow-out cycle, but also the species that were present varied throughout the cycle. Figure 6 shows two typical total ion chromatograms (TIC’s) from one of the sampling locations. Both samples were collected under identical conditions, on the same day, from the same duty fan on the same shed at the same ventilation rate. The only difference was the sample volume,
22
GAVIN PARCSI AND RICHARD M. STUETZ
the Carbotrap300 was 2.91L and the Tenax TA was 3L. The compounds labelled are A – 1-butanol, B – dimethyl disulphide, C – toluene, D – styrene, E – N-butyl-1butanamine, F – 4-ethyl-decane, G – butylated hydroxytoluene (BHT). Table 1 shows a list of predominant NMVOC compounds that were isolated and identified within the matrix of the exhaust emissions from the poultry sheds.
Figure 6. GC-MS analysis of sorbent tubes: Carbotrap300 (top spectra) and Tenax TA (lower spectra). (A – 1-butanol, B – dimethyl disulphide, C – toluene, D – styrene, E – N-butyl1-butanamine, F – 4-ethyl-decane, G – butylated hydroxytoluene).
IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE
23
3.2 GC-MS-O ANALYSIS GC-MS-O analysis allows the simultaneous collection of olfactory and mass spectral data from GC analysis. Figure 7 shows a typical total ion chromatogram with the odour chromatogram overlayed to identify the odorants within the matrix. The results shows that only a small number of the compounds present are identified by the operator as odorous, and therefore could be potentially responsible for the odorous emissions from the poultry shed samples. Figure 7 also shows that the intensity of odorous compounds can be scaled from 0-3 thereby identifying the most odorous compounds and the one’s that are more likely to cause offensive to local receptors. Table 2 lists the NMVOCs that were isolated and identified by the ODP operator as being odorous. The most predominant odorants in the poultry emission matrix was determined to be dimethyl disulphide and 2, 3-butanedione (diacetyl). The ODP operator can also include voice activated odour descriptors to describe the character of odorants (Figure 8). 3.3 VARIATIONS IN ODORANT PROFILES The correlation of dominant odorants from the poultry shed emissions (Table 2) with the results of dilution olfactometry has shown that odour emission trends can be strongly linked to the abundance of these specific compounds. Figure 9 illustrates the relationship over the grow-out cycle between the abundance of dimethyl disulphide as acquired by mass spectral data and odour concentrations (determined by dilution olfactometry). The results have been normalised to the volume of air that was being exhausted from the shed at the time of sampling and shows that the variations in odour and NMVOC emissions can be linked to the either the bird age or bird mass. Figure 10 supports these observations and shows that the emissions of two key odorants (dimethyl disulphide and 2, 3-butanedione) are also subject to diurnal variations which is most likely the result of bird activities within the shed over the 24 hours due to feeding and lighting cycles.
24
GAVIN PARCSI AND RICHARD M. STUETZ
Table 1 Non-methane volatile organic compounds identified using GC-MS. Compound Family Aromatics
Sulphur
Aldehydes
Ketones
Nitrogen Alcohols Carboxylic Acids Terpines
Other Hydrocarbons
Compounds Isolated Toluene o-Xylene p-Xylene Benzene 1-ethyl-4-methyl-benzene 1-ethyl-2-methyl-benzene Acetophenone Benzaladehyde Phenol Styrene Dimethyl Sulphide Dimethyl Disulphide Dimethyl Trisulphide Butanal 3-methyl-butanal Cyclohexanal Hexanal 2-ethyl-1-hexanal 2-butanone Diacetyl 3-methyl-2-butanone 3-hydroxy-2-butanone Trimethylamine 1-butanol Cyclohexanol Acetic Acid α-pinene β-pinene Limonene Camphene Camphor Carene Eucolyptol Tetradecane Hexadecane Tetrahydrofuran
IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE
25
Table 2 Odorants identified using olfactory detection port (from Figure 7). Compound family Sulphur Ketones
Compound Dimethyl Disulphide Dimethyl Trisulphide 2,3-butanedione (diacetyl) 2-butanone Acetophenone 3-hydroxy-2-butanone
Odour Threshold Value (ppb)1 0.16 – 12 0.005 – 0.10 2.3 – 6.5 50,000 65 800
Figure 7. GC-MS-O analysis showing total ion chromatogram and odour chromatogram (A – 2-butanone, B – 2, 3-butanedione, C – dimethyl disulphide D – 3-hydroxy-2-butanone E – dimethyl trisulphide and F – acetophenone).
1
Odour Detection Values reported by Leffingwell & Associates http://www.leffingwell.com/ odorthre.htm
26
GAVIN PARCSI AND RICHARD M. STUETZ
Figure 8. GC-MS-O analysis showing the additions of odour descriptors on the odour chromatogram.
Figure 9. Variations of odour and dimethyl disulphide at different stages of a typical chicken grow-out cycle.
IMPROVING ODOUR MANAGEMENT AND ABATEMENT PERFORMANCE
27
Figure 10. Diurnal variations of odour and two key odorants.
4 CONCLUSIONS The GC-MS analysis of samples from different poultry sheds revealed that that there is a complex matrix of non-methane volatile organic compounds that form the emissions from these facilities. The simultaneous collection of olfactory and mass spectral data via GC-MS-O analysis demonstrated that only a small number of the NMVOC’s present in the matrix are responsible for the resulting odorous emissions. Olfactory-GC-MS analysis was able to identify the key odorants in the poultry emissions samples as dimethyl disulphide and 2, 3-butanedione. These compounds were determined to be the most odorous over the chicken grow-out cycle and showed that distinct odorant profiles occur due the different growth stage during poultry shed production (i.e. the age of the bird or the total mass of birds within the shed). The GC-MS-O analysis also showed that diurnal variations in odorants compositions where also influenced by chicken activity within the poultry sheds. As odour abatement process failure is often due to inadequate characterisation of the emission source in terms of odour composition. The application of olfactoryGC-MS analysis offer a potential approach to identify key odorous markers from different emission sources as demonstrated with the analysis of poultry shed emissions.
28
GAVIN PARCSI AND RICHARD M. STUETZ
The ability to identify compounds that have greater receptor impact will enable improved design of odour abatement systems to remove specific odorous compounds. Improved characterisation of odorous emissions will also enable more effective evaluation of odour control systems performance during its operation.
5 ACKNOWLEDGEMENTS The authors acknowledge the financial support of the Australian Poultry CRC (Project 04-45 – Dust and Odour Emissions from Poultry Sheds) and thanks the Queensland Department of Primary Industries and Fisheries (Erin Gallagher, Neale Hudson, JaeHo Sohn and Mark Dunlop), the Victorian Department of Primary Industries (Maurie Miles) and UNSW (Xinguang Wang and Gautam Chattopadhyay) for their assistance in VOC sampling and analysis. Gavin Parcsi was financially supported through a PhD scholarship from the Australian Poultry CRC.
REFERENCES Gostelow, P., Longhurst P.J., Parsons, S.A. and Stuetz, R.M. (2003) Sampling for Measurement of Odours. IWA Scientific and Technical Report No.17, IWA Publishing, London. Gostelow, P., Parsons, S.A. and Stuetz, R.M. (2001) Odour measrement in sewate treatment – a review. Water Res. 35(3): 579-597. Hochereau, C. and Bruchet, A. (2004) Design and application of a GC-Sniff/MS system for solving taste and odour episodes in drinking water. Water Sci. Technol. 49(9): 81-87 Kai, P. and Schäfer, A. (2004) Identification of key odour components in Pig House Air using hyphenated gas chromatography olfactometry. Agricultural Engineering International: the CIGR Journal of Scientific Research and Development. VI(04 006). Suffet, I.H., Khiari, D. and Bruchet, A. (1999) The drinking water and odour wheel for the millennium: beyond geosmin and 2-methylisoborneol. Wat. Sci. Technol. 40(6): 1-13. Wright, D.W., eaton, D.K., Nielsen, L.T., Kuhrt, F.W., Koziel, J.A., Spinhirne, J.P. and Parker, D.B. (2005) Multidimensional gas chromatography-olfactometry for the identification and prioritization of malodors from confined animal feeding operations. J. Agric. Food Chem. 53: 8663-8672.
29
A statistical perspective on biofilter performance in relation to the main process parameters and characteristics of untreated flows ANTON PHILIP VAN HARREVELD Odournet S.L., Ctra. de l’Esglèsia 60 bis, Barcelona 08017, Spain
ABSTRACT A large number of olfactometric measurements of odour removal efficiency of municipal waste organic fraction and green waste composting installations were compiled and analysed graphically. The number of measurements and installations is >50 for treated gas characteristics and >15 for filter media characteristics. All installations concerned were located in the Netherlands and Belgium. All untreated and treated gas odour concentrations were measured in duplicate or triplicate, according to olfactometry standard EN13725 or its predecessor NVN2820. The data were then analysed using graphical methods to identify trends and relation between effectiveness of performance and a large number of operational parameters and characteristics, including: • Area flow loading (m3·m-2 filter area·hour-1) • Contact time • Temperature of untreated flow • Odour concentration of untreated flow • Ammonia concentration of untreated flow • Dry matter content of filter media • pH of filter media
1 INTRODUCTION Biofilters are widely used to treat odorous process flows from composting installations used for processing the organic fraction of municipal waste and green wastes (separately collected vegetable, fruit and garden waste). The main objective of treatment is reduction of odour emissions and their related impacts. The criteria for satisfactory treatment result vary from country to country. For example:
29
30
ANTON PHILIP VAN HARREVELD
• The Netherlands Emissions Guideline (Infomil, 2000) considers an odour concentration in the treated gas of < 2500 ouE·m-3 indicative for an efficient biofilter. • In Germany (30.BimSchV, 2001) a biofilter is expected to produce treated gas flows with a concentration of < 500 ouE·m-3. • In Spain Odournet consultants frequently see environmental license conditions and contract conditions for waste management installations that require the concentration from biofilters not to exceed 1000 ouE·m-3. These widely differing performance criteria beg the question: What is a consistently attainable endpoint for treatment of organic fraction and green waste composting process flows in biofilters? This question leads to the related question: Which are the parameters that determine the effectiveness of treatment of a biofilter? A large number of supposedly relevant parameters have been identified in literature (IPPC BREF, 2003). Some of these factors relate to the design parameters of the biofilter, and others to the characteristics of the raw gas to be treated. 1. Biofilter design parameters: a. Area flow loading. The volume flow of gas treated by 1 square meter of biofilter surface, typically expressed as m 3·m-2·hour -1. This design parameter may vary over a range of 50 to 600 m3·m-2·hour-1. In practical applications for composting installations a range of 100 to 200 m3·m-2· hour-1 is more typical. b. Time of residence. The time that the gases remain in contact with the filter medium is a function of the area flow loading and the filling height of the media material. The parameter is expressed in seconds. As the actual biological oxidation occurs in the liquid adhered to the media, the time permitted for transfer of odorous components from the gas to the liquid phase is of obvious importance. Typically, the gas retention time in the biofilter media is between 30-60 seconds. c. Temperature of filter media. The temperature of the filter media must be suitable for the biomass to thrive. At low temperatures, the metabolism will occur at a slower rate. At higher temperature, above the mesophile range of 10 to 45ºC, the population will shift from mesophile to thermophile species, while at even higher temperatures biomass will not be able to exist. Typically, the optimum temperature in a biofilter is considered to be within the range of 30 to 38ºC.
A STATISTICAL PERSPECTIVE ON BIOFILTER PERFORMANCE
31
2. Biofilter media characteristics: a. Type of filter media. There is a wide range of filter media, most of organic origin, e.g. peat, heather, wood chips, root chips, bark chips, compost, etc. Inorganic media are used in some cases, e.g. extruded clay pellets, lava rock, typically mixed with some organic material. The media provides for the basic life support of the biomass, and can contribute by providing: physical support surface, moisture, nutrient supply and pH control. b. Nutrient availability. The biomass should satisfy part of its nutrient requirements by using the components provided by the raw gas flow. However, these may not provide a balanced diet and deficiencies in nutrients such as phosphorus, nitrogen and trace elements may limit biomass development. In that case nutrient availability must be supplemented from the media of by irrigation with a nutrient solution. c. Dry matter. The dry matter content should be such that the filter structure has a homogeneous structure, while not becoming soggy, with the risk of creating anaerobic zones. An optimum dry matter content is typically in the range of 25 to 40%, in the experience of Odournet, while the literature indicates a wider range of 40-60% dry matter (IPPC BREF, 2003). d. Irrigation water. The quality of the water used for irrigation can have an impact on the availability of nutrients. Well water usually contains high iron levels, which can bind phosphate and cause a nutrient deficiency in the biofilter. e. pH. The optimum range is between pH 6 and 8. When the acidity moves outside this optimum range the filter can still function, but the population will become more specialised and hence limited in its biodiversity, which may influence its effectiveness. f. pH buffer capacity. Some materials have a larger capacity to buffer pH in situations where acids are being formed by the biological oxidation of reduced sulphur compounds (e.g. H2S) and ammonia (NH3). g. Nitrate. The total content of ammonium (NH4+), nitrite (NO2-) and nitrate (NO3-), expressed as NH4+NOx-N in g per kg wet material, can impact on the microbial activity. Literature values indicate that at values in excess of 4 g·kg-1 wet material the nitrification in the filter is completely inhibited. An alternative indicator is that a concentration in excess of 6 g·litre-1 of nitrate in the liquid fraction cause a reduction in microbial activity. A range of between 0.25 and 3.5 g of NH4+NOx-N per kg wet material should be maintained.
32
ANTON PHILIP VAN HARREVELD
h. Sulphate. High concentrations of sulphate (SO42-), formed by biological oxidation of reduced sulphur compounds can cause low pH and an excess of conductivity, which in turn indicated suboptimal osmotic conditions for the biomass. i. Conductivity. The electric conductivity of the media is indicative for the salinity of the liquid in the media. If the salinity, or the concentration of dissolved ions, becomes too high, the osmotic properties of the liquid become unfavourable for microbial activity. As a rule of thumb the electrical conductivity should be less than 1000 μS·cm-1. 3. Raw gas characteristics: a. Temperature. The raw gas should enter the biofilter at a temperature that is suitable to maintain the optimum media temperature as discussed above. b. Relative humidity. The raw gas should enter the biofilter media fully saturated with humidity to avoid any drying out of the filter. Drying out of the filter media is the most common cause of biofilter failure, in the experience of Odournet consultants. c. Odour concentration. The odour concentration of the raw gas will determine the job at hand of the biofilter. As the range of odorants can be extremely variable it is difficult to provide a full discussion. Generally speaking, biofilters typically achieve higher efficiencies when used at higher concentrations (100000 to 1000000 ou E·m -3) than at lower concentrations. Compounds that are less soluble are treated with less efficiency, for obvious reasons. d. Absence of toxic or inhibiting compounds. Toxic compounds in the raw gas should be avoided (CO, chlorinated compounds). Large concentrations of less odorous compounds, such as CH4, may reduce the effectiveness of the filter in treating the more odorous compounds. This paper explores the data obtained from a large number of measurements conducted in the Netherlands and Belgium in a variety of biofiltration units servicing composting processes. Some of the units were assessed repeatedly over a period of approx. 5 years. The measurements were all conducted by dynamic olfactometry according to the EN13725:2003 standard, in laboratories with an accredited quality system according to ISO17025. The concentration of the treated gas in relation to some of the parameters discussed above is explored in a number of graphs, to provide an insight based on statistics obtained under real-world conditions. As these observations were obtained from actual performance evaluations, and not from a controlled experimental setup, it was not possible to keep all variables constant and vary one parameter of interest.
A STATISTICAL PERSPECTIVE ON BIOFILTER PERFORMANCE
33
Covariance between non-independent variables is therefore a factor to be kept in mind.
2 MATERIALS AND METHODS 2.1 OLFACTOMETRY The odour concentrations in ouE·m-3 of raw gas and treated gas flows, collected before and after the biological treatments, were measured in strict adherence to the NVN2820:1996 and later the EN13725:2003 standard for olfactometry (Van Harreveld, 1998). All measurements were conducted in laboratories with a quality system according to ISO17025 in place, and accredited on this basis by the Accreditation Council of the Netherlands, under the umbrella of European Accreditation. All measurement results are based on a minimum of duplicate samples, and typically on triplicate samples of raw gas and treated gas. The geometric mean of the individual measurements is presented in the graphs. The measurements were conducted during the period between 1995 and 2000.
3 RESULTS AND DISCUSSION 3.1 TREATED GAS ODOUR CONCENTRATION IN RELATION TO AREA FLOW LOADING Figure 1 shows the relation of outgoing odour concentration at different levels of area loading, in the range between 40 and 350 m3·m-2·hour-1.
Figure 1. Relation between the odour concentration in the treated gas and area flow loading.
34
ANTON PHILIP VAN HARREVELD
The graph in Figure 1 does not show a clear correlation between area flow loading and treated gas odour concentration in the range considered. However, concentrations of the treated gas below 1000 ouE·m-3 are only seen to occur at area flow loadings below approximately 180 m3·m-2·hour-1. 3.2 TREATED GAS ODOUR CONCENTRATION IN RELATION TO RESIDENCE TIME Figure 2 shows the relation of outgoing odour concentration at different residence times, in the range between 20 and 90 seconds.
Figure 2. Relation between the odour concentration in the treated gas Cod in ouE·m-3 and residence time in seconds.
The graph does not show a clear correlation between residence time and treated gas odour concentration in the range considered. 3.3 TREATED GAS ODOUR CONCENTRATION IN RELATION TO THE TEMPERATURE OF THE RAW GAS
Figure 3 shows the relation of outgoing odour concentration at different temperatures of the raw gas, in the range between 24 and 8ºC.
A STATISTICAL PERSPECTIVE ON BIOFILTER PERFORMANCE
35
Figure 3. Relation between the odour concentration in the treated gas and the raw gas temperature.
The graph does not show a clear correlation between raw gas temperature and treated gas odour concentration in the range considered. It should be emphasized that the large majority of the observations are below 40ºC. Treated gas concentrations below 1.000 were not observed at temperatures in excess of 37ºC. 3.4 TREATED GAS ODOUR CONCENTRATION IN RELATION TO DRY MATTER CONTENT OF THE MEDIA
Figure 4 shows the relation of outgoing odour concentration at different levels of dry matter content of the media, in the range between 25% and 40%. The graph does not show a clear correlation between dry matter content and treated gas odour concentration in the range considered, although less favourable treated gas values appear to occur at low dry matter content. It should be noted that the range considered coincides with the range of advisable values.
36
ANTON PHILIP VAN HARREVELD
Figure 4. Relation between the odour concentration in the treated gas and dry matter content of the media.
3.5 TREATED GAS ODOUR CONCENTRATION IN RELATION TO PH OF THE MEDIA Figure 5 shows the relation of outgoing odour concentration at different levels of pH of the media, in the range between 5.0 and 7.5.
Figure 5. Relation between the odour concentration in the treated gas and pH of the media.
A STATISTICAL PERSPECTIVE ON BIOFILTER PERFORMANCE
37
The graph does not show a clear correlation between pH and treated gas odour concentration in the range considered, although less favourable treated gas values appear to occur at the extremes of the range considered. It should be noted that the range considered almost coincides with the range of advisable values of pH between 6 and 8. 3.6 T REATED
GAS ODOUR CONCENTRATION IN RELATION TO RAW GAS AMMONIA
CONCENTRATION
Figure 6 shows the relation of outgoing odour concentration at different levels of raw gas ammonia concentration, in the range between 0 and 120 ppm, with one observation at approximately 200 ppm.
Figure 6. Relation between the odour concentration in the treated gas and ammonia concentration.
The graph does not show a clear correlation between the ammonia concentration in the raw gas and the treated gas odour concentration in the range considered. Even at concentrations in excess of 50 ppm, that are generally considered undesirable, the treated gas odour concentration appears to remain at fairly typical levels. However, it should be noted that these are snapshot values. The detrimental effect of sustained high ammonia loads will show itself with time, when the electrical conductivity is increased due to increased levels of NH4+NOx-N, leading to reduced
38
ANTON PHILIP VAN HARREVELD
microbial metabolism and hence treatment capacity. For the observations presented in the figure above information on the standing time of the filter medium is not available. 3.7 TREATED
GAS ODOUR CONCENTRATION IN RELATION TO THE CONCENTRATION OF THE
RAW GAS
Figure 7 shows the relation of outgoing odour concentration at different concentrations of raw gas odour concentration, in the range between a few thousands and 500000 ouE·m-3.
Figure 7. Relation between the odour concentration in the treated gas and the raw gas.
Even at very high raw gas concentrations, of several hundreds of thousands of ouE·m-3, the treated gas will have a concentration rarely exceeding 10000 ouE·m-3. Of the measured odour concentration in the treated gas (n = 123) of biofilters treating organic green waste, 4% of the odour concentrations in the treated gas were found to be less than 500 ouE·m-3, a requirement often derived from German regulations. Only 14% of treated gas measurements was < 1000 ouE·m-3, while 52% was < 2500 ouE·m-3, a criterion used in the Netherlands for a as indicative for a well functioning biofilter. Based on the same data, the removal efficiency of the biofilters is plotted in relation to the odour concentration of the raw gas in Figure 8.
A STATISTICAL PERSPECTIVE ON BIOFILTER PERFORMANCE
39
Figure 8. Relation between the odour removal efficiency an the odour concentration in the raw gas.
The graphs show that biofilters have a capability to treat a wide range of raw gas odour concentrations with a remarkably consistent efficiency, typically at >80%. The lower efficiencies are presumably an indication for suboptimal operational conditions, as they typically occur at lower odour loads.
4 CONCLUSIONS The observations presented in this study lead to the following conclusions: 1. Biofilters appear to be a very robust treatment method suitable for reducing odour emissions in raw gas from green waste and organic fraction composting processes, ranging from a few tens of thousands of ouE·m-3 to half a million of ouE·m-3 with odour removal efficiencies in the majority of cases in excess of 90%. 2. For the parameters considered in this study, which were generally within the suggested range of values in the literature, no clear relation between the parameter value and the treated gas odour concentration could be observed. 3. The values of residual odour concentration are often in excess of the targets that are applied in various countries as emission criteria for composting odours after treatment with biofiltration. Of the data considered:
40
ANTON PHILIP VAN HARREVELD
a. 4 % satisfied the German criterion of Cod, treated < 500 ouE·m-3 b. 14 % complied with the criterion of Cod, treated < 1000 ouE·m-3 that is often seen in Spanish waste management contracts and environmental licences. c. 52 % of the observations was found to satisfy the criterion of Cod, treated < 2500 ouE·m-3 that is mentioned in the Netherlands Emission Guideline as indicative for a well functioning biofilter. 4. A performance criterion more restrictive than a treated gas concentration of Cod, treated < 2500 ouE·m-3 appears to be too restrictive and not viable as a value that operators need to attain consistently.
5 ACKNOWLEDGEMENTS The data for this study were provided by anonymous clients of the Odournet group and compiled by Margrethe Bongers, formerly a senior consultant at Odournet.
REFERENCES 30.BimSchV, Bundesimmissionsschutzgesetz, Dreißigste Verordnung zur Durchführung des BundesImmissionsschutzgesetzes paragraph 6, Emissionsgrenzwerte, 2001, http://bundesrecht. juris.de/bimschv_30/index.html EN13725:2003. Air Quality – Determination of Odor Concentration by Dynamic Olfactometry, Brussels: CEN. Infomil. (2000) NeR Nederlandse Emissie Richtlijn. (Netherlands Emissions Guideline), ISBN 90 76323 01 1, 2000 (English version available via www.infomil.nl). IPPC BREF 2003, Reference Document on Best Available Techniques in Common Waste Water and Waste Gas Treatment / Management Systems in the Chemical Sector, European Commission, subclause 3.5.2.1., Seville, http://eippcb.jrc.es NVN2820:1990 Luchtkwaliteit. Sensorische geurmetingen met olfactometer (Air Quality – Sensory measurement of odour using an olfactometer), NEN Netherlands Normalisation Institute, Delft, 1990. www.nen.nl Van Harreveld, A.Ph. (1998) A review of 20 years of standardization of odour concentration measurement by dynamic olfactometery in Europe. J. Air Waste Manage. Assoc. 49: 705-715.
41
Refurbishment of an odour collection and biofilter treatment system at a municipal solid waste composting facility in Perth, Western Australia TERRY SCHULZ AND STUART MCALL Principal of the Odour Unit Pty Limited, Sydney, Australia and CEO of Southern Metropolitan Regional Council, Perth, Australia
ABSTRACT This paper discusses a case study at a large, regional Municipal Solid Waste composting plant in Perth, Western Australia, where odour emissions from the facility have resulted in adverse odour impacts in the surrounding community. The plant was designed as a fully-enclosed operation, where all ventilation air from two major processing buildings was to be collected and treated in four large biofilters. The paper describes a detailed investigation into the collection and treatment systems at the plant, which identified deficiencies in the design of both systems. It contains the results of the physical and olfactory investigations, and documents the design for an upgraded collection and biofilter-treatment system that has subsequently been installed.
1 INTRODUCTION Southern Metropolitan Regional Council owns and operates a large Municipal Solid Waste Composting Facility (WCF) in Canning Vale, Perth, Western Australia. The WCF has a capacity of 150,000 tonnes/year of MSW, using a fully-enclosed, ‘invessel’ composting process. It is one of four large MSW composting plants in Australia. The location of the Canning Vale WCF is within a larger Regional Resource Recovery Centre, in the suburbs of Perth, the capital of Western Australia. It has been operating since late-2002.
41
42
TERRY J. SCHULZ AND STUART McALL
Perth is a city of approximately 1 million people and is the capital of Western Australia. It has a very hot and dry climate, with summer temperatures regularly above 350 Celcius, and ambient relative humidities that can stay below 40 per cent for several months at a time. As will be explained, these weather conditions can have a strong influence on the design and performance of biofilter-based odour control systems in the area. Since the commissioning of the Canning Vale WCF a high level of adverse odour impacts has been experienced in the residential areas up to 400 metres from the plant. This situation lead to the development of a strong and well-organised community opposition to the continued operation of the facility, and resulted in increasing pressure from the WA environmental regulatory authority to mitigate the odour emissions. In February 2006, following a number of earlier attempts to identify the reasons for the unacceptable odour emissions from the plant, investigations commenced on a comprehensive assessment of the design and performance of the entire odour control collection and treatment system at the plant. The methodology used, the findings and the implementation of the findings of that study are described in this paper. The outcome of the study has been a comprehensive re-design of the odour collection system and a total refurbishment of the four biofilters that treat all odour emissions to atmosphere. A full report (The Odour Unit, 2006) documenting the full investigation and results is available on the SMRC website.
2 DESCRIPTION OF THE FACILITY The Canning Vale WCF was designed using the proprietary ‘Bedminster’ composting system. This system currently receives 120,000 tonnes per year of raw MSW from six local government areas in Perth, and processes it into a fully composted product, in the process separating plastics and metals for subsequent recycling. While there are several different types of MSW composting processes in use around the world, the features of this particular WCF composting process may be summarised as follows: • A separate MSW receivals hall (the Tipping Building), having a volume of approximately 20,000 cubic metres; • Four, rotating tunnel digesters, in two banks of two digesters, each bank receiving MSW in the Tipping Building and unloading into different ends of the Aeration Building; • The Aeration Building, in which the partially digested material is processed further using a forced aeration technique for a minimum of 45 days. The volume of this building is 200 metres long, 50 metres wide and has a volume
REFURBISHMENT OF AN ODOUR COLLECTION AND BIOFILTER TREATMENT SYSTEM
43
of approximately 100,000 cubic metres. This paper will focus predominantly on the capture and treatment of the odour emissions from this building; • The Load-out wing of the Aeration Building, where fully composted material is stored and loaded into trucks; and • The biofilters. There are four biofilters in total. Two smaller units service the Tipping Building, while two larger biofilters receive and treat air drawn from the Aeration Building. The two Tipping Building biofilters are each fed by two identical fans, each having a capacity of 55,000 m3/hr. One of these fans extracts air from above the inlet to the digester vessels, while the other draws air from within the body of the building. The discharge from each fan enters a packed-column scrubbing unit, designed to remove particulates from the air stream and provide some humidification. The design of the building is symmetrical, having two digesters, two fans, two scrubbers and a biofilter on each side. The original design of the Tipping Building had a single roller-shutter doorway for truck access. The Tipping Building biofilters (#3 and #4), like the Aeration Building biofilters (#1and #2), are open bed, roofed units, featuring a foul air distribution system based on the use of multiple, slotted PVC pipes buried in gravel, and fed from a longitudinal distribution header duct. Each fan feeds to its own header duct and biofilter section. As a result, each of the two Tipping Building biofilters behaves as two separate units, although visually appearing to be a single biofilter. There is no capability to direct the output from one fan to another section of the biofilter. The Aeration Building foul air extraction system uses six 72,000 m3/hr fans mounted outside the longest side of the building. Fans 1 to 4 direct air to Biofilter #1, (the largest biofilter – 1,500 m2), while fans 5 and 6 service Biofilter #2 (1,000 m2). The original design of the extraction system involved the six fan-suction ducts terminating inside the wall, approximately five metres above ground level. A number of wall-mounted inlet vents were located along the wall opposite the fans, with the intention of affecting a flow of air laterally across the building. The two ends of the building, containing the odorous digester outlets, the compost conveyors and trommel screens, were not fitted with any form of collection system. The roof height at the ends of the building is higher than in the aeration/maturation section, rising to a maximum height immediately adjacent to the roller shutter doors used to load-out screenings. As with the Tipping Building biofilters, the two Aeration Building biofilters behave as six independent units, each fed by its own fan, with no cross connection between these units. A summary of the odour collection and treatment system ‘as installed’ at the WCF is shown in Table 1.
44
TERRY J. SCHULZ AND STUART McALL
Table 1. Canning Vale WCF odour control system summary. Tipping Building Ventilation
Tipping Building Biofilters
Aeration Building Ventilation
Aeration Building Biofilters
20,000 m3 building volume 4 fans, each 55,000 m3/hr 10.5 air changes/hour 2 biofilters, 800 m2 bed area per biofilter plenum pipe-in-gravel air distribution wood chip/compost bed medium 120,000 m3 building volume 6 fans, each 72,000 m3/hr fan suction ducts in wall cross flow air movement 4.3 air changes/hour 2 biofilters, 1,500 m2 and 1,000 m2 bed area plenum pipe-in-gravel air distribution wood chip/compost bed medium
3 PROBLEMS IDENTIFIED IN PREVIOUS STUDIES As mentioned in Section 1, the operation of the WCF had resulted in the detection of nuisance odours identified as having originated from the facility, up to 400 metres from the plant. Unfortunately the odour complaint records retained by the regulator and SMRC contained little odour descriptor information and therefore did little to identify the source activity within the WCF responsible for the nuisance. Typical descriptors used by complainants were ‘garbage’, ‘waste’, ‘landfill’ and ‘rotting waste’. A previous odour impact assessment study carried out in 2005 identified three key odours that were judged to be responsible for the bulk of the odour complaints. These odours, and their sources within the WCF, are summarised in Table 2. The clear conclusion from this earlier investigation was that the biofilters and Aeration Building were the most problematical odours, and that the odours from the Tipping Building, were not likely to be detectable in the community. At that time the biofilters were known to be operating sub-optimally, due to bed moisture control problems and the resulting tendency for foul air short-circuiting. In an effort to counter this problem SMRC elected to decrease the fan speed, and hence output, of each of the ten biofilter fans. This decision was taken in early 2005.
REFURBISHMENT OF AN ODOUR COLLECTION AND BIOFILTER TREATMENT SYSTEM
45
At the time of this study there appeared to be little understanding of the magnitude of any fugitive odour releases from the Aeration Building. Table 2. Odour types within the Canning Vale WCF (2005 Study). Odour character Garbage, rotting vegetative material Fruit cake, fermented Fruit Stale air or water, rubber tyre Cheesy, baby vomit Earthy, damp, mild garbage Forest floor, mouldy, medicinal,
Likely odour source Tipping Floor, Digester discharge area, Aeration floor Aeration floor Aeration floor Biofilters 1 and 2 Biofilters 3 and 4
4 INVESTIGATION METHODS The investigation was carried out from February to April 2006. In the course of this investigation a number of methodologies were used to assess both odour impacts and the performance of the odour collection and treatment systems. These are summarised below. 4.1 ODOUR IMPACT ASSESSMENT 4.1.1 OLFACTORY ASSESSMENT AT THE SOURCE This procedure involved The Odour Unit staff assessing the character, intensity and relative emission rate of odours generated within the facility. While subjective in nature, olfactory assessment is able to differentiate between different odour characters. If a clearly defined odour can be detected at a processing source, this enables this same odour and source to be identified if detected downwind of the facility during an ambient odour assessment. 4.1.2 AMBIENT ODOUR SURVEYS The use of quantitative ambient odour assessments is now commonplace in odour impact assessment worldwide. In the absence of an Australian Standard method for this type of assessment a simple but robust procedure was developed, involving a trained and calibrated assessor systematically traversing the downwind areas from an odour source, logging wind data and the incidence, character and intensity of odours present. In this project the ambient surveys were carried out in the neighbouring down wind residential areas.
46
TERRY J. SCHULZ AND STUART McALL
4.1.3 COMMUNITY FEEDBACK The scope of work for this project required TOU to interact with community members, as a means of identifying the most problematical odours from the facility. On two occasions samples of odours from a range of processing sources within the WCF were presented to community representatives. The sources assessed were: • • • •
Tipping Floor odour; Aeration Floor odour; Tipping Floor biofilter-treated air; and Digestion odour (collected during a digester unloading event).
On the second occasion a community representative was presented with samples collected at source and therefore at ‘full strength’, as well as the same samples diluted to a level likely to be experienced in the community (10 to 20 odour units). This was done to cover the possibility that odour character could change with dilution. 4.2 ODOUR CONTROL SYSTEM ASSESSMENT 4.2.1 DUCT AIRFLOW MEASUREMENT Airflow rates in the various ducts in the collection system and into the biofilters were measured using a hot-wire anemometer that measured gas velocity. Where possible the velocity measurements were taken in a straight section of ducting to maximise accuracy of the measurement. All flow readings presented in this report are expressed at the temperature prevailing in the duct. 4.2.2 DUCT PRESSURE MEASUREMENT Pressure or vacuum was measured simultaneously with velocity measurement in each of the ducts. A simple water-filled manometer was used. The pressure readings were used to determine the pressure duty on each of the 10 fans tested, as well as the back-pressures upstream of each of the cells in the biofilters. Relative humidity was measured into each of the biofilters, and in the Tipping Floor and Aeration Buildings. Ambient readings were also taken. A TSI VelociCalc 8386A instrument was used for this purpose. 4.2.3 BIOFILTER SURFACE FLOW MEASUREMENTS A technique was developed to assess the extent to which spatial variation in airflow rates occurred across the surface of the biofilter beds. This technique involved the use of a cylindrical, open-bottomed ‘hood’ (0.8 m diameter) that was placed on the surface of the biofilters. A 100 mm PVC pipe mounted on the lid of the hood accelerated the low velocity of air leaving the bed, enabling a velocity reading to be accurately taken.
REFURBISHMENT OF AN ODOUR COLLECTION AND BIOFILTER TREATMENT SYSTEM
47
This technique has been found to accurately assess relative changes in airflow leaving the surface of a biofilter bed. In this respect it is useful in assessing whether short-circuiting or blockages exist in biofilter beds. A variation in velocity readings of greater than 30 percent would indicate an air distribution problem. However the technique is not able to measure the true airflows leaving a biofilter bed, due to the intrusive nature of the device. 4.2.4 BIOFILTER PRESSURE LOSS MEASUREMENTS A means of quantifying the pressure losses across Biofilter 1 was needed in order to differentiate between the losses through the plenum distribution pipes and the biofilter medium. Two holes were dug in Biofilter 1, one in Cell 1 serviced by Fan 1, and the second at Cell 4. Cell 4 is serviced by Fan 4, which in turn draws particulateladen air into the biofilter from the final product screening area. Two sets of manometer tubing were inserted into the bed at these two locations. The first tube at each location was located immediately above the plenum pipe and beneath the shade cloth mesh layer, in the gravel layer, beneath the biofilter medium. The second tube was positioned above the shade cloth. The holes were backfilled and compacted back to as near their original condition as was possible, and pressure readings taken 4.2.5 BULK AIR MOVEMENT ASSESSMENT This study found it imperative that the general airflow patterns be assessed within the Aeration Building. A portable smoke generating machine was used for this purpose. The machine is able to generate copious quantities of smoke for 15-20 seconds at a time, and was sufficient to fill a section of the building and determine airflow patterns. This method was used with great success and was easily able to identify deficiencies in the odour collection system design. 4.2.6 ODOUR TESTING A decision was taken early in this study not to use olfactometry testing. While such testing would have quantified the odour removal performance of the biofilters it was felt that the performance of the biofilters was already known to be sub-optimal.
5 RESULTS 5.1 ODOUR IMPACT ASSESSMENT The results of the three methods used to assess the WCF processing sources emitting the odours being detected in the community showed clearly that two odours present in the Aeration Building were responsible. These sources were the aerated
48
TERRY J. SCHULZ AND STUART McALL
compost piles (a compost odour) and the digester unloading and screening operation (a fermenting/fruity odour). The source of both of these odours was determined to be fugitive odour emissions from the Aeration Building. It was also determined that incomplete odour destruction in the biofilters was also causing odour impacts, suggesting that the biofilters needed refurbishment. 5.2 DUCT AIRFLOW AND PRESSURE MEASUREMENTS The results of airflow and pressure testing are shown in Table 3. Table 3. Airflow and pressure measurements WCF biofilters (16 March 2006). Location Biofilter 1 – Fan 1 Biofilter 1 – Fan 2 Biofilter 1 – Fan 3 Biofilter 1 – Fan 4 Biofilter 2 – Fan 5 Biofilter 2 – Fan 6 Fan specifications Biofilter 3 – Fan 11 Biofilter 3 – Fan 12 Biofilter 4 – Fan 9 Biofilter 4 – Fan 10 Fan specifications
Airflow (m3/hr) 45,600 45,200 43,900 40,700 * 49,600 72,000 35,700 46,500 30,700 34,500 55,000
Biofilter delivery pressure (kPa) 3.9 3.9 3.6 3.9 * 4.0 4.5 (suction plus delivery) 1.6 0.9 ** 1.8 1.6 4.5 (suction plus delivery)
* not accessible for measurement ** biofilter bed was dry, scrubber out of service
5.3 FOUL AIR HUMIDITY MEASUREMENTS The results of the humidity testing on the Tipping Floor Biofilters 3 and 4, which were equipped with scrubber/humidifier units for each fan, indicated that acceptable relative humidity (RH) readings above 90% could only be achieved when the ambient RH was above 60%. The overall finding was that the scrubber/humidifiers were only able to achieve an RH net increase of 31.3 percentage units. Given that the WCF operates in Perth’s very dry climate, where summer RH values in the range 2030% are normal, it was determined that the Tipping Floor Biofilters 3 and 4 were operating under considerable stress. The problem was found to be even worse for the Aeration Building Biofilters 1 and 2 where no form of foul air humidification was provided.
REFURBISHMENT OF AN ODOUR COLLECTION AND BIOFILTER TREATMENT SYSTEM
49
5.4 BIOFILTER SURFACE FLOW MEASUREMENTS These results showed erratic and unacceptable airflows across the surface Biofilters 1 and 2. A five-fold variation in surface flows was measured, indicating poor distribution of foul air beneath the biofilter beds. 5.5 BIOFILTER PRESSURE LOSS MEASUREMENTS Table 4 contains the results of pressure loss testing in two sections of Biofilter 1. The section serviced by Fan 1 was receiving air typical of that contained in the bulk of the Aeration Building. The section serviced by Fan 4 was receiving air laden with a high particulate loading, drawn from the product screening area. Table 4. Pressure losses across Biofilter 1 (5 April 2006). Location Southern end of Biofilter 1 (Fan 1) Pressure into Plenum Pipe Outside pipe and below shade cloth mesh (pressure loss through slotted pipe) Above shade cloth mesh, below medium At biofilter medium surface Northern end of Biofilter 1 (Fan 4) Pressure into Plenum Pipe Outside pipe and below shade cloth mesh (pressure loss through slotted pipe) Above shade cloth mesh, below medium At biofilter medium surface
Pressure (kPa) +3.80 +2.00 (1.80) +1.80 0.00 +4.30 +1.10 (3.20) +1.10 0.00
Note: these readings were taken at the reduced airflows shown in Table 3
5.6 BULK AIR MOVEMENT ASSESSMENT The smoke testing carried out inside the Aeration Building revealed highly significant adverse airflows were occurring, as follows: • There was no consistent airflow pattern evident; • There was no cross-flow of air from the inlet louvres to the fan suction ducts; • Thermal buoyancy of the air above the compost piles was dominating the air movement within the building; and • The internal, auxilliary ‘push’ fans, designed to assist the biofilter fans in the cross-flow of air from the inlets to the fan suction ducts were having a negative effect on this desired air movement.
50
TERRY J. SCHULZ AND STUART McALL
It was concluded from the smoke testing that a thermally induced, positive air pressure was occurring under the roof of the Aeration Building, despite the action of the fans. The pressure at floor level was found to be neutral (neither positive nor negative).
6 DISCUSSION The technical component of the study were able to provide evidence that confirmed the findings of the odour impact assessment investigations, namely that fugitive odour emissions from the Aeration Building at the WCF were likely to be causing adverse odour impacts in the community. The cause of these fugitive emissions was determined to be an inability to achieve and maintain negative pressure conditions inside the building, due primarily to two factors - the reduced outputs from the six Aeration Building fans, and the thermally buoyant plumes rising from the compost piles. It was also determined that this situation was being exacerbated by the failure of the existing odour collection system within the building to capture the highly odorous emissions from the digester load-out and screening operations, at each end of the building. The pressure and flow measurements for the Aeration Building biofilters revealed a serious problem and justified the decision to decrease the fan outputs, in the short term. The results showed excessive pressure losses through the plenum pipe air distribution system, due to a combination of high design air velocities and large loses through the slotted orifices in the pipes. While the pressure loses through the biofilter beds were acceptable, the medium was found to be at the end of its useful life.
7. REFURBISHMENT DESIGN AND PERFORMANCE As a result of the findings of this study the odour collection system within the Aeration Building and all four biofilters have been refurbished, as follows: • The fans have been returned to their design airflows; • A longitudinal header duct has been installed along the length of the building, immediately under the roof apex. All six fans draw from this header duct. Secondary ducts have been installed at each end of the building to capture the air from the digester load-out and screening emissions; • The biofilter plenum pipes have been fitted with 30 to 40 mm holes at 500 mm spacing, to reduce pressure losses into the biofilter medium. Approxi-
REFURBISHMENT OF AN ODOUR COLLECTION AND BIOFILTER TREATMENT SYSTEM
51
mately 50,000 holes have been installed in 560 plenum pipes in the four biofilters; • The medium in all four biofilters has been replaced; and • A high-pressure misting system has been installed in the Aeration Building. This system is able to raise the ambient humidity in the building to at least 90%. This system was commissioned in early 2007 and has been found to achieve strong biofilter performance and measurable negative pressures at all levels and locations within the Aeration Building. While this represents an effective technical solution to the odour emissions problem and has resulted in a sharp decrease in odour complaints, the local community opposition to the continued operation of the WCF is still significant. SMRC is continuing to work with the community to allay its residual concerns.
REFERENCE The Odour Unit (2006) Canning Vale waste composting facility Odour control system review. (www.smrc.com.au/pdf/SMRC_FinalReportForCommunity18052006.pdf).
This page intentionally left blank
53
Odour and H2S degradation in a full scale biofilter with a mineral based organic coated filter media FRANZ-BERND FRECHEN1, WOLFRAM FRANKE1
AND
BJÖRN SCHOLL2
1
Department of Sanitary and Environmental Engineering, University of Kassel, Kurt-Wolters-Strasse 3, 34125 Kassel, Germany 2 Magistrate of the city of Frankfurt/Main, Goldsteinstrasse 160, 60528, Frankfurt/Main, Germany
ABSTRACT In order to minimize the odorous emissions from sludge storage tanks on the waste water treatment plant «Niederrad» of the city of Frankfurt/Main, about 12.000 m3/h of foul air has to be treated. Due to high Hydrogen-Sulphide (H2S) loads, the installed standard biofilter systems failed operation after one year. Thus, one of the existing filter beds was filled with a mineral based organic coated material; the other one was re-filled at the same time with a standard biofilter media to allow a comparative study. In a long term monitoring program from May 2006 to June 2007, both media were compared regarding degradation of H2S and odour. The one-year measurement program revealed that the mineral high performance media performs much better then the standard organic media.
1 INTRODUCTION In order to minimize the odorous emissions from sludge storage tanks on the wastewater treatment plant «Niederrad» of the city of Frankfurt/Main, about 12.000 m3/h of foul air has to be treated. A first step was the installation of a biofilter with a humidifier, designed in 2004 according to estimated loads. Shortly after start-up of the biofilter operation, severe odorous emissions from the biofilter were recognized. In 2005, a first filter monitoring campaign, performed by the operator, showed that the H2S concentration in the raw gas was much higher than estimated during design and was
53
54
FRANZ-BERND FRECHEN, WOLFRAM FRANKE AND BJÖRN SCHOLL
also much higher than recommended in the VDI guideline 3477 (2004) for biofilter design. As a replacement of the biofilter with a chemical scrubber would have caused serious problems in terms of space restrictions and economic aspects, the first choice was to undertake a test with a new, high performance biofilter media, which, according to the manufacturer, would be able to deal with high H2S loads. This so called IHCS (inert hydrophilic compound structure) media was filled in one of the two existing filter beds in spring 2006; the other one was re-filled at the same time with a standard biofilter media to allow a comparative study. In a long term monitoring program from May 2006 to June 2007, performed by the University of Kassel, both media were compared regarding degradation of odour and H2S.
2 MATERIALS AND METHODS 2.1 TREATMENT PLANT LAYOUT The technical specifications of the air treatments system is given in Table 1. Table 1. Technical specification of air treatment. Aspect Air stream Filter size Filter volume Filter media
Unit m3/h m2 m3 -
line 1 6.000 85 150 Mineral based organic coated IHCS media by Otto Industries, Bad Berleburg (D)
line 2 6.000 85 150 Wood mix (root wood, chipped wood and bark mulch)
The IHCS media was developed by Otto Industries, Bad Berleburg (D), and it was the first time that its performance was comparatively tested in a large scale application on a wastewater treatment plant. The core of the IHCS media consists of clay granulate and the coating is a mixture of cement, fertilizer, activated carbon and chipped wood (OTTO, 2005a). The material could be recycled by the producer (OTTO, 2005b). Figure 1. shows a sketch of the installed air treatment system.
ODOUR AND H2S DEGRADATION IN A FULL SCALE BIOFILTER WITH A MINERAL
55
Figure 1. Scheme of the air treatment system with sampling points.
All sampling points (behind ventilator line 1 [bv1], behind humidifier line 1 [bh1], behind filter line 1 [bf1], bv2, bh2 and bf2) are marked in the scheme. For comparison of the filter performance, this paper focuses on sampling points bh1, bh2, bf1 and bf2. 2.2 ODOUR MEASUREMENT The quantification of odour is the key to describe the odorous performance of an air treatment system. In order to fulfil European standards, the odour concentration was measured according to DIN EN 13725:2003, updated with DIN EN 13725:2006. The basis of this measurement method is that odour could be recognized by a person if the individual odour threshold is exceeded. For example, if an odorous air needs to be diluted 2n times for just not exceeding and n times for exceeding the odour threshold, the odour concentration is the geometric mean of n and 2n, which is 1.41n. For improving statistical accuracy on the whole, twelve or more single
56
FRANZ-BERND FRECHEN, WOLFRAM FRANKE AND BJÖRN SCHOLL
measurement results are needed to calculate the dilution number. Although a dilution number has no unit, the odour concentration cod is denoted with European odour units per cubic meter (ouE/m3). 2.3 H2S-MEASUREMENT H2S-Measurement was performed with Odalog measurement devices by AppTek (AUS). An OdaLog detects H2S-concentrations from 0 ppm to 200 ppm with a resolution of 0,1 ppm. A data logger is included, which can store 32.000 sets of values (date/time, concentration of H2S and temperature). The device has an IR-connector for periodical data acquisition and optionally (OdaTrak) a fibre optics connector for online data transfer.
3 RESULTS AND DISCUSSION 3.1 ODOUR MEASUREMENT The result of the odour measurement program is shown in Figure 2.
Figure 2. Results of the odour measurement program.
On the whole, eleven samples were taken and analysed during one year. The IHCS-filter (bf1) always showed a higher efficiency and most of the time reached
ODOUR AND H2S DEGRADATION IN A FULL SCALE BIOFILTER WITH A MINERAL
57
significant lower clean gas concentrations than the standard woodmix biofilter (bf2). The difference between the two materials tended to become bigger near the end of the measurement campaign. The deviation in concentration values in raw gas between both air treatment lines was caused by the disadvantageous connection of the pumping station exhaust air outlet behind the air splitter into line 1 pipe (Figure 1) that made this highly polluted air stream move only into line 1 and not into line 2. 3.2 H2S MEASUREMENT The online monitoring delivered 350,000 to 430,000 sets of values for each sampling point during one year. The difference was caused by availability of devices and some necessary service stops. In Figure 3, the frequencies of the H2S-concentrations is presented. The raw gas concentration of H2S showed an average value of 25.7 ppm for the IHCS filter (bh1) and 23.9 ppm for the wood mix biofilter (bh2). It can be seen clearly that H2S degradation was performed successfully in both filter units, as H2S values of less than 1 ppm were detected in the clean gas. The effect of connecting the pumping station exhaust air to the pipe system behind the air splitter can be seen clearly. Only line 1 employing the IHCS filter material is loaded with this highly polluted air.
Figure 3. Results of H2S-Monitoring.
58
FRANZ-BERND FRECHEN, WOLFRAM FRANKE AND BJÖRN SCHOLL
An estimation of the total H2S load for both filters can give an impression of the possible load capacities. Other sulphurous gases like thiols or dimethyl sulfide (DMS) were disregarded as well as the sulphur runoff from the two biofilter line effluents that will definitively contain sulphuric acid; because no data were available on this streams. Keeping this in mind, the estimated H2S loads are 1,919 kg H2S/a for the IHCS filter and – slightly less – 1,786 kg H2S/a for the wood mix biofilter (6,000 m3/h, mean values of H2S concentrations). With regard to filter volumes of 150 m3, the specific H2S load capacity of the IHCS was 12.8 kg/m3 without filter breakdown and for the standard wood mix biofilter 11.9 kg/m3 with operation failure in the end.
4 CONCLUSIONS Both air treatment systems operated successfully during most of the time. Although the IHCS filter was loaded with more H2S and odour than the standard filter, it was more advantageous concerning degradation performance, even increasing towards the end of the one year operation. Hence, the mineral based organic coated biofilter media seems to be an interesting way of improving standard biofilters. This material allows operation of biofilters with average H2S concentrations of 20 ppm, which is much higher than VDI guideline 3477 (2004) recommends. Economic aspects depend on the alternatives. The IHCS media is more expensive (specific material cost 270 €/m3) than a standard biofilter media (specific material cost 60 €/m3). In this case the upgrade of the biofilter with the new IHCS media was more economic than the installation of a chemical scrubber regarding investment as well as annual cost. REFERENCES DIN EN 13725:2003: Air quality. Determination of odour concentration by dynamic olfactometry; German version EN 13725:2003. Beuth Verlag. Berlin, Germany. 2003. DIN EN 13725:2006: Air quality. Determination of odour concentration by dynamic olfactometry; German version EN 13725:2003. Corrigenda to DIN EN 13725:2003-07. German version EN 13725:2003/AC:2006. Beuth Verlag. Berlin, Germany. 2006. Otto. (2005a) Safety data sheet of IHCS Media. Otto Luft und Klimatechnik GmbH & Co. KG. Bad Berleburg, Germany. 28.02.2005. Otto. (2005b) Biological air pollution control. Product information. Otto Luft und Klimatechnik GmbH & Co. KG. Bad Berleburg, Germany. 2005. VDI Guideline 3477. (2004): Biological waste gas purification - Biofilters. Beuth Verlag. Berlin, Germany. 2004.
59
High throughput biofiltration for odour control at water purification plant VITALY ZHUKOV1, ANDREY VEPRITZKY2, LEONID MITIN2, VLADIMIR POPOV1 1 A.N.Bakh Institute of Biochemistry, Russian Academy of Sciences, Leninskiy pr., 33 119071 Moscow, Russia 2 Innovational Biotechnologies Ltd., Leninskiy pr., 33 119071 Moscow, Russia
ABSTRACT A high throughput trickling biofilter for odour control was designed basing on the principles of biotrickling filter technology developed in Moscow Bakh Institute of Biochemistry. All the necessary blocks except a fan: temperature and humidity control unit, a biofilter bed, an irrigation system, a control block and display unit are combined within one compact biofiltration module – a standard container 6000x2400x2400 mm. The plant is thermo-insulated that enables outdoor installation. The biofilter is easily scaled up by adding extra filtration beds. A typical biofiltration module rated for 5,000-10,000 m3/h has a contact time of 3-6 s (biofilter bed total volume – 10.5 m3) and a maximum footprint of 14.5 m2. After extensive pilot plant studies the first 5000 m3/h trickling biofilter easily scalable to 20000 m3/h was installed at Moscow Water Works in spring 2007 to control odour emissions - hydrogen sulfide, mercaptanes and other malodorous volatile organic compounds in up to 60 mg/m3 concentration. The performance results of the industrial biofilter are discussed.
1 INTRODUCTION Biofiltration – is an established technique to control odours (Devinny et al., 1999). Various types of biofilters have been suggested and successfully applied varying from the most simple open air units with soil/compost beds (www.bohnbiofilter.com) to sophisticated fully automated enclosed plants using proprietary artificial media and enabling full process control (Popov and Zhukov, 2005). The area of odour control is rather competitive and only most economic solutions have a chance of surviving on the market. Moreover deodouration is regarded less technically challenging than VOC 59
60
VITALY ZHUKOV, ANDREY VEPRITZKY, LEONID MITIN AND VLADIMIR POPOV
control enabling simple biofiltration plants to perform quite satisfactory and comply with existing legislation and end-user expectations. Trickling bed biofilters offer a number of advantages over conventional biological methods to treat off-gases. In some cases the contact time between the VOC laden air and the biocatalyst in trickling bed biofilters may be reduced to below 10 s range (Gabriel and Deshusses, 2003) thus minimizing the overall system footprint, dimensions and power requirements. For a number of years we are perfecting the design of the trickling biofilters (Zhukov et al., 1998; Popov and Bezborodov, 1999) and explore their potential in various applications: e.g. water soluble VOCs used in flexographic printing (Popov et al., 2004) , chlorinated compounds (Popov et al., 2005), formaldehyde (Popov et al., 2000), BTEX (Bezborodov et al., 1998), etc. Here we report an All-in-One high-throughput biotrickling filter for odour control applications and present its preliminary performance characteristics at Moscow Water Works.
2 MATERIALS AND METHODS 2.1 MICROORGANISMS As a base of the microbiological consortium used to populate the carrier in the biofilter the thionic bacterium Thiobacillus novellas has been used which efficiently degraded hydrogen sulfide. It was complimented by other strains from the in-house collection capable to utilise mercaptanes and volatile compounds present in the emissions of the water purification plants. All the strains used have been tested in a specialised certified laboratory properly authorised to perform such studies and were proved to be non-pathogenic, non-virulent and non-toxic for mammals. 2.2 CARRIER An inert polymer carrier with open-pore foam-like structure was used to immobilize the microbial consortium. 2.3 ANALYTICAL METHODS AND MONITORING Biofilter performance was monitored organoleptically and instrumentally. The personnel present on site evaluated the intensity of the smell at the outlet of the biofilter. The inlet and outlet concentrations of H2S (main source of the obnoxious odour) were routinely measured on-site electrochemically by a portable electrochemical analyzer Colion-1. The electrochemical cell analyzer was from time to time calibrated against a standard laboratory calorimetric procedure for quantifying SH-compounds. Prior to standard laboratory assay the SH-containing compounds were trapped by passing the predetermined volume of the air through the cartridges filled with glass spheres covered with solution of lead or mercury acetate.
HIGH THROUGHPUT BIOFILTRATION FOR ODOUR CONTROL AT WATER PURIFICATION PLANT
61
The assay of the gas mixtures was performed via GC/MS with Shimadzu GC2010 instrument equipped with GCMS-QP2010 Thermal Desorber (Markes International Ltd., UK) and capillary columns. The sample for GC/MS was obtained by adsorbing the volatiles by aspirating the controlled volume of air through a Tenax© column.
3 RESULTS AND DISCUSSION 3.1 CONSTRUCTION OF THE BIOFILTER To make biofilter more user-friendly and versatile a novel layout of the plant was used. An All-in-One principle was followed. All the main functional units of the plant: inlet air distribution and conditioning system; biofilter bed; irrigation and pHcontrol system; air transport system and droplet remover; control unit; etc. are enclosed within a standard 20-feet container (6000x2400x24000 mm) (Figure 1). The plant is properly thermo-insulated that enables outdoor installation. This is most important for the regions with a severe climate to which Russia evidently belongs to. When not in operation the plant can run in a stand-by mode maintaining its internal temperature regimes and thus precluding freezing or overheating.
Figure 1. Schematic outline of the All-in-One biotrickling filter.
62
VITALY ZHUKOV, ANDREY VEPRITZKY, LEONID MITIN AND VLADIMIR POPOV
The plant is very simple to start and operate. It requires just proper tap water and power supply and connection to the on-site existing drainage/sewage system. The air ducts are easily fitted to the inlet flange of the system. The system can be regarded of the type «install-and-run». The main parameters of the plant are shown in Table 1, while its general view and location on site appear in Figure 2. Table 1. Parameters of the biofiltration plant. Parameter and unit Nominal air flow, m3/h EBRT, s Optimal inlet concentrations of volatiles, mg/m3 Pressure drop, Pa Installed power, kWt Irrigation flow, m3/h Maximal air linear velocity, m/s Consumption of tap water (including evaporation), l/h Power consumption for outdoor installation (summer / winter), kWt*h Temperature of the incoming air, °C Allowed ambient temperature, °C Dimensions (HxLxW), mm Operating foot-print, m2 Dry weight, kg Wet weight (operating), kg
Value 4,000 – 11,000 3-10 < 250 350 – 1,250 < 18,5 < 12,5 < 0,70 < 110 < 3.5 / 15.5 0 – 50 (-20) – (+40) 2,475x6,300x2,400 14,5 < 3,700 < 10,500
The compact arrangement of all the components of the biofilter within one unit provides a number of technological advantages: easy transportation, erection and mounting; easy interfacing with the infrastructure existing on site (power, water, sewage, etc.); easy maintenance and service. The plant can be easily scaled up by adding either additional filtration beds or by adding extra complete units. 3.2 OPERATION OF THE BIOFILTER Main operation facilities of the sewage department of the Moscow Water Works are located rather close to the housing area. It is not feasible to relocate them, thus management of Moscow Water is looking for efficient and cost-effective technology that will enable to secure that neighbouring households do not complain about the irritating odours. An extensive programme of testing and pilot runs was launched that
HIGH THROUGHPUT BIOFILTRATION FOR ODOUR CONTROL AT WATER PURIFICATION PLANT
63
enabled Moscow Water to select a trickling biofilter technology as the best available technology to control odours. The final phase of tests with a full-scale biofiltration plant, Figure 2, is currently in progress.
Figure 2. Biofilter at the premises of Moscow Water Works.
The biofilter was mounted close to the workshop for the residue sedimenting and dewatering that produces concentrated malodourous emissions. More than 70 components that varied in concentration quite considerably over time were identified in the emission by GC/MS technique. The dominant one was methane (1500-1600 mg/m3), while hydrogen sulfide – the major irritating pollutant was present in 12-16 mg/ m3 concentration with peaks up to 60-90 mg/ m3. The profile of the inlet and outlet H2S concentrations measured on-line with electrochemical sensor is presented in Figure 3. The results of the start-up and several months monitoring confirmed reliability of the technology. After about one month acclimation period the plant reached its target performance efficiency and is able to remove >98-99 % of the hydrogen sulfide coping with the spike emissions of about 30-80 mg/m3.
64
VITALY ZHUKOV, ANDREY VEPRITZKY, LEONID MITIN AND VLADIMIR POPOV
The instrumental GC – GC/MS assays showed that such pollutants as mercaptanes, amines, limonene, aromatic compounds were degraded with the efficiency of >99 % and could not be determined at the biofilter outlet (Table 3). Thioglycolic acid – one of the major pollutants with an average concentration of 5.5-6.0 mg/ m3 and peaks up to 10 mg/ m3 was usually degraded with efficiency of 80-99 %. Considerable depletion of the methane content - 40-70 % was also noted.
Figure 3. Start-up and performance of the trickling biofilter for odour control operating at Moscow Water Works. EBRT = 6 s.
The long-term (up to one year period) monitoring of the biotrickling filter plant is required to draw sound conclusions on the efficiency and the running costs. However already preliminary results show that the new All-in-One biotrickling filter could provide a viable solution to odour control at municipal water treatment plants.
HIGH THROUGHPUT BIOFILTRATION FOR ODOUR CONTROL AT WATER PURIFICATION PLANT
65
Table 2. Efficiency of VOC degradation in the biofilter. Compound 1. Methane 2. Isobutyl amine 3. Ethyl sulfide 4. Thioacetic acid 5. Thioglycolic acid 6. Dioxane 7. Toluene 8. Ethylene diamine 9. Methyl mercaptane 10. Butyric acid 11. Dibutyl amine 12. Ethanol amine 13. Limonene 14. Dibutyl sulfide
Retention time, min 00:55 01:09 01:43 01:57 02:19 02:37 02:47 02:53 02:59 03:54 05:27 05:38 08:26 08:50
Peak area, Inlet 11226.2 72.32 5.89 211.80 86.32 22.84 32.53 27.11 21.65 6.67 8.50 11.24 13.19 3.61
arb.un. Outlet 2557.85 – – 47.54 – – – – – – – – – –
Conversion, % 77 >99 >99 78 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99
4 ACKNOWLEDGEMENTS The project was supported in part by a grant from the Russian Federal Agency of Science and Innovations 02.447.11.3001. The authors thank management of Moscow Water, in particular Dr. Dmitry Danilovich and Mr. Michael Kozlov, for co-operation and good will.
REFERENCES Bezborodov, A.M., Rogojin, I.S., Ushakova, N.A., Kurlovitch, A.E. and Popov, V.O. (1998) Optimization of the microbiological technology for purification of air emissions from the mixture of benzene, toluene and xylene. Appl. Biochem. Microbiol.(Russia) 34: 265-269. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. Boca Raton: Lewis Publishers. 299 p. Gabriel, D. and Deshusses, M.A. (2003) Retrofitting existing chemical scrubbers to biotrickling filters for H2S emission control. Proc. Natl. Acad. Sci. USA. 100: 6308-6312.
66
VITALY ZHUKOV, ANDREY VEPRITZKY, LEONID MITIN AND VLADIMIR POPOV
Popov, V, and Bezborodov, A. (1999) Industrial technology for microbiological abatement of air emissions. Appl. Biochem. Microbiol. (Russia) 35: 570-577. Popov, V.O., Bezborodov, A.M., Cross, P. and Murphy, A. (2000) Industrial trickling bed biofilters for abatement of VOCs from air emissions. In (ed. Ed.D.Wise e.a.) Remediation of Hazardous Waste Contaminated Soils, Marcel Decker Inc., 2nd ed., pp 449-473. Popov, V.O., Bezborodov, A.M., Cavanagh, M. and Cross, P. (2004) Evaluation of industrial biotrickling filter at the flexographic printing facility. Env. Prog. 100: 39-44. Popov, V., Khomenkov, V., Zhukov, V., Cavanagh, M. and Cross, P. (2005) Design, construction and long-term performance of novel type of industrial biotrickling filters for VOC and odour control. In: (Kennes C. and Veiga M.C., Eds), International Congress on Biotechniques for Air Pollution Control, La Coruòa, Spain, pp. 257–262. Popov, V.O. and Zhukov, V. (2005) Odour removal in industrial facilities. In (Shareefdeen Z. and Singh A., Eds.), Biotechnology for Odour and Air Pollution Control. Springer-Verlag, Heidelberg-New-York, Germany, pp. 305-326. Zhukov, V.G., Rogojin, I.S., Ushakova, N.A., Zagustina, N.A., Popov, V.O. and Bezborodov, A.M. (1998) Development of the technology of air deodouration and its field testing with the use of a pilot plant. Appl. Biochem. Microbiol. (Russia) 34: 370-376.
67
Continuous monitoring of odours at a biofilter outlet S ELENA S IRONI 1 , L AURA C APELLI 1 , P AOLO C ÉNTOLA 1 , R ENATO D EL R OSSO 1 , MASSIMILIANO IL GRANDE2 1
Olfactometric Laboratory, Department CMIC «Giulio Natta», Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milan, Italy 2 Progress S.r.l., Via N.A. Porpora 147, 20131 Milan, Italy
ABSTRACT The experience matured in the field of biofiltration applied to odour removal enabled to study the correlation between some of the characteristic parameters of the biofilter bed (e.g. T, RH, ΔP) with the emitted odour concentration. Today odour measurement can be performed only by classical olfactometry. Classical olfactometry is an expensive and time-consuming method that is not suitable for continuous monitoring as needed by operators of compost facilities. This paper describes the experimental approaches adopted for the development of a system for the continuous monitoring of odour emissions, i.e. an instrument for the repeated air analysis, capable of qualitatively and quantitatively recognizing odours. This work shows the results of the first experiments carried out with the purpose of developing an electronic nose to be applied at a biofilter outlet for the real-time odour concentration measurement and for the detection of the exceeding of a given odour «alarm threshold».
1 INTRODUCTION Biofiltration is a process commonly used for the removal of odorous compounds (Devinny et al., 1999). The biofilter efficiency depends on several key parameters such as the moisture content of the medium (Van Lith et al., 1997; Leson and Winer, 1991; Quiniam et al., 1999; Krailas and Pham, 2002; Morales et al., 2003), temperature (Yang and Allen, 1994) and pH (Devinny et al., 1999). Typical organic media may be 40-80% water (by weight) when saturated. The recommended water content is commonly evaluated at 50% of the water-holding
67
68
SELENA SIRONI, LAURA CAPELLI, PAOLO CÉNTOLA, RENATO DEL ROSSO
AND
M. IL GRANDE
capacity of the material. The rough surface and porosity provide extensive microbial habitats that may improve colonisation and bacterial diversity (Delhomenie et al., 2002). Most biofilters are colonised with mesophilic microbial communities, that is to say microbes that thrive at temperatures between 20 and 45°C (Darlington et al., 2001). In general for this temperature range, enzymatic activity increases as a function of temperature. Finally most biofilters are designed to operate in a near-neutral pH range (6-8) (Mac Nevin and Barford, 2001). Also the nature of the packing material is of great importance for the service life of the biofilter, microbial growth, removal performance and operational cost. Thus, the choice of biofilter medium (size of particles and their organic or inorganic nature) is a fundamental factor in the successful running of biofilters (Krailis et al., 2000; Tawil, 2001). Biofilters are widely used as odour abatement systems at different kinds of industrial plants. Odour control at the outlet of biofilters is very important, especially in the case of municipal solid waste (MSW) treatment plants, because odour emissions from this kind of plants are subject to limits fixed by the competent authority. The efficiency of biofilters in odour removal can be evaluated by monitoring the above mentioned parameters or by the direct analysis of odour concentration. Dynamic olfactometry is the technique commonly used for odour measurements at biofilters inlets and outlets, but olfactometric analyses are expensive and time consuming, and they can’t be carried out continuously. The aim of this paper is to describe the approaches adopted for developing an electronic nose suitable for the continuous monitoring of the odour concentration at a biofilter outlet, with the purpose of getting over the discontinuity bound to dynamic olfactometry. Nonetheless, the development of such a system requires a validation, which might be achieved by the execution of periodical olfactometric analyses for verifying the instrument accuracy in odour concentration determination.
2 MATERIALS AND METHODS 2.1 OLFACTOMETRIC ANALYSES The collection of samples to be analyzed by dynamic olfactometry on the biofilter outlet is carried out by means f a static hood, which has the function of isolating the sampling point from the external conditions, and to channel the air stream in a stack from where the sample is collected with a depression pump inside a NalophanTM bag with a TeflonTM inlet tube (Bockreis and Steinberg, 2005) (Figure 1).
CONTINUOUS MONITORING OF ODOURS AT A BIOFILTER OUTLET
69
The odour concentration of the samples collected at the biofilter outlet is determined by dynamic olfactometry, in conformity with the European Norm EN 13725:2003. The odour concentration is expressed in European odour units per cubic metre (ouE/m3), and it represents the number of dilutions with neutral air that are necessary to bring the concentration of an odorous sample to its odour perception threshold concentration. The analysis is carried out by presenting the sample to a group of selected panellists at increasing concentrations by decreasing serial dilutions, until the panel members perceive an odour that is different from the reference neutral air. The odour concentration is then calculated as geometric mean of the odour threshold values of each panellist, multiplied by a factor that depends on the olfactometer dilution step factor. An olfactometer ECOMA model TO8, based on the «yes/no» method, was used as dilution device (Figure 2). This instrument with aluminium casing has 4 panellists places in separate open boxes. Each box is equipped with a stainless steel sniffing port and a push-button for «yes» (odour threshold). The measuring range of the TO8 olfactometer goes from a minimum dilution factor of 4(=22) to a maximum dilution factor of 65536(=216), with a dilution step factor 2. All the measurements were conducted within 30 hours after sampling, relying on a panel composed by 4 panellists. The odour concentration was calculated as geometric mean of the odour threshold values of each panellist, multiplied by 2 .
Figure 1. Equipment for odour sampling at a biofilter outlet.
70
SELENA SIRONI, LAURA CAPELLI, PAOLO CÉNTOLA, RENATO DEL ROSSO
AND
M. IL GRANDE
Figure 2. The ECOMA TO8 olfactometer.
2.2 D EVELOPMENT
OF A SYSTEM FOR THE CONTINUOUS ODOUR CONCENTRATION
DETERMINATION
2.2.1 GENERAL PRINCIPLES The development of this system is based on the use of an already existing technology, i.e. the electronic nose. The first prototype of electronic nose was described by Persaud and Dodd in 1982, and it consists of an instrument which comprises an array of electrochemical sensors with partial specificity and an appropriate pattern recognition system, capable of recognizing simple or complex odours. Even though electronic noses are studied since several years, most studies concern applications that are very different from the one proposed in this work. For these reasons, a complete re-design of the instrument is required in order to make it suitable for the specificities associated with the application in the environmental sector. The work required for the development of a similar system is composed of two fundamental activities, which are strictly interconnected, i.e. the instrument design and the definition of its utilization procedures. The instrument design comprises the following aspects: the choice of the gas sensors; the implementation of a suitable software for the instrument operation and the data acquisition and processing, and finally the study of technical features which are needed in order to make the instrument usable not only in laboratory but also in the field. Parallel to the instrument design activity it is extremely important to define the electronic nose utilization procedures.
CONTINUOUS MONITORING OF ODOURS AT A BIOFILTER OUTLET
71
The use of an electronic nose provides to relate an unknown «match data set» to a «training data set» acquired by the instrument in the so-called «training» phase, during which a number of odorous samples are analyzed in order to create the database of odour concentration values used as a reference by the instrument for further odour concentration estimation. The principal utilization procedures to be defined concern therefore the instrument settings, the training methods and the data processing methods. 2.2.2 ELECTRONIC NOSE DESCRIPTION The instruments used for this study have been developed in collaboration with Sacmi s.c.a.r.l. and the Sensor Laboratory of the University of Brescia (Falasconi et al., 2005) (Figure 3).
Figure 3. Electronic nose used for the study.
The system includes a pneumatic assembly for dynamic sampling (pump, electro-valve, and electronic flow meter), a thermally controlled sensor chamber with 35 cm3 of internal volume and an electronic board for controlling the sensor operational conditions. The electronic noses have been equipped with an array of six thin film MOS (Metal Oxide Semiconductor) sensors, which make the system sensitive to a large spectrum of volatile compounds, and a humidity sensor. For the analyses, the carrier flow rate was 150 cm3 min-1 and the temperature of the sensor chamber was kept constant at 50°C.
72
SELENA SIRONI, LAURA CAPELLI, PAOLO CÉNTOLA, RENATO DEL ROSSO
AND
M. IL GRANDE
2.2.3 TRAINING FOR ODOUR QUANTIFICATION The quantification of odour concentration by means of an electronic nose requires a particular training: odorous gas samples must be diluted at different odour concentration values and analyzed in order to create a database that can be used for the estimation of the odour concentration of unknown air samples by interpolation of the training points. For this reason, after collection, the odour concentration of the samples is measured by dynamic olfactometry. Once the odour concentration of the samples is known, each sample is diluted at different odour concentration values, by means of the same dilution device used for the olfactometric analyses (olfactometer), in order to obtain samples with odour concentration values included in the typical odour concentration range of odorous ambient air (20-1000 ouE m-3) (Sironi et al., 2007). 2.2.4 INSTRUMENTAL SENSITIVITY TOWARDS ODOUR CONCENTRATION VARIATIONS In order to develop a system for the continuous monitoring of odours at a biofilter outlet it is necessary to evaluate its sensitivity towards odour concentration variations. For this reason, we are currently carrying out a set of experimental tests, in order to evaluate how different sensors respond at the analysis of odorous gas samples at different odour concentration values. The instrumental sensitivity is tested using different typologies of odours: pure compounds and gas samples collected at biofilter outlets. As pure compounds, we decided to consider ammonia (NH3) and hydrogen sulphide (H2S), because these are the compounds for which an concentration limit is fixed by the competent authority at the outlet of biofilters installed at plants for the mechanical biological treatment of MSW.
3 RESULTS AND DISCUSSION This paragraph reports the results of some of the studies conducted with the aim of developing an electronic nose suitable for the continuous monitoring of odours at a biofilter outlet. More in detail, the results reported concern the studies relevant to the experimental verification of the instrumental sensitivity towards odour concentration variations. It is worth to remember that this results are preliminary and therefore partial, because our work in this field is still in progress. As mentioned in the previous paragraph, the sensitivity of different sensors is tested using samples of pure compounds and of gaseous mixtures collected at biofilter outlets. As an example, we consider the tests conducted with a set of samples at different odour concentration values (22, 34, 44 and 110 ouE/m3) obtained by dilution of a
CONTINUOUS MONITORING OF ODOURS AT A BIOFILTER OUTLET
73
sample collected at the outlet of a biofilter installed at a plant for the mechanical and biological treatment of MSW. Figure 4 shows the responses of a Nickel Oxide sensor (p-type) relevant to the analyses of the above mentioned samples. It is possible to observe that the amplitude of the response curves increases with the odour concentration values. Based on this observations it is possible to affirm that the studied sensor is sensitive even to very small variations of the odour concentration (i.e. 10 ouE/m3).
Figure 4. Response curves relevant to the analysis of gas samples at different odour concentration values.
Figure 5 shows a PCA relevant to the same test. It is possible to observe that the odour concentration values grow along the dircetion of the first principal component.
74
SELENA SIRONI, LAURA CAPELLI, PAOLO CÉNTOLA, RENATO DEL ROSSO
AND
M. IL GRANDE
Figure 5. PCA relevant to the analysis of gas samples at different odour concentration values.
The data relevant to the analyses of the samples at different odour concentration values obtained by dilution of a sample collected at a biofilter outlet can be splitted in two, a test training set composed by 5 measures, and a test match set composed by the remaining 5 measures. The obtained test training set is used as a reference for the estimation of the odour concentration values of the test match set, in order to evaluate the instrument accuracy in odour quantification. Table 1 shows the results of this test, i.e. the odour concentration values attributed by the instrument to the measures that form that test match set. The extrapolation of statistical data from the test results, e.g. mean, maximum and minimum error, and correlation index between true and estimated values, allows to gain some information about the estimation accuracy (Table 2). The calculated correlation index is equal to 0.99923, indicating a very good correlation between estimated and true and estimated values.
75
CONTINUOUS MONITORING OF ODOURS AT A BIOFILTER OUTLET
Table 1. Results of the odour concentration estimation test. N. 1 2 3 4 5
Match Value 19.46 35.25 32.31 43.41 109.48
True Value 22 34 34 44 110
Error -2.54 1.25 -1.69 -0.59 -0.52
Table 2. Statistical parameters relevant to the odour concentration estimation test. Mean absolute error Max error Min error Correlation index
1.3173 2.5398 0.51962 0.99923
4 CONCLUSIONS Our current work consists in the execution of a large number of tests with the aim of studying the optimal combination of sensors for the development of an electronic nose for the continuous monitoring of odours at a biofilter outlet. The most important characteristics for such a system are high sensitivity and accuracy. The studies we already conducted allowed to exclude some sensors, with too low sensitivity or stability, considering that sensors to be applied for real-time odour concentration measurements should be very sensitive even to very small odour concentration variations (<10 ouE/m3), whereas for the development of an instrument to be used just as an alarm for the exceeding of a given odour concentration threshold it might be sufficient to use sensors with a lower sensitivity to odour concentration variations (40-50 ouE/m3). Our future work will consists in the realization of the described system for the continuous monitoring of odours at a biofilter outlet and, after an accurate training phase, in its application on field. This will allow to overcome the difficulties associated with the intrinsic discontinuity of the olfactometric analyses. Nonetheless, because of the reliability of dynamic olfactometry, this technique will still be needed for the instrument validation, by the execution of periodical contemporaneous analyses by dynamic olfactometry and electronic nose, in order to evaluate the instrument accuracy in the odour concentration estimation.
76
SELENA SIRONI, LAURA CAPELLI, PAOLO CÉNTOLA, RENATO DEL ROSSO
AND
M. IL GRANDE
REFERENCES Bockreis, A. and Steinberg, I. (2005) Measurement of odour with focus on sampling techniques. Waste Manage. 25: 859-863. Darlington, A.B., Dat, J.F. and Dixon, M.A. (2001) The biofiltration of indoor air: air flux and temperature influences the removal of toluene, ethylbenzene and xylene. Environ. Sci Technol. 35: 240-246. Delhomenie, M.C., Bibeau, L. and Heitz, M. (2002) A study of the impact of particle size and adsorption phenomena in a compost based biological filter. Chem. Eng. Sci. 57: 4999-5010. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control, Lewis Publisher Ed. EN 13725 (2003) Air quality - Determination of odour concentration by dynamic olfactometry, Comité Européen de Normalisation, Brussels, Belgium. Falasconi, M., Pardo, M., Sberveglieri, G., Riccò, I. and Bresciani, A. (2005) The novel EOS835 electronic nose and data analysis for evaluating coffee ripening. Sens. Actuators B Chem. 110: 73-80. Krailas, S. and Pham, Q.T. (2002) Microkinetic determination and water movement in a downward flow biofilter for methanol removal. Biochem. Eng J. 10: 103-113. Krailis, S., Pharm, Q.T., Amal, R., Jang, J, K. and Heitz M. (2000) Effects of inlet mass loading, water and total bacterial count on methanol elimination using upward flow biofilter. J. Chem. Technol. Biotechnol. 75: 299-305. Leson, G. and Winer, A.M. (1991) Biofiltration: an innovative air pollution control technology for VOC emissions. J. Air Waste Manage Assoc. 41: 1045-1054. Morales, M., Hernández, S., Cornabe, T., Revah, S. and Auria, R. (2003) Effect of drying on biofilter performance modelling and experimental approach. Environ. Sci. Technol. 37: 985-992. Mac Nevin, D. and Barford, J. (2001) Inter-relationship between adsorbtion and pH in peat biofilters in the context of a cation-exchange mechanism. Water Res. 35: 736-744. Persaud, K. and Dodd, G.H. (1982) Analysis of discrimination mechanisms in the mammalian olfactory system using a model nose. Nature 299: 352-355. Quiniam, C., Strevett, K. and Ketcham, M. (1999) VOC elimination in a compost biofilter using a previously acclimated bacterial inoculum. J. Air Waste Manage Assoc. 49: 544-553. Sironi, S., Capelli, L., Céntola, P. and Del Rosso, R. (2007) Development of a system for the continuous monitoring of odours from a composting plant: Focus on training, data processing and results validation methods. Sens Actuators B Chem. 124: 336-346. Tawil, A.A. (2001) Bioxidation of n-propanol and acetone in biofilters and bioscrubbers, PhD Thesis University College Dublin. Yang, Y. and Allen, E.R. (1994) Biofiltration control of hydrogen sulphide design and operational parameters, J. Air Waste Manage Assoc. 44: 863-868. Van Lith, C., Leson, G. and Michelsen, R. (1997) Evaluating design options for biofilters. J. Air Waste Manage Assoc. 47: 37-48.
77
Removal of odour and ammonia in ventilation air from growing-finishing pig units using vertical biofilters ANDERS LEEGAARD RIIS Danish Pig Production, Axeltorv 3, 1609 Copenhagen V, Denmark
ABSTRACT The aim of this study was to investigate the removal of odour and ammonia from outlet air using vertical biofilters in two units with growing-finishing pigs in the winter. Woodchips were used as media in the wall of the biofilters. The air from the pig units was humidified by a high-pressure water system before it reached the biofilters. A total of 56 odour and ammonia measurements were taken at an average outdoor temperature of 5.4 C. The biofilters significantly reduced the odour concentration (OUE/m3) in the outlet air (P<0.001). The measured odour removal efficiency averaged 60 %. In contrast, the biofilters did not reduce the ammonia concentration (ppm) significantly in the outlet air. The hedonic tone of the odour of the air was determined before and after the biofilter. The untreated air was recorded as more unpleasant than the air that had passed through the biofilters. In conclusion, the biofilters were capable of reducing the odour concentration in the outlet air from units with growing-finishing pigs in the winter. The biofilters’ treatment of the air made the odour less unpleasant. However, the biofilters were not capable of reducing the ammonia concentration in the outlet air in the winter.
1 INTRODUCTION In recent years, there has been a growing interest in reducing odour and ammonia from pig production in Denmark. This interest has been heightened by more stringent environmental regulations in Denmark, which were most recently intensified in January 2007 (Danish Ministry of the Environment, 2006). This has resulted in several companies developing technologies for odour and ammonia reduction of air from pig
77
78
ANDERS LEEGAARD RIIS
units. In the United States, the research has mainly been on biological air treatment. Most of the focus has been on horizontal biological air filters in which the filter material consists of a mixture of woodchips and compost (Nicolai and Schmidt, 2004). However, these solutions have a large space requirement, and experiences have shown that they are a popular place for rodents. Therefore, research in the United States has recently focused on vertical biofilters, where the concept of woodchips and compost are used as filter material (Nicolai et al., 2005). In comparison with the horizontal biofilters, the filter material is put into the wall of a vertical silo. The air from the pig facility enters the vertical biofilter and is then forced through the wall of the biofilter. Vertical biofilters reduce the space requirement and minimize the risk of rodents invading the biofilter. The present study was carried out to investigate the removal of odour and ammonia from outlet air using vertical biofilters in two units with growing-finishing pigs in the winter.
2 MATERIALS AND METHODS 2.1 EXPERIMENTAL SETUP Two units, each housing 180 growing-finishing pigs, with partially slatted floors were used. Two vertical biofilters were placed outside the two units. The ventilation system was based on the principle of negative pressure ventilation. Fresh air entered the unit through a diffuse inlet in the ceiling. The outlet air was sucked out through the roof and into ventilation channels which were connected to the biofilters. The biofilters were round-shaped and hollow with an outer diameter of 3.90 m and a height of 4.60 m. The ventilation fan was placed in the channel directly before the filters, and the outlet air was forced through the filters. Woodchips were used as media in the wall of the biofilters (Figure 1).
Figure 1. Design of the vertical biofilters.
REMOVAL OF ODOUR AND AMMONIA IN VENTILATION AIR
79
The wall of the biofilters was 0.70 m wide at the top and 0.30 m wide at the bottom to avoid variable airflow through the biofilter wall due to settling of the media over time. The design was in accordance with recommendations by Nicolai et al. (2005). Before the air from the pig units reached the inside of the biofilters, it was humidified by a high-pressure water system (Agrofilter GmbH, Alfstedt, Germany) to ensure that the woodchips were moistened. As the outlet air passed through the wall of the biofilters, dust, ammonia and odour compounds in the air stream were degraded and metabolized by the biofilm on the woodchips. The biofilters were set up by September 1st 2006. The measurement period went from February to March 2007 and included one batch of pigs. A total of 56 odour and ammonia measurements were taken on seven days spread over a period of four weeks during the measurement period. 2.2 ANALYTICAL METHODS The odour samples were collected in 30 L Tedlar® odour bags. The bags were placed in an airtight container and filled by creating a vacuum in the airtight container with a pump. On each of the seven days of measurement, two pairwise odour samples were taken from the air stream before and after each biofilter, respectively. The first pairwise odour samples were taken between 11.00 a.m. and 11.30 a.m. and the second pairwise odour samples were taken between 12.30 p.m. and 1.00 p.m. The collection of odour samples and analyses of odour concentration (OUE/m3) were performed in compliance with European olfactometric standard EN:13725 (CEN, 2003). The ammonia concentration (ppm) was measured in the air stream before and after the filters using Kitagawa gas detector tubes 105SD (Komyo Rikagaku Kogyo K.K., Kawasaki-City, Japan). The relative humidity in the air stream was measured using a TSI VelociCalc Plus 8386 instrument (TSI Incorporated, Minnesota, U.S.A.). On the sixth day of measurement, one pairwise odour sample from the air stream before and after one biofilter was sent for analysis of hedonic tone of the odour. The hedonic tone describes the pleasantness or the unpleasantness of a given odour. The hedonic tone was determined at the concentration level of 1 to 15 OUE/m3. The determination of the hedonic tone was performed in compliance with the German guidance, VDI 3882, sheet 2. The ammonia concentration and the logarithmically transformed odour concentration were processed with an analysis of variance in the MIXED Procedure in SAS (SAS Inst. Inc., Cary, NC).
3 RESULTS AND DISCUSSION During the days of measurement, the weight of the pigs increased from 43.1 to 69.8 kg. The measurements were taken at an average outdoor temperature of 5.4 °C and
80
ANDERS LEEGAARD RIIS
ranged from between 2.9 and 8.4 °C. The biofilters significantly reduced the odour concentration (OUE/m3) in the outlet air during the winter period (P<0.001), and there was no statistical difference between the two biofilters regarding efficiency. The measured odour removal efficiency averaged 60 % (95 % confidence limits: 53 - 67). However, the removal efficiency in per cent depended on the odour concentration in the outlet air with high removal efficiency at a high odour concentration versus low removal efficiency at a low odour concentration. The odour concentration in the outlet air from the pig units generally increased over time. The relatively high variation in odour concentration in the pig facility during the period was apparently due to the growth of the pigs, which resulted in more faeces on the floors and walls and thereby an increasing odour concentration. The mean odour concentration in the outlet air from the pig units was 2065 OUE/m3 and ranged from between 1034 and 3604 OUE/ m3 during the measurements. However, after the air had passed through the biofilters, the odour concentration was reduced to an average of 783 OUE/m3 and ranged from between 418 and 1447 OUE/m3 during the measurements (Figure 2).
Figure 2. Odour concentration in the air from the pig units before and after it has passed through the biofilters.
In relation to the odour removal efficiency, the results of the hedonic tone showed that the odour panel perceived the odours as unpleasant to various extents since higher
REMOVAL OF ODOUR AND AMMONIA IN VENTILATION AIR
81
concentrations were perceived as more unpleasant (Figure 3). However, the untreated air was recorded as more unpleasant than the air that had passed through the biofilters. In contrast, the biofilters did not reduce the ammonia concentration (ppm) significantly in the outlet air during the winter period. This was in contrast to Jensen et al. (2005) and Riis et al. (2006), who found an ammonia reduction with the use of biofilters. The average measured ammonia concentration in the outlet air from the pig units was 6.0 ppm and in the range of 3.5 to 9.5 ppm.
Figure 3. Hedonic tone of the odour in the air from the pig units before and after it has passed through the biofilters.
During the measurement period, the high-pressure water system ran for 8.5 hours a day in relation to the timetable in Table 1. The consumption of water was 170 litres of water added to the air of each biofilter per day. However, the consumption of water was no more than the amount of water that continuously evaporated from the biofilters. This means that there was no accumulation of water in the bottom of the biofilters.
82
ANDERS LEEGAARD RIIS
Table 1. Timetable for the high-pressure water system during the measurement period (February to March 2007). Start 1.00 AM 7.00 AM 9.00 AM 11.00 AM 4.00 PM 7.00 PM
Stop 2.00 AM 8.00 AM 10.00 AM 2.30 PM 5.00 PM 8.00 PM
In Figure 4, the relative humidity in the air is shown for the air before and after it has passed through the biofilters. The average measured relative humidity in the air before it reached the biofilters was 67.3 % and ranged from between 58.3 and 76.0. After the air had passed through the biofilters, the relative humidity was increased to 86.6 % and ranged from between 72.3 and 97.4. Humidifying the air by the high pressure water system increased the relative humidity in the air by an average of 24.5 % (95 % confidence limits: 20.7 - 28.3). Appropriate humidification is necessary to achieve a good odour reduction with biofilters, which was also experienced by Jensen et al. (2005), Riis et al. (2006) and Nicolai and Smith (2004).
Figure 4. Relative humidity in the air from the pig units before and after it has passed through the biofilters.
REMOVAL OF ODOUR AND AMMONIA IN VENTILATION AIR
83
In conclusion, the vertical biofilters with woodchips as media were capable of reducing the odour concentration in the outlet air from units with growing-finishing pigs during the winter period. Based on the measurements of the hedonic tone, the biofilters treatment of the air made the odour less unpleasant. However, the biofilters were not capable of reducing the ammonia concentration in the outlet air in the winter.
4 ACKNOWLEDGEMENTS The vertical biofilters were designed in cooperation with Dick Nicolai, South Dakota State University.
REFERENCES CEN (2003) Air quality determination of odour concentration by dynamic olfactometry (EN13725). Brussels, Belgium: European Committee for Standardization. Danish Ministry of the Environment (2006) Lov om miljøgodkendelse m.v. af husdyrbrug. http:// www.ft.dk/?/Samling/20061/lovforslag/L55/index.htm (June 22, 2007). Jensen, T.L., Riis, B.L. and Feilberg, A. (2005) Reduction af lugt og ammoniak med Oldenburg biofilter, Agrofilter GmbH. Meddelelse nr. 727. Landsudvalget for Svin. Nicolai, R.E., Lefers, R. and Pohl, S.H. (2005) Configuration of a vertical biofilter. Proceedings of the Seventh International Symposium Livestock Environment VII. Beijing, China: 358-364. Nicolai, R. and Schmidt, D. (2004) Biofilters. Extension Fact Sheet 925-C. Cooperative Extension Service, Ag & Biosystems Engineering. South Dakota State University. 18 p. Riis, B.L., Hansen, A.G. and Jensen, T.L. (2006) Luftrensning til stalde. Videreudvikling af biofiltre til eksisterende staldanlæg med akutte lugtproblemer. –Pilottest ved svinestald. Arbejdsrapport fra Miljøstyrelsen Nr. 31. Miljøministeriet.
This page intentionally left blank
85
Membrane inlet mass spectrometry (MIMS) as a tool for evaluating biological air filters in agriculture ANDERS FEILBERG Danish Technological Institute, Teknologiparken, Kongsvang Allé 29, DK – 8000 Aarhus, Denmark
ABSTRACT Membrane inlet mass spectrometry (MIMS) is presented as a new tool for monitoring the removal efficiency of biofilters with respect to odour compounds. The MIMS technique is based on the separation of volatile chemicals and gases from an air stream by a thin silicone membrane adjacent to the ion source of a standard quadropole mass spectrometer. The vacuum conditions of the MS forces the separated compounds to diffuse through the membrane and evaporate in the ion source. The compounds are detected by MS by means of specific molecular ions or fragment ions. It is possible to monitor a number of individual compounds or compound groups contributing to the odour nuisance of livestock facilities. 4-Methylphenol (p-cresol), skatol, indol, 4-ethylphenol, phenol and dimethyltrisulfide give rise to specific signals corresponding to molecular fragments. A variety of carboxylic acids can be detected by signals corresponding to three subgroups of this compound group. The sum of reduced organic sulphur compounds (ROS) are measured by a common signal. The contribution of individual sulphur compounds (mainly methanethiol and dimethyl sulphide) to ROS can be estimated from supplementary MS fragments. MIMS is not suitable for measuring ammonia and hydrogen sulphide for which other methods must be applied. MIMS is suitable for continuous monitoring on-site and has been applied for evaluation of a number of biofilters in the Danish agricultural sector, primarily for treating ventilated air from pig barns.
85
86
ANDERS FEILBERG
1 INTRODUCTION Odour nuisance is an increasing problem for the local population in many areas due to intensified livestock production. In order to maintain a cost-effective agricultural production in accordance with the local society, there is therefore a need for efficient odour abatement technologies. A number of technologies are proposed, developed and implemented, e.g. biofilters and chemical scrubbers. Testing, evaluating and optimizing such technologies require reliable and accurate methodologies for measuring odour and the efficiencies of odour abatement technologies. Traditionally, the efficiency of odour reduction has been measured by olfactometry based on the odour perception of a trained human panel. Olfactometry is of relevance because it is a direct measure of the actual physiological effect of odour mixtures. However, it is generally accepted that there is substantial uncertainty associated with the measurements. This is particularly a disadvantage in relation to technology optimization where small improvements thus cannot be identified with certainty. Chemical measurements of odour compounds can in principle be carried out with great accuracy. Provided that the major odour compounds can be identified and the relationship between odour compounds and odour perception can be established with a reasonable statistical certainty, chemical measurements therefore has the potential to supplement and partly replace olfactometry. Measurements based on gas chromatography with mass spectrometric detection (GC/MS) has been used to detect and identify odour compounds from manure and livestock facilities (Blunden et al., 2005; Hobbs et al., 1998; Schiffman et al., 2001). Here, we present a related technique, membrane inlet mass spectrometry (MIMS) (Ketola et al., 1997; Ketola et al., 2002). Compared to GC/MS, MIMS is a direct measurement technique, which allows for on-line measurements of odour compounds. Examples of measurements include monitoring the efficiency of biofilters and the effect of cooling ventilation air.
2 MATERIALS AND METHODS A mass spectrometer (Balzers, QMG 420) equipped with a membrane inlet consisting of a temperature controlled membrane (50 μm polydimethyl siloxane) and a stainless steel air intake was used for the measurements presented in this paper. An air pump was used to pass unfiltered air by the membrane at a flow rate of ~200 ml/min providing turbulent conditions near the membrane surface, which gives the most efficient uptake of volatile compounds. Odour compounds are sampled from the air by absorption into the membrane, and because of the low pressure inside the mass spectrometer compounds diffuse through the membrane and evaporate into the ion source of the mass spectrometer.
MEMBRANE INLET MASS SPECTROMETRY (MIMS) AS A TOOL FOR EVALUATING BIOLOGICAL
87
In order to monitor the efficiency of odour abatement technologies, a multiposition valve was used to multiplex between air influenced by the technology, reference air and background (clean) air. In case of a biofilter, for example, a measurement cycle consists of 1) clean air, 2) air sampled before the filter and 3) air sampled after the filter. The efficiency with respect to a certain compound is estimated from the background-subtracted signals before and after the filter, respectively.
3 RESULTS AND DISCUSSION 3.1 BASIC STUDIES OF MIMS RESPONSE TO ODOUR COMPOUNDS In Figure 1, a mass spectrum obtained by sampling the headspace of manure from a Danish pig farm is presented. The sample represents a concentrated odour sample compared to typical emission levels from pig farms and is useful for evaluating which compounds contribute to the main MIMS-signals.
Figure 1. Assignment of MIMS-signals to odour compounds or groups of odour compounds. The x-axis is the mass/charge ratio (m/z) of the ions formed.
It can be seen from Figure 1, that several of the major MIMS-signals can be assigned to important odour compounds or structurally related groups of compounds. Some signals, e.g. from m/z 50 – 58 and from m/z 77 – 80 are less specific and therefore less useful for monitoring odour compounds. The composition of the sample was confirmed by GC/MS. There are a number of carboxylic acids present in typical samples from pig facilities. These give rise to overlapping MIMS-signals at m/z 60, 73 and 74. Therefore carboxylic acids are mainly measured as groups of compounds according to these signals. Likewise, reduced organic sulphur compounds are detected by a common
88
ANDERS FEILBERG
signal at m/z 47 corresponding to compounds containing the CH3S- moiety, which mainly comprise methyl mercaptan, dimethyl sulphide and dimethyl disulphide. These sulphur compounds give rise to other MS-fragments which can be used to estimate the contributions of individual sulphur compounds to m/z 47. Other odour compounds, like skatol and indol, possess more specific MS signals. The sensitivity of MIMS is compound specific and is related to the volatility and polarity of the compound and hence its ability to be absorbed into the non-polar membrane. Carboxylic acids, for example, are highly abundant (as quantified by GC/ MS) but do not give rise to the largest MIMS signals due to their high polarity. An important feature of MIMS is the linearity of the response as a function of concentration. In Figure 2, examples of calibration curves for carboxylic acids (RCOOH; the sum of carboxylic acids containing the mass spectrometric fragment with m/z 60) and p-cresol (4-methylphenol) are shown. The data in Figure 2 was obtained by injecting various amounts of a synthetic mixture of 10 important odour compounds into a Tedlar sampling bag and analyzing by MIMS. The concentration levels were selected in order to cover real-world concentration levels.
Figure 2. Calibration curves of RCOOH (carboxylic acids) and p-cresol. The straight curves are linear regression data.
3.2 APPLICATION OF MIMS FOR MONITORING ODOUR REDUCTION An example of results from the application of MIMS for monitoring the efficiency of a biofilter is presented in Figure 3. The biofilter tested is a vertical cellulose pad biofilter with a relatively short residence time (< 1 second). Recirculated water is applied on top of the filter matrix and flows down mainly on the inlet side of the filter. It appears that the efficiency of the filter towards different compounds is quite stable, whereas there is more variation regarding the efficiency towards different types of compounds. P-cresol and carboxylic acids are efficiently removed, whereas organic
MEMBRANE INLET MASS SPECTROMETRY (MIMS) AS A TOOL FOR EVALUATING BIOLOGICAL
89
sulphur compounds are less reduced and with much more variability in the results. These results are not surprising since it is a prerequisite for efficient removal that the odour compounds are absorbed into the aqueous phase of the biofilter. Organic sulphur compounds, such as dimethyl sulphide, are much less water soluble than carboxylic acids or p-cresol.
Figure 3. Biofilter efficiencies towards p-cresol (black circles), RCOOH (open circles) and reduced organic sulphur compounds (ROS; grey circles). Data from March 2005. The efficiency is estimated from the MIMS signals before and after the filter.
Another example is shown in Figure 4 and 5. In this case a vertical filter consisting of two filter walls placed inside a large bin (2.5x10m) was tested. The filtermaterial was light expanded clay aggregates (LECA©) with a filter thickness of 15 cm and a residence time of approximately 5 seconds. Air is added between the two filter walls and is humidified with spray nozzles. Water is normally recirculated but during the measurements presented here, water was frequently changed and clean water added to the system.
90
ANDERS FEILBERG
Figure 4. Biofilter efficiencies towards reduced organic sulphur compounds (ROS; black circles), dimethyl sulphide (open circles) and skatol (grey circles). Data from February 2007. The efficiency is estimated from the MIMS signals before and after the filter.
Figure 5. Biofilter efficiencies towards carboxylic acids (black circles), 4-ethylphenol (open circles) and p-cresol (4-methylphenol; grey circles). Data from February 2007. The efficiency is estimated from the MIMS signals before and after the filter.
MEMBRANE INLET MASS SPECTROMETRY (MIMS) AS A TOOL FOR EVALUATING BIOLOGICAL
91
Similar to other filters tested, water soluble odour compounds such as carboxylic acids and phenols are consistently removed efficiently (80-100 %). In the LECA© filter type, however, it is documented that organic sulphur compounds are also removed with an efficiency of ~40-60 %. This is most likely due to a longer residence time combined with relatively clean water used for wetting the filters. 3.3 COMPARISON WITH OLFACTOMETRY A dataset consisting of olfactometric measurements and MIMS-measurements were subject to multivariate statistical analysis in order to attempt to elucidate a statistical relationship between odour reduction observed by olfactometry and odour reduction predicted from MIMS-measurements. The analysis was based on 23 data points from 3 different types of odour abatement technologies and resulted in a correlation coefficient (R2) of 0.56. Although this is statistically significant (P<0.001), it is also clear that a more extensive dataset is needed for this purpose. Part of the difficulty with this approach is the inherent uncertainty associated with olfactometry.
4 CONCLUSIONS Membrane Inlet Mass Spectrometry (MIMS) has been tested as a technique for continuous measurements of odour compounds in or emitted from pig facilities. MIMS is able to detect a range of compounds either as individual compounds or as groups of structurally related compounds. MIMS has been demonstrated to be sufficiently robust to be used on-site and to carry out measurements unattended for extended periods of time. Examples of MIMS-measurements include monitoring of the efficiencies of biofilters installed in pig facilities. A tentative comparison of MIMS with olfactometry indicates a potential for predicting odour reduction from MIMS-measurements, although more data is needed to confirm this.
5 ACKNOWLEDGEMENTS This work has been partly supported by the Danish Environmental Protection Agency and the Danish Ministry of Food, Agriculture and Fisheries.
92
ANDERS FEILBERG
REFERENCES Blunden, J.V.P. Aneja and Lonneman, W.A. (2005) Characterization of non-methane volatile organic compounds at swine facilities in eastern North Carolina. Atmos. Environ. 39: 6707-6718. Hobbs, P.J., Misselbrook, T.H. and Pain, B.F. (1998) Emission rates of odorous compounds from pig slurries. J. Sci.Food Agric. 77: 341-348. Ketola, R.A., Kotiaho, T., Cisper, M.E. and Allen, T.M. (2002 Environmental applications of membrane introduction mass spectrometry. J. Mass Spectrom. 37: 457-476. Ketola, R.A., Mansikka, T., Ojala, M., Kotiaho, T. and Kostiainen, R. (1997) Analysis of volatile organic sulfur compounds in air by membrane inlet mass spectrometry. Anal. Chem. 69: 4536-4539. Schiffman, S.S., Bennett, J.L. and Raymer, J.H. (2001) Quantification of odors and odorants from swine operations in North Carolina. Agric. Forest Meteorol. 108: 213-240.
93
MODELING AND TECHNOLOGICAL ASPECTS
This page intentionally left blank
95
Modeling of fungal biofilter for the abatement of hydrophobic VOCs ALBERTO VERGARA-FERNÁNDEZ1,2 AND SERGIO REVAH3 1
Escuela de Ingeniería Ambiental, Facultad de Ingeniería, Universidad Católica de Temuco, Manuel Montt 56, Casilla 15-D, Temuco, Chile 2 Departamento de Ingeniería de Procesos e Hidráulica, Universidad Autónoma MetropolitanaIztapalapa, Apdo. Postal 55-534, CP 09340, México DF, México 3 Departamento de Procesos y Tecnología, Universidad Autónoma Metropolitana-Cuajimalpa, c/o IPH, UAM-Iztapalapa, Av. San Rafael Atlixco No. 186, 09340 México, D. F., México
ABSTRACT This work describes the growth of filamentous fungi in biofilters for the degradation of hydrophobic VOCs. The study system was n-hexane and the fungus Fusarium solani B1. The system is mathematically described and the main physical, kinetic data and morphological parameters of aerial hyphae were obtained by independent experiments for model validation. The model proposed in this study describes the increase in the transport area by the growth of the filamentous cylindrical mycelia and its relation with n-hexane elimination in quasi -stationary state in a biofilter. The model describing fungal growth includes Monod-Haldane kinetic and hyphal elongation and ramification. The reduction in the permeability caused by mycelial growth was further related to pressure drop by Darcy’s equation. The model was verified with biofiltration experiments using perlite as support and gaseous n-hexane as substrate.
1 INTRODUCTION Biofiltration is one of the main techniques for the control of volatile organic compounds (VOCs) present in low concentrations in industrial gaseous emissions. The high flowrate of such emissions means that the investment and operating costs for conventional systems are high. In these systems, microorganisms fixed in a solid support oxidise the VOCs, principally to CO2 and water.
95
96
ALBERTO VERGARA-FERNÁNDEZ
AND
SERGIO REVAH
Due to the complexity of the system, resulting from its inherent heterogeneity and the diversity of the microbial populations which may become established, biofilters are difficult to model. The modelling of these systems involves physical and biochemical steps, liquid flow and diffusion, the properties of the microbial community and the solid support, prediction of the area and active thickness of the biofilm, etc (Bibeau et al., 2000). One of the main considerations when modelling biofilters is the assumption that the biomass and the liquid film which surrounds it form a single pseudo-homogeneous phase known as the biofilm. Ottengraf and van der Oever (1983) developed a solution to analyse the concentration profile in the biofilm and throughout the biofilter column to obtain the quantity of contaminant biodegraded in the biofilter, using first order and order zero growth kinetics. Shareefdeen and Baltzis (1994) developed a model for a fixed bed biofilter with transitory state operation for the treatment of toluene, implementing mass balances in the biofilm, the gas phase and the solid support, and using a Monod microbial growth kinetic. Hodge and Devinny (1995) and Jorio et al. (2003) developed a model using four different types of support material, to describe the mass transfer between the air phase and solid/water, the biodegradation of the substrate, CO2 production and changes in the pH as a result of CO2 accumulation. They also assumed that the filter medium and the distribution and density of the biomass in the biofilm is homogeneous and that the adsorption is reversible. The same assumptions are made by Deshusses et al. (1995), using a Monod type growth kinetic, with competitive inhibition for a mixture of methyl isobutyl ketone and methyl ethyl ketone. The mathematical model to describe the biofiltration of mixtures of hydrophilic and hydrophobic compounds used by Mohseni and Allen (2000) is based on the biophysical model proposed by Ottengraf and van den Oever (1983) for a VOC. The steady state model was developed considering the biofilm as an organic matrix and using Monod growth kinetics with inhibition. Iliuta and Laranchi (2004) describe the growth of the biofilm and its effect on the aerodynamics and clogging of the biofilter. The model considers a uni-directional flow based on the volumetric average of the balance of mass, momentum and species, linked to the conventional equations for diffusion/reaction in biological systems. Spigno et al. (2003) made a simple, steady state, axial dispersion model to evaluate the n-hexane elimination in a biofilter using the fungus Aspergillus niger. This model makes the same assumptions as those used for microbial consortia or bacterial biofilters, working with a constant, homogeneous fungus biofilm. The balance in the biological phase included a Monod type biodegradation with substrate inhibition. In general, it may be observed that the models which have been developed are based on the structure of a biofilm as a pseudo-homogeneous phase. However in the case of aerial biofilms, as those generated by filamentous fungi, this definition is not
MODELING OF A FUNGAL BIOFILTER FOR THE ABATEMENT OF HYDROPHOBIC VOCs
97
readily applicable. Although good results are obtained, they do not provide good information on the actual phenomena occurring in the inter-particular spaces of the biofilter and the growth characteristics of the fungus inside the biofilter. The objective of this work is to describe the growth of the filamentous fungi on a biofilters for the degradation of hydrophobic VOCs. The study system was n-hexane, as a model substrate and the fungus Fusarium solani B1.
2 MATHEMATICAL MODEL 2.1 DEFINITION OF THE SYSTEM The model proposed in this study describes the increase in the transport area by the growth of the filamentous cylindrical mycelia and its relation with n-hexane elimination in quasi -stationary state in a biofilter. To mathematically describe the system, we considered four processes: (1) mass transfer of VOCs in the bulk gas, (2) mass transfer of VOCs into the gas layer around the mycelium, (3) mass transfer and reaction of the nitrogen source through the elongating mycelia, (4) and the kinetic of mycelial growth. Processes (2) and (3) include movable boundary conditions to account for the mycelial growth. The model describing fungal growth includes Monod-Haldane kinetic (Shuler et al., 2003) and their elongation and ramification and is further related to macroscopic parameters such as pressure drop. The basic concept of the model develop was obtained using the Figure 1. 2.2 MECHANISM OF GROWTH The biomass and total length were determinate considering the principal hyphae and the branching.
(1)
2.3 MASS BALANCE IN THE BIOFILTER Unidirectional gas flow was considered:
(2)
98
ALBERTO VERGARA-FERNÁNDEZ
AND
SERGIO REVAH
Figure 1. Schematic representation of growth.
Axial boundary condition that consider the continuity of flux to the right of z = 0 and z = H (Illiuta and Larachi, 2004). (3)
(4)
To determine the flow regime around the hyphae, the criterion reported by Slattery (1999) was used and found that it can be considered creeping flow (Reynolds number around hyphae of 0.002).
MODELING OF A FUNGAL BIOFILTER FOR THE ABATEMENT OF HYDROPHOBIC VOCs
99
(5)
The boundary conditions incorporate the interaction in the gas-hyphae interphase. (6)
(7)
(8)
(9)
In the development of the mass balance of the nitrogen source, diffusion and reaction throughout hyphae were considered.
(10)
100
ALBERTO VERGARA-FERNÁNDEZ
AND
SERGIO REVAH
The boundary condition and the initial condition for this equation are: (11)
(12)
(13)
2.4 EFFECT OF THE GROWTH ON THE PRESSURE DROP The pressure drop was evaluated as a function of the reduction in the permeability caused by mycelial growth in the bioreactor and the Darcy´s equation.
3 MATERIALS AND METHODS 3.1 MICROORGANISMS AND INOCULUM Fusarium sp. was isolated as described by Arriaga and Revah (2005a). Its preservation, cultivation conditions and spore production was similar to reported for García-Peña et al. (2001). The biofilter was inoculated with a mineral medium solution and 2×107 spores mL-1. 3.2 CARBON SOURCES AND MINERAL MEDIUM The carbon source used was n-hexane (Baker, 98.5%). The mineral medium for fungi maintenance and cultivation was reported previously by Arriaga and Revah (2005a). 3.3 BIOFILTER SYSTEM The gas-phase biofilter consisted of a 1 m cylindrical glass column with inner diameter of 0.07 m, incubated at 30(±3)°C. The biofilter was packed with 250 g of
MODELING OF A FUNGAL BIOFILTER FOR THE ABATEMENT OF HYDROPHOBIC VOCs
101
perlite (bed void fraction of 68% and particle size of 3.4 – 4.8 mm) mixed with the mineral medium and the spore solution. Hexane-saturated air was mixed with moistened air and introduced at the top of the biofilter with a flow rate of 1.2 L min-1, with a residence time of 1.3 min, to reach an inlet n-hexane load of 325 g m-3 h-1. 3.4 ANALYTICAL METHODS Hexane concentration in biofilter system were measured with FID-GC and CO2 production by TCD-GC. The biomass in the perlite was measured as volatile solids with a thermogravimetric analyzer. Measurements were done in triplicate. The pressure drop was measured online by using pressure transducer with a data acquisition system online.
4 RESULTS AND DISCUSSION 4.1 VALIDATION OF THE MATHEMATICAL MODEL The Figure 2 and 3 shows the comparisons between the experimental results obtained in biofilter and model simulation. For the model simulation the data shown in Table 1 were used. 4.1.1 ELIMINATION CAPACITY (EC) Figure 2 compares the experimental data and the mathematical model for different cellular yield coefficients. In biological systems, when growth is uncoupled with the energy metabolism, the constant cellular yield for growth does not represent the reality of biomass production. This can be one of the explanations of the greater EC observed in Figure 2 with respect to the model prediction for low values of cellular yield (0.1 g-1g-1) when the growth in the fungi starts, existing differences of 2% between the experimental data and the model. Similarly it is possible to explain the low EC obtained with the simulation during the latency stage, greater cellular yield coefficients (0.8 g-1 g-1), existing differences of 12% between the experimental data and the model. In general, it is possible to observe an average deviation between the model and the experimental data for the EC of 7%. 4.1.2 PRESSURE DROP The Figure 3 shows the comparison of the experimental results of the pressure drop (ΔP) in the biofilter and the model simulation. The Figure 3 shows that in the first days of operation, before obtaining an important biomass growth, the ΔP determined by the model was in average 11% lower than experimental data, presumably due to the static liquid present in the bed, necessary to maintain the humidity of the biofilter, which was not considered in the model. On the other hand, it is possible to
102
ALBERTO VERGARA-FERNÁNDEZ
AND
SERGIO REVAH
observe that for the ΔP simulated after 30 days operation was obtained an average deviation of the experimental data of 3%. Table 1. Parameters used in simulations. Parameter Kinetic parameter KAH KNH KI mC μmax YN YA Morphological parameters λ γ dh Lav Lmax,m Lmax,B LC Lo NTB N0 ρh Physical-chemical parameters H εR (initial) dp Dg vg ρg μg DNH DDz kg HPC
Value
Unit
Reference
1.9 500 30.1 1.51×10-4 0.0518 2.546 0.824
g m-3 g m-3 g m-3 g g h-1 h-1 g g-1 g g-1
[1] [1] [1] [1] [1] [2] [2]
0.35 2.47 2.10 280.1 1477 452.1 665.6 8.34 7.0 1.0×104 1.1×10-9
——μm μm μm μm μm μm ——mg μm-3
[2] [2] [2] [2] [2] [2] [2] [2] [2] [2] [3]
1.0 0.685 0.004 0.029 48.91 1160 64.98 5.7×10-6 0.079 36.56 0.20
m m3 m-3 m m2 h-1 m h-1 g m-3 g m-1 h-1 m2 h-1 m2 h-1 m h-1 —-
—————[4] [4] ———[1]
[1] Vergara-Fernández et al. (2006), [2] Vergara-Fernández, (2007), [3] López-Isunza et al. (1997), [4] Hartmans y Tramper (1991).
MODELING OF A FUNGAL BIOFILTER FOR THE ABATEMENT OF HYDROPHOBIC VOCs
103
Figure 2. Experimental EC and model simulation. (•) Experimental results and (–) simulation. Variation of the EC simulation for a range of cellular yield in F. solani between 0.1 and 0.8 g-1g-1.
Figure 3. Experimental evolution of the pressure drop in the biofilter and model prediction. (•) Experimental data and (–) simulation.
104
ALBERTO VERGARA-FERNÁNDEZ
AND
SERGIO REVAH
5 CONCLUSIONS Growth of fungi and n-hexane elimination was modeled in biofiltration systems connecting a growth model based on microscopic parameters and the different mass balances describing the main transport phenomena occurring inside a biofilter. The independent evaluation of the parameters allowed a small deviation with experimental data below 10% for the elimination capacity and the pressure drop.
6 NOMENCLATURE av CAb CAG CAH CNH CNL dh DDz Dg DNH H kg KAH KNH KI Keq1 Keq2 Lav Lmax,m Lmax,B Lh Lh,Total mC NTB N0 VR VE vg υg
: Specific area, [L2/L3] : n-hexane concentration in the bulk, [M/L3] : n-hexane concentration in the gas film, [M/L3] : Extra-cellular COV concentration, [M/L3] : Nitrogen source concentration in the hyphae, [M/L3] : Nitrogen source concentration in liquid, [M/L3] : Average diameter of the hyphae, [L] : Axial dispersión coefficient, [L2/T] : n-hexane diffusivity, [L2/T] : Nitrogen source diffusivity in the hyphae, [L2/T] : Biofilter height, [L] : Mass transfer coefficient of gas, [1/T] : Affinity constant of n-hexane, [M/L3] : Affinity constant of nitrogen, [M/L3] : Inhibition constant, [M/L3] : Equilibrium constant of n-hexane/hyphae : Equilibrium constant of nitrogen source/hyphae : Average length of the hyphae, [L] : Average maximum distal length of the individual hyphae, [L] : Average maximum distal length of the branching, [L] : Individual total length of hyphae, [L] : Total length of the hyphae, [L] : Cellular maintenance coefficient, [1/T] : Branching number in individual hyphae : Initial number spores : Reactor total volume, [L3] : Total volume of the support, [L3] : Average lineal rate of the gaseous phase, [L/T] : Gas film rate, [L/T]
MODELING OF A FUNGAL BIOFILTER FOR THE ABATEMENT OF HYDROPHOBIC VOCs
105
Xh,Total : Total biomass, [M] : n-hexane cellular yield, [M/M] YA : Nitrogen source cellular yield, [M/M] YN Symbols : Hyphae density, [M/L3] [1.1´10-9 mg.μm-3] ρh μ : Growth specific rate, [1/T] μmax : Maximum growth specific rate, [1/T] ϕ : Branching frequency, [1/T] : Thickness of the gas film, [L] δg ε R : Bed void fraction γ : Branching proportionality constant β : Principal hyphae fraction
REFERENCES Alonso, C., Suidan, M.T., Sorial, G.A., Smith, F.L., Biswas, P., Smith, P.J. and Brenner, R.C. (1998) Gas treatment in trickle-bed biofilters: biomass, how much is enough? Biotechnol. Bioeng. 54: 583-594. Arriaga, S. and Revah, S. (2005a) Removal of n-hexane by Fusarium solani with a gas-phase biofilter. J Ind. Microbiol. Biotechnol. 32: 548-553. Arriaga, S. and Revah, S. (2005b) Improving n-hexane removal by enhancing fungal development in a microbial consortium biofilter. Biotechnol. Bioeng. 90(1): 107-115. Auria, R., Ortiz, I., Villegas, E. and Revah, S. (1995) Influence of growth and high mould concentration on the pressure drop in solid state fermentations. Proc. Biochem. 30(8): 751-756. Bibeau, L., Kiared, K., Brzenzinski, R., Viel, G. and Heitz, M. (2000) Treatment of air polluted with xylenes using a biofilter reactor. Water, Air, and Soil Poll. 118: 377-393. Deshusses, M.A., Hamer, G. and Dunn, I.J. (1995) Behavior of biofilters for waste air biotreatment. 1. Dynamic model development. Environ. Sci. Technol. 29: 1048-1058. Fuller, E.N., Schettler, P.D. and Giddings, J.C. (1996) Ind Eng Chem 58: 19. García-Peña, E.I., Hernández, S., Favela-Torres, E., Auria, R. and Revah, S. (2001) Toluene biofiltration by the fungus Scedosporium apiospermun TB1. Biotechnol. Bioeng. 76(1): 61-69. Hartmans, S. and Tramper, J. (1991) Dichloromethane removal from waste gases with a trickle-bed bioreactor. Bioproc. Eng. 6: 83-92. Hodge, D.S. and Devinny, J.S. (1995) Modeling removal of air contaminants by biofiltration. J Environ. Eng. 121(1): 21-32.
106
ALBERTO VERGARA-FERNÁNDEZ
AND
SERGIO REVAH
Iliuta, I. and Laranchi, F. (2004) Transient biofilter aerodynamics and clogging for VOC degradation. Chem. Eng. Sci. 59: 3293-3302. Jorio, H., Payre, G. and Heitz, M. (2003) Mathematical modeling of gas-phase biofilter performance. J. Chem. Technol. Biotechnol. 78: 834-846. López-Isunza, F., Larralde-Corona, C.P. and Viniegra-González, G. (1997) Mass transfer and growth kinetics in filamentous fungi. Chem. Eng. Sci. 52(15): 2629-2639. Lobo, R.O. (2004). Principios de transferencia de masa. Universidad Autónoma Metropolitana Unidad Iztapalapa. Segunda impresión, México, pp. 532. Mohseni, M. and Allen, D.G. (2000) Biofiltration of mixtures of hydrophilic and hydrophobic volatile organic compounds. Chem. Eng. Sci. 55: 1545-1558. Ottengraf, S.P.P. and van den Oever, A.H.C. (1983) Kinetics of organic compound removal from waste gases with a biological filter. Biotechnol. Bioeng. 25: 3089-3102. Shareefdeen, Z. and Baltzis, B.C. (1994) Biofiltration of toluene vapor under steady-state and transient conditions: theory and experimental results. Chem. Eng. Sci. 49: 4347-4360. Shuler, M.L., Kargi, F. and Lidén, G. (2003) Bioprocess Engineering, Basic concepts. Second Edition. Prentice Hall of India. New Delhi. Kluwer Academic/Plenum Publishers. New York, USA, Slattery, J.C. (1999) Advanced Transport Phenomena. Cambrige University Press. New York, USA.. pp. 709. Spigno, G., Pagella, C., Daria Fumi, M., Molteni, R., de Faveri, D.M. (2003). VOCs removal from waste gases: gas-phase bioreactor for the abatement of n-hexane by Aspergillus niger. Chem. Eng. Sci. 58: 739-746. Vergara-Fernández A., Van Haaren, B. and Revah, S. (2006) Phase partition of gaseous hexane and surface hydrophobicity of Fusarium solani when grown in liquid and solid media with hexanol and hexane. Biotechnol. Lett. 28: 2011-2017.
107
Gasoline biofiltration: an analytic model ANDREW MARK GERRARD1, MARTIN HALECKY2 AND JAN PACA2 1 2
University of Teesside, Middlesbrough, TS1 3BA, England Institute of Chemical Technology, 166 28 Prague, Czech Republic
ABSTRACT We present an analytical method of solution for a pair of models representing the removal of gasoline from air by biofiltration. The experimental data showed that the aromatic components of the gasoline were more readily digested than the aliphatics. The models, involving two or three fitted parameters, fitted the laboratory data well.
1 INTRODUCTION Biofiltration is an increasingly used process to remove volatile organics (VOC) from a vapour stream. This involves the packed bed absorption of the VOC, followed by its biodigestion. The process can be competitive with condensation, incineration, adsorption and scrubbing when the VOC concentrations are low and the gas flow rates are high. If biofiltration is used to remove gasoline vapour from air, it is found that the aromatic fraction is relatively easily and rapidly digested but the aliphatic fraction is consumed only when a substantial amount of the aromatics have been eliminated. Experimental work done at the ICT, Prague (Halecky et al., 2006) has measured these effects by sampling down the length of a bioreactor bed and obtaining the concentration profiles for both fractions of the gasoline.
107
108
ANDREW M. GERRARD, MARTIN HALECKY AND JAN PACA
2 MODELLING THE SYSTEM This system can be modelled by a pair of component mass balances. Over a differential element in the bioreactor, we can write for the aromatics that: (1) where rar is the apparent rate of removal of the aromatics per unit of biomass, C is a concentration at depth h and Q is a volumetric flow-rate. (The remainder of the terms are defined in the nomenclature section.) This can be rewritten as: (2) and similarly for the aliphatics: (3) The rates of removal were previously modelled by the following expressions:
(4)
and
(5)
Clearly, the first part of these expressions has the form of Monod kinetics, however note that, as a simplification, they are based on gas phase concentrations. The interference between the fractions was handled by allowing each fraction to inhibit the others’ rate of digestion in the form shown by the second part of these equations. This leads to a maximum of six kinetic parameters being needed to characterise the system. However, when the data was analysed, only three parameters were needed: two pseudo first order constants and a further constant to represent the effect of the aromatics on the aliphatic removal (with the multiplicative form: [K4/(K4+Caromatics)].
GASOLINE BIOFILTRATION: AN ANALYTIC MODEL
109
The equations have been solved (Gerrard et al., 2006) using Euler’s numerical technique. The purpose of this paper is to derive some analytic solutions to the fitting of this smaller, (three constant) kinetic model. 2.1 FIRST ANALYTIC SOLUTION The two component mass balances can be written in full as:
(6)
(7)
As mentioned above, the earlier work showed that equation (6) could be simplified to first order kinetics with no inhibition term, thus: (8)
where (9)
Clearly, equation (8) can be immediately solved to give:
(10) For the aliphatics, we can write the equation (7) with first order kinetics again plus the inhibition term to represent the effect of the aromatics on the rate of removal of the aliphatic fraction:
(11)
110
ANDREW M. GERRARD, MARTIN HALECKY AND JAN PACA
(12) Separating the variables:
(13)
Thus,
(14)
where: (15)
(16)
Thus,
(17)
(18) then (19)
GASOLINE BIOFILTRATION: AN ANALYTIC MODEL
111
Hence integral becomes: (20)
Of course, this can be directly integrated to give:
(21)
Finally:
(22)
This model will be fitted to the experimental data later. 2.2 SECOND ANALYTIC SOLUTION A new formulation of the model to represent the effect of the aromatics on the aliphatic rate of removal uses a different multiplying factor namely: [1-k*Caromatic]. Clearly, as the aromatic concentration falls, there is a progressively smaller effect on the rate of aliphatic removal. (This will allow an analytical solution to be found which can predict the concentration profiles almost as well as the solution given above, but with a simpler equation.) In more detail, we now re-write equation (11) as (23)
Introducing equation (10) here we have: (24)
112
ANDREW M. GERRARD, MARTIN HALECKY AND JAN PACA
Separating the variables gives: (25)
Integrating gives:
(26)
or: (27)
This equation can also be used to predict the shape of the concentration profiles.
3 FITTING DATA TO THE MODELS The experimental data from the Institute for Chemical Technology, Prague came from a small biofilter of height equalling 75 cm, with a diameter of 10 cm. The concentrations of aliphatic and aromatic gasoline fractions were measured at the inlet, outlet and four other intermediate points. The trials were organised to give one set of results with approximately constant inlet concentration and a varying gas flow-rate together with another set having constant organic load (eg a high flow-rate and a correspondingly low inlet concentration). The table lists the seven experiments used in the analysis. The third row gives the sum of squares of the errors, SOS, (between the actual and predicted concentration profiles) assuming the simple first order models for both gasoline fractions, ignoring any interaction. The fourth row shows the considerable improvement when equation (22) was used, (which, of course, includes both the first order characteristic plus the inhibition effect of the aromatics on the aliphatic components). The values of the parameters are given in rows five to seven. These are the same as previously reported (Gerrard et al., 2006) which required the use of a numerical solution of the equations. As expected, the aromatic (first order) constants are typically around three times the values of the aliphatic. The K4 values for the inhibition constant are seen to average 16 mg m-3.
113
GASOLINE BIOFILTRATION: AN ANALYTIC MODEL
Table 1. Results of curve fitting. Run 1
2
3
4
5
6
7
Q m /hr Car and al mg/m3 SOS no inhibition
0.015 120, 129
0.03 122, 127
0.06 136, 113
0.006 267, 231
0.015 106, 90
0.03 54, 44
0.06 25, 20
323.
203.
78.
1557.
337.
81.
27.
SOSeqn 22 m1/K1 m2/K3
87. 21.1 4.0
54. 18.4 5.4
75. 9.2 5.4
443. 7.4 1.2
80. 16.9 5.3
23. 21.8 10.2
10. 33.4 11.0
K4
29.9
24.1
18.1
8.10
18.9
6.64
9.5
SOSeqn 27 m1/K1 m2/K3 k
112. 20.8 3.5 0.0083
66. 18.3 3.8 0.0082
75. 9.2 1.8 0.0054
1060. 7.3 0.59 0.0037
103. 16.6 4.1 0.0094
33. 21.6 5.9 0.018
8. 32.1 10.5 0.04
3
The last four rows give the values of the fitted parameters which minimised the SOS using equation (27). (The Solver routine in Excel was used here.) As before, the m1/K1 parameter is again larger than the m2/K3 constant. This again reflects the relative ease of removal of the aromatic components. (The m1/K1 values for both models are very close, the m2/K3 values do differ somewhat.) The values of k used as the parameter to quantify the inhibition effect in the second model were small and positive, as expected. If we recollect the form of the equation used: [1-k*Caromatic], then clearly, the minimum value of this term is zero. This would occur at the highest aromatic concentrations, ie at the inlet, thus k <= 1/ Car, in. Most of the k values were indeed at this upper limit. Hence, the multiplicative inhibition factor can be written as [1-Car/Car,in]. Thus, the rate of digestion of the aliphatics is linearly related to the aromatic concentration and the model reduces, in this instance, to a two parameter system.
4 CONCLUSIONS Two simple, analytic solutions to the modelling of gasoline biofiltration have been produced. The kinetic parameters are listed to aid the designer of such bio-reactors.
114
ANDREW M. GERRARD, MARTIN HALECKY AND JAN PACA
5 ACKNOWLEDGEMENTS The helpful observation of Mr M Lazonby is appreciated here. The study was financially supported by the Czech Science Foundation, Joint Project 104/05/0194and the Ministry of Education of the Czech Republic, Research Project MSM 6046137305.
6 NOMENCLATURE A a B C h k K m Q r u ε μ
area of cross section of the bed, m2 microbial mass per volume of solids, mgbio m-3 constants defined in equations (15 and 16), m-1 concentration in gas phase, mg m-3 bed height, m inhibition constant, m3 mg-1 kinetic constant, mg m-3 constants defined in equations (9 and 12), mg m-3 h-1 volumetric flow-rate, m3 h-1 rate of bio reaction, mg mgbio-1 h-1 defined in equation (18) bed voidage kinetic constant, mg mgbio-1 h-1
Subscripts al aliphatic compounds ar aromatic compounds bio microbial mass in inlet
REFERENCES Gerrard, A.M., Halecky, M. and Paca, J. (2006) Simple model for the biofiltration of the aliphatic and aromatic fractions of gasoline. Paper given to Chisa 2006 Conference, Prague, 30 August 2006. Halecky, M., Paca, J. and Gerrard, A.M. (2006) Gasoline vapor biodegradation along bed height: Part 1: Performance characteristics. Paper given to ACHEMA Congress, 19 May 2006, Frankfurt-am-Main, Germany.
115
Modeling of biomass accumulation and filter bed structure change in biofilters for gaseous toluene removal JINYING XI, HONG-YING HU AND CAN WANG Department of Environmental Science and Engineering, Tsinghua University, Beijing 100084, China
ABSTRACT In this study, a proper microbial growth model was established and analyzed to investigate the biomass accumulation process in biofilters. Four biofilters treating gaseous toluene were set up in parallel and were operated under different inlet toluene loadings for 100 days. Based on the experimental data of microbial biomass and toluene removal rate, the kinetic parameters were decided by either estimation from literature or parameter regression. The calculation results based on the model showed a good agreement with the experimental data of biomass change. By applying the model, it is found that lower than 50% of biomass in the filter bed was active during the last 50 days for the four biofilters. In addition, the void fraction of the filter bed with highest loading was only 55% of the initial level at the end of the operation. All the experimental and calculation results indicated that the microbial growth model could successfully describe the biomass accumulation process and have the potential to predict the long-term performance of biofilters.
1 INTRODUCTION Over the past two decades, biofiltration was recognized as a cost-effective method to treat volatile organic compounds (VOCs) (Leson and Winer, 1991). Many full-scale biofilters were set up and successfully used to treat different organic waste gases emitted from industrial process or waste treatment process (Swanson and Loehr, 1997). However, the VOCs removal capacity of the biofilter was often found to get lower while filter bed pressure drop get higher during long-term operation due to different reasons (Son et al., 2005; Weber and Hartmans, 1996). Bed clogging caused by excess
115
116
JINYING XI, HONG-YING HU
AND
CAN WANG
biomass accumulation was one of the most important reasons for long-term performance decline. To predict long-term performance change and optimize the operation conditions, it is important to know the biomass accumulation pattern and its impact on filter bed structure. Several microbial growth models were established in the literature to describe the biomass accumulation process in the filter beds of biofilters (Iliuta and Larachi, 2004; Song and Kinney, 2002). Some model considered the inert biomass production and its microbial growth pattern was quite different from those models with no inert biomass description. Furthermore, most of the studies did not monitor the biomass accumulation process and verified the model by experimental data. To decide the impact of biomass accumulation to the filter bed and biofilter performance, Alonso et al. (1998) compared three physical filter bed models representing the geometric structure of porous media and found that spheres and pipes model could explain the performance decline of the biofilter. The aim of this study is to describe the microbial growth pattern by developing a proper microbial growth model considering inert biomass production and verify the model based on the experimental data. The filter bed structure change due to biomass accumulation was also analyzed by combining the microbial growth model and a physical filter bed model.
2 MATERIALS AND METHODS 2.1 EXPERIMENTAL SETUP Four paralleled biofilters, identified as 1#, 2#, 3# and 4# respectively were build in this experiment. The diagram of the biofilter system was shown in Figure 1. The packed filter bed had a diameter of 120 mm and a height of 200 mm. Wood chips were used as the organic packing medium and propylene spheres were added into the filter bed (Xi et al., 2005). The void fraction of the filter bed was 0.60 at the beginning of the operation period. A gas feeding apparatus was set up to produce a toluene gas with desired concentration and flow rate (Xi et al., 2005). 2.2 OPERATION CONDITIONS After being inoculated with activated sludge, the four biofilters were operated continuously for more than 100 days. The operating conditions during this period are shown in Table 1.
MODELING OF BIOMASS ACCUMULATION AND FILTER BED STRUCTURE CHANGE
117
Figure 1. Diagram of the experimental system.
Table 1. Operating conditions of the four biofilters. Items or parameters Temperature Relative humidity Inlet toluene concentration
Flow direction Flow rate Empty bed retention time Superficial velocity Spraying intervals Quantity sprayed
Value 20~25 ºC 50~85 0~50d: 950 mg·m-3; 50~100d: 1#:1000mg·m-3, 2#:600mg·m-3, 3#: 270 mg·m-3, 4#:90 mg·m-3 Upflow 0.3 m3·h-1 27 s 26.5 m·h-1 Once every 8 hours, 1min for each time 9~10 L· m-2·min-1
The experimental period was divided into two phase. In Phase I (day 1~50), the inlet concentrations of the four biofilters were same and remained around 950 mg·m-3. In Phase II (day 51~100), the inlet toluene concentrations of the four biofilters were set to different levels as shown in Table 1.
118
JINYING XI, HONG-YING HU
AND
CAN WANG
2.3 ANALYTICAL METHODS During the experimental period, the inlet and outlet toluene concentrations of the four biofilters were monitored by gas chromatograph (Shimadazu GC-14B, Japan) periodically to evaluate the performance of the biofilters. Weighing method was carried out to measure the amount of biomass in the filter bed (Okkerse et al., 1999). The total dry biomass concentration in the filter bed could be calculated by following equation (1).
(1)
where Xt is the biomass concentration in the filter bed, g/m3; Wf is the weight of the wet biofilm on the surface of the packing media, g; Wt is the total weight of the biofilter column, g; Wr is the weight of the empty column, g; Wp is the weight of the wet packing media, g; h is the moisture content of the biofilm, dimensionless; V is the volume of the filter bed, m3; Assuming that the values of Wr ,Wp and h were constant during the experimental period, the value of X can be calculated by measuring Wt.
3 MODEL DEVELOPMENT 3.1 MICROBIAL GROWTH MODEL Assuming that the dry biomass in the filter bed is consisted of active biomass and inert biomass. Thus the dry biomass concentration Xt could be expressed by the following equation: (2) where Xa is the concentration of active biomass, g·m-3; Xi is the concentration of the inert biomass, g·m-3. The variation of Xa and Xi could be expressed as: (3)
(4)
MODELING OF BIOMASS ACCUMULATION AND FILTER BED STRUCTURE CHANGE
119
where Y is the yield coefficient of active biomass, dimensionless; -r is VOCs removal rate, g·m-3·h-1; b is shear/decay coefficient, h-1;β is the ratio of produced inert biomass to lost active biomass, dimensionless; t is time, h. Assuming that Y, b, β is constant for the microbial community in the filter bed and –r is constant with constant operation conditions under pseudo-steady state, the following equations could be drawn:
(5)
(6)
(7)
Where Xa0 is the initial active biomass concentration when t=0, g·m-3; Xi0 is the initial inert biomass concentration when t=0, g·m-3. 3.2 FILTER BED STRUCTURE MODEL The physical filter bed model is similar with the pipes model in the literature (Alonso and Suidan, 1998). Assuming that the biofilm density is constant during microbial growth, the total biomass concentration Xt could be expressed as: (8)
(9)
Where ε0 is the initial void fraction of the filter bed with no biomass, dimensionless; ε is the void fraction of the filter bed with biomass, dimensionless; a0 is the initial specific surface area of the filter bed with no biomass, m2·m-3; a is the specific surface area of the filter bed with biomass, m2·m-3; Xf is the biomass density, g·m-3.
120
JINYING XI, HONG-YING HU
AND
CAN WANG
4 RESULTS AND DISCUSSION 4.1 Model parameter estimation To describe the microbial growth process, the value of Xa0, Xi0, Y, -r, b and β should be estimated by either experiment result or referring to literature. The values of these model parameters were listed in Table 2. Table 2. Values of model parameters Model parameter
value
unit
Estimation method
Xa0 (initial active biomass concentration) Xi0 (initial inert biomass concentration) Y(yield coefficient)
0
g·m-3
Assumption
0
g·m-3
Assumption
0.6
dimensionless
-r(VOCs removal rate)
*
g·m-3·h-1
β (ratio of produced inert biomass to lost active biomass) b (shear/decay coefficient)
0.2
dimensionless
From literature (Rittmann, B.E. and McCarty, P.L. 2001) Averaged by experimental data regression from the experimental data (R2=0.96)
0.005
h-1
regression from the experimental data (R2=0.96)
* The average toluene removal rates were different in different phase for the four biofilters.
4.2 MODEL SIMULATION AND VALIDATION With the parameters, the biomass concentrations of the four biofilters during the operation period were calculated and compared with the experimental data. The results was shown in Figure 2. The results demonstrated that the microbial growth pattern of the biofilter could be well simulated by the microbial growth model established in this study. By the microbial growth model (equation 5, 6), the active and inert biomass concentrations in the filter bed were also calculated. The results were show in Figure 3. The results demonstrated that lower than 50% of biomass in the filter bed was active during the last 50 days of operation. Okkerse et al. (1999) and Song et al. (2002) also established microbial growth model for biofilter or biotrickling filter considering inert biomass production, their work did not show much experimental data. The result in this study verified that the inert biomass could not be omitted and most of the accumulated biomass was inert biomass after long time operation.
MODELING OF BIOMASS ACCUMULATION AND FILTER BED STRUCTURE CHANGE
Figure 2. Variations of the predicted and experimental biomass concentration.
Figure 3. Variation of the active and inert biomass concentration of the four biofilters.
121
122
JINYING XI, HONG-YING HU
AND
CAN WANG
4.3 FILTER BED STRUCTURE CHANGE Based on the filter bed structure model (Equations 8, 9), the characteristic parameters (ε and a) of the filter bed could be calculated. The value of Xf and ε0 was measured and the results were 1×105 g·m-3 and 0.6 respectively. The calculated void fraction and specific surface area during the operation period were shown in Figure 4.
Figure 4. Variation of a/a0 and ε/ε0 during the operation period.
The void fraction and the specific surface area of the filter bed can affect the VOCs removal capacity and bed pressure drop of the biofilter (Morgan-Sagastume et al., 2001; Song and Kinney, 2000). The results in Figure 4 illustrated that the void fraction and the specific surface area of the filter bed would be lowered after the excess biomass accumulated. For biofilter 1#, the void fraction was only 55% of the initial level at the end of the operation.
MODELING OF BIOMASS ACCUMULATION AND FILTER BED STRUCTURE CHANGE
123
5 CONCLUSIONS The following conclusions can be drawn from the results presented in this study: (1) The microbial growth model could successfully describe the biomass accumulation process under different operation conditions. (2) After long time operation, the accumulated biomass in the filter bed was mainly inert biomass and the ratio of inert biomass would raise continuously. (3) The filter bed structure model could quantitatively determine the void fraction and specific surface area of the filter bed decrease due to excess biomass accumulation.
REFERENCES Alonso, C. and Suidan, M. (1998) Dynamic mathematical model for the biodegradation of VOCs in a biofilter: biomass accumulation study. Environ. Sci. Technol. 32: 3118-3123. Iliuta, I. and Larachi, F. (2004) Transient biofilter aerodynamics and clogging for VOC degradation. Chem. Engin. Sci. 59: 3293-3302. Leson, G. and Winer, A.M. (1991) Biofiltration: an innovative air pollution control technology for VOC emissions. J. Air Waste Manage. Assoc. 41: 1045-1054. Morgan-Sagastume, F., Sleep, B.E. and Allen, D.G. (2001) Effects of biomass growth on gas pressure drop in biofilters. J. Environ. Eng. 127: 388-396. Okkerse, W.J.H., Ottengraf, S.P.P., Osinga-Kuipers, B. and Okkerse, M. (1999) Biomass accumulation and clogging in biotrickling filters for waste gas treatment. Evaluation of a dynamic model using dichloromethane as a model pollutant. Biotechnol. Bioeng. 63: 418-430. Rittmann, B.E. and McCarty, P.L. (2001) Environmental Biotechnology: Principles and Applications, vol. 129, McGraw-Hill, New York, USA. Son, H.K., Striebig, B.A. and Regan, R.W. (2005) Nutrient limitations during the biofiltration of methyl isoamyl ketone. Environ. Prog. 24 (1): 75-81. Song, J.Y. and Kinney, K. (2000) Effect of vapor-phase bioreactor operation on biomass accumulation, distribution, and activity: linking biofilm properties to bioreactor performance. Biotechnol. Bioeng. 68 (5): 508-516. Song, J.Y. and Kinney, K. (2002) A model to predict long-term performance of vapor-phase bioreactors: A cellular automaton approach. Environ. Sci. Technol. 36: 2498-2507. Swanson, W.J. and Loehr, R.C. (1997) Biofiltration: fundamentals, design and operations principles, and applications. J. Environ. Eng. 123 (6): 538-546.
124
JINYING XI, HONG-YING HU
AND
CAN WANG
Weber, F.J. and Hartmans, S. (1996) Prevention of clogging in a biological trickle-bed reactor removing toluene from contaminated air. Biotechnol. Bioeng. 50(1): 91-97. Xi, J.Y., Hu, H.Y., Zhu, H.B. and Qian, Y. (2005) Effects of adding inert spheres into the filter bed on the performance of biofilters for gaseous toluene removal. Biochem. Eng. J. 23: 123-130.
125
Modelling the adsorption of styrene and acetone on activated carbon and perlite beds ANDREW MARK GERRARD1, SEBASTIAN MOLLENCAMP1, KEHINDE MAKINDE1, JAN PACA2 AND ONDREJ MISIACZEK2 1 2
University of Teesside, Middlesbrough TS1 3BA, England Institute of Chemical Technology, 16628 Prague, Czech Republic
ABSTRACT Experimental data from the Institute of Chemical Technology, Prague were analysed to derive some simple models of the rate of adsorption of either styrene or acetone from dilute air mixtures on sterile beds of activated carbon or perlite. It was found that the rate of progress towards the saturation of these beds could be reasonably accurately represented by a first order Lagergren model, where the rate constant had a value in the range 0.0039 to 0.29 min-1.
1 INTRODUCTION Biofiltration is a commonly used process to remove volatile organic compounds (VOC) from air streams (Gerrard et al., 2005). It is particularly appropriate for applications with low inlet concentrations and high gas flows. The VOC is absorbed into a moist layer containing the micro-organisms and is biodegraded by them. In unsteady state conditions, adsorption onto the inert solid bed is also significant. In order to better understand this, the poster will describe the simple modelling of a number of experiments which measured the rate of adsorption of styrene and acetone (as both dry and moist gases) onto (biologically inactive) packed beds of activated carbon and of perlite.
125
126
A. M. GERRARD, S. MOLLENCAMP, K. MAKINDE, J. PACA AND O. MISIACZEK
2 EXPERIMENTAL The beds were fed with a steady, incoming flow-rate of gases at a constant inlet concentration of solute. The inlet and outlet concentrations were measured. As time progressed, the bed became progressively saturated with the VOC, thus, the exit concentration steadily rose in a sigmoidal fashion. In total, fourteen experiments were carried out, (by varying the solute and the bed material, the inlet concentration level and the condition of the mixture at the inlet, (i.e. wet or dry).
3 THEORY Where there was enough consistent data, the Freundlich isotherm: (1) was used to represent the equilibrium data, where qe= the (equilibrium) loading of solute on the bed divided by the bed mass, Cin=gas inlet concentration, and m and N are constants). For the kinetic investigation, two models were used, the Lagergren and the Vermuelen. Lagergren proposed to relate the total amount of solute adsorbed by the packing with time. He suggested the rate of change of adsorbed solids in the bed was proportional to the unsaturated state of the bed, raised to a power, thus: (2)
Here, qe is the maximum (equilibrium) amount adsorbed per mass of bed, t is the time. k is the adsorption rate constant (k1 for pseudo first order and k2 for the pseudo second order approach) and n is the model’s order. This equation is easily solved for n = 1 and n = 2. For the pseudo first order model (n = 1) the equation gives: (3) For the pseudo second order model (n = 2) the equation gives: (4)
MODELLING THE ADSORPTION OF STYRENE AND ACETONE ON ACTIVATED CARBON
127
Clearly, both equations are easily linearised. Vermuelen (1953) proposed an equation to describe the fraction of adsorbed solute compared to equilibrium as a function of time, (F(t)=q/qe). He formulates his equation thus:
(5)
where D= effective particle diffusivity and = is the particle size. We can define.
Then substitution and rearrangement lead to a straight line formula: (6)
This can be plotted so that the slope gives A directly.
4 RESULTS For the equilibrium data, the styrene-perlite experiments were the most consistent and gave the following logged graph, see Figure 1, for the five experimental runs. For the dry system, the Freundlich equation has an index of around 1.4 and for the wet conditions, it was closer to unity.
Figure 1. Freundlich diagram for styrene on perlite.
128
A. M. GERRARD, S. MOLLENCAMP, K. MAKINDE, J. PACA AND O. MISIACZEK
Moving to the modelling of the kinetic data, Figures 2 and 3 show log- linear plots showing the above styrene on perlite data fitted to a pseudo first order Lagergren model:
Figure 2. First order Lagergren model for styrene on dry perlite.
Figure 3. First order Lagergren model on wet perlite.
All the coefficients of determination are well above 90%, indicating a reasonable fit.
MODELLING THE ADSORPTION OF STYRENE AND ACETONE ON ACTIVATED CARBON
129
Figure 4. Second order Lagergren model for styrene on wet perlite.
Figure 5. Vermeulen model for styrene on dry perlite.
Figures 4 and 5 show another pair of graphs where some of the styrene on perlite data is fitted to the other two models. Again, the goodness of fit values are pleasing. The Table 1 below summarises all 14 experimental runs and lists the kinetic parameters (k1, k2 and A) and the coefficient of determination, R2, for the three models tested. For example, the third, fourth and fifth rows of numerical data, show the effect of changing inlet concentration for the wet styrene on perlite system. As noted above, all three approaches give reasonable fits to the data, with the average R2 value being 93% for the first order Lagergren model, 88% for the Vermeulen and 86% for the second order Lagergren model. In addition, the range of values of the constants is somewhat smaller for the first order model, which is a further reason to (slightly) prefer it to the alternatives.
130
A. M. GERRARD, S. MOLLENCAMP, K. MAKINDE, J. PACA AND O. MISIACZEK
In more detail: we note for the styrene on perlite results that the inlet concentration does not have a large effect on the value of k1, especially for the wet system. The results for wet acetone on perlite also give almost constant values for this parameter. Table 1. Summary of the constants found in the three models. Inlet Concentration
k1
R2
k2
R2
A
R2
Lagergren
First order
Lagergren
2nd order
Vermeulen
model
0.079 0.035
0.992 0.950
83.5 3.24
0.997 0.986
0.064 0.028
0.994 0.905
0.11 0.079 0.12
0.929 0.994 0.947
151. 31.9 3.22
0.986 0.989 0.975
0.089 0.06 0.10
0.900 0.989 0.900
0.049 0.0069
0.964 0.741
0.016 0.000041
0.967 0.572
0.043 0.0052
0.933 0.653
0.0039
0.985
0.00019
0.916
0.0041
0.845
0.085 0.29
0.865 0.979
0.261 6.56
0.452 0.792
0.072 0.23
0.816 0.960
0.044 0.047
0.987 0.963
0.0266 0.015
0.966 0.968
0.039 0.041
0.964 0.932
0.0069
0.741
0.00004
0.5724
0.0052
0.653
0.0041
0.983
0.00021
0.936
0.0042
0.861
Dry styrene on perlite 49 203 Wet styrene on perlite 55 217 780 Dry styrene on carbon 2498 928 Wet styrene on carbon 693 Dry acetone on perlite 645 1677 Wet acetone on perlite 920 2517 Dry acetone on carbon 929 Wet acetone on carbon 681
MODELLING THE ADSORPTION OF STYRENE AND ACETONE ON ACTIVATED CARBON
131
5 CONCLUSIONS It was found that the pseudo first order Lagergren model gave a good fit to most of the data. Typically, the values for k1 ranged from around 0.0039 to 0.29 min-1 and the coefficients of determination were usually over 90%.
6 NOMENCLATURE A C D F k m n N q ro R2 t
constant concentration diffusivity fractional saturation of bed rate constant equilibrium constant order of model power in Freundlich equation solute adsorbed on bed radius of pore coefficient of determination time
subscripts e equilibrium in inlet condition
REFERENCES Gerrard, A.M., Misiaczek, O., Hajkova, D., Halecky, M. and Paca, J. (2005) Steady state models for the biofiltration of styrene/air mixtures. Chemical and Biochemical Engineering Quarterly. 19 (2): 185-190. Vermuelen, T. (1953) Theory for Irreversible and Constant- Pattern Solid Diffusion. Industrial and Engineering Chemistry. 45: 1664-1665.
This page intentionally left blank
133
Co-treatment of benzene and toluene vapours in a biofilter: A factorial design approach ELDON R. RENE AND T. SWAMINATHAN* Department of Chemical Engineering, Indian Institute of Technology Madras, Chennai – 600036, India
ABSTRACT Biofiltration has now become an indispensable treatment technique for the removal of low concentrations of Volatile Organic Compounds (VOCs) from process vent streams. This study involves performance evaluation of a laboratory scale compost based biofilter for the treatment of mixtures of benzene and toluene (BT) vapours. Experiments were conducted as per a statistical design of experiment, the 2k full factorial design, with the initial concentrations of benzene and toluene and the gas flow rate as the independent variables and the elimination capacity (EC) and removal efficiency (RE) as response variables. The maximum EC attained was 31.7 g/m3.h for benzene and 85.9 g/m3.h for toluene, while the total maximum EC at an inlet loading rate (ILR) of 150.2 g/m3.h was 91.2 g/m3.h. It was also observed that while there was mutual inhibition, benzene removal was severely inhibited by the presence of toluene than toluene removal by the presence of benzene. Statistical analysis in the form of analysis of variance (ANOVA) was carried out to determine the main and interaction effects of variables on the RE and EC values. This study establishes the potential application of biofilters to handle mixtures of VOCs effectively through a statistically authentic approach.
1 INTRODUCTION The growth of industries and heavy vehicular traffic has contributed tremendously to the decline in ambient air quality and pollution of the environment. Among these, the Volatile Organic Compounds (VOCs) emitted from process industries pose a significant threat to human health and environment. Benzene and toluene (BT) are two commonly noticed VOCs that arise from the petrochemical industry, pharmaceutical and printing works. BT, even in low concentrations has been found to cause significant damage to the liver and kidney and paralyse the central nervous system (Martin et al., 1998; Murata et al., 1999). In practice, emissions from these process industries are often
133
134
ELDON R. RENE AND T. SWAMINATHAN
characterized by the presence of mixtures of VOCs rather than as just one pollutant. Biological techniques are attracting growing interest for control of VOC emissions. Biofiltration, a process that removes VOCs from air stream by passing it through a packed bed of biofilm grown on an inert support, has been scaled up to industrial scale. Biological treatment begins with the treatment of contaminants from the air phase to the water phase (Devinny et al., 1999). The efficiency of this process primarily depends on the kinetics of micro processes such as absorption, adsorption, diffusion and biodegradation. However, while treating mixtures interaction effects between pollutants can play an important role in both mass transfer and biodegradation steps of the biofiltration process (McNevin and Barford, 2000). This study primarily aims in evaluating the removal pattern of BT mixture at different concentrations and flow rates in a lab scale compost biofilter.
2 THE 2k FULL FACTORIAL DESIGN Experiments with biofilters would normally involve in studying the effects of two or more factors (parameters) on a response variable. Factorial design is widely used in experiments where main and interaction effects of factors are likely to have a significant effect on the final response. For example, in biofiltration of mixtures of toluene and xylene, Jorio et al. (1998) reported that the degradation of toluene was inhibited by the presence of xylene in mixture. Similarly in a binary mixture of ethanol and methanol, Arulneyam and Swaminathan (2004) reported that the presence of methanol inhibited the degradation of ethanol and vice versa. The 2 level factorial design central composite design was used in this study with three factors. The factors that were chosen were benzene concentration, toluene concentration and flow rate, while the response variables were removal efficiency (RE, %) and elimination capacity (EC, g/m3.h).
3 MATERIALS AND METHODS 3.1 CHEMICALS Laboratory grade chemicals of benzene (>99%) and toluene (>99%, sulphur free) were purchased from Ranbaxy Fine Chemicals Limited, India. 3.2 MICROORGANISM AND MEDIA The mixed microbial consortium obtained from a sewage treatment plant was acclimatized with benzene as the carbon source in 250 ml flasks. The nutrient solution had the following composition (in g/l); K2HPO4 – 0.8, KH2PO4 – 0.2, CaSO4⋅2H2O – 0.05, MgSO4⋅7K2O – 0.5, (NH4)2SO4- – 1.0 and FeSO4 – 0.01 in distilled water.
CO-TREATMENT OF BENZENE AND TOLUENE IN VAPOURS A BIOFILTER
135
3.3 FILTER MATERIAL The packing material consisted of sieved compost (3-6 mm) and ceramic beads (4-6 mm) mixed in a 6:4 volume ratio. The filter material was inoculated with the acclimatized mixed consortia and loosely packed into the biofilter. 3.4 EXPERIMENTAL SETUP Figure 1 illustrates the schematic of the experimental setup. The biofilter was made of poly acrylic tubes (5×70 cm) having 6 sampling ports sealed with a rubber septa at 10 cm along the biofilter height. The filter material supported on a perforated plate was packed to a height of 50 cms. Mixtures of BT were generated by passing humidified air through one of the VOC reservoirs and mixing with other vapour stream in a mixing chamber. The ranges of factors selected for factorial design are shown in Table 1. Samples were collected at regular time intervals using a gas tight syringe and analyzed for residual benzene and toluene concentrations.
Figure 1. Schematic of the experimental setup.
136
ELDON R. RENE
AND
T. SWAMINATHAN
Table 1. Range of factors selected for factorial design. Factors 3
Benzene, g/m Toluene, g/m3 Flow rate, m3/h
-1 (low) 0.12 0.14 0.024
Range and levels of factors 0 (center point) 0.54 0.81 0.036
+1 (high) 0.95 1.48 0.072
3.5 ANALYTICAL METHODS Benzene and toluene concentrations in the gas were measured using a Gas Chromatograph (Model 5765, Nucon gas chromatograph, Nucon Eng. India) with a poropak column (1/8" ID, liquid – 10% FFAP, solid – Ch-WIHP, 80/100 mesh) and flame ionization detector. Nitrogen was used as the carrier gas at a flow rate of 20 ml/ min. The temperatures of the injection port, oven and detection port were 150, 120 and 250 °C respectively. The elution times were 1.1 min for benzene and 1.7 min for toluene respectively. 3.6 SOFTWARE USED Statistical calculations and analysis (F – Fischer’s variance ratio; P – probability value and T – test of significance) were done using the software MINITAB (version 12.2, PA, USA).
4 RESULTS AND DISCUSSIONS 4.1 BIOFILTER STARTUP Prior to carrying out experiments with mixtures of benzene and toluene (BT), the biofilter was operated for about 5 months under varied operating conditions with benzene vapours (data not shown). This biofilter was fed with near equal proportions of benzene and toluene for 18 days at a flow rate of 0.024 m3/h and concentrations varying between 0.3-0.4 g/m3. A slow gain in the removal of toluene could be noticed from the 6th day of operation and this subsequently reached a steady state value of 74% on the 18th day. This removal pattern indicates good acclimatization and high biological activity of the filter bed (graph not shown). 4.2 REMOVAL OF MIXTURES OF BENZENE AND TOLUENE Continuous experiments were carried out according to the 23 full factorial design (k=3), at different initial concentrations of benzene and toluene and at different flow
137
CO-TREATMENT OF BENZENE AND TOLUENE IN VAPOURS A BIOFILTER
rates after acclimatizing the biomass to a mixture of benzene and toluene. The low and high conditions for the variables are: benzene concentration – 0.12 and 0.95 g/m3; toluene concentration – 0.14 and 1.48 g/m3; flow rate – 0.024 and 0.072 m3/h. Each experimental run was run for a period of 5 to 8 days to achieve steady state removal profiles and the average of these values were taken for calculating the RE and EC of the biofilter. The entire design of experiments along with their respective RE and EC is shown in Table 2. Table 2. Range of operating variables, removal efficiency and elimination capacity of biofilter for benzene and toluene removal. Expt. No 1 2 3 4 5 6 7 8 9 10
Initial concentration, g/m3 Benzene Toluene 0.12 0.14 0.95 0.14 0.12 1.48 0.95 1.48 0.12 0.14 0.95 0.14 0.12 1.48 0.95 1.48 0.54 0.81 0.54 0.81
Flow rate, m3/h 0.024 0.024 0.024 0.024 0.072 0.072 0.072 0.072 0.048 0.048
RE, % Benzene 72.7 54.5 68.6 29.5 63.2 41.5 56.2 13.8 30.9 28.4
Toluene 81.1 80.9 73.8 62.3 75.4 71.1 62.3 36.9 65.9 66.8
EC, g/m3.h Benzene 2.31 13.35 2.21 7.32 6.22 31.66 5.69 10.41 9.85 8.71
Toluene 2.92 2.88 29.44 24.11 8.79 8.56 76.47 42.48 27.91 25.91
The results expressed in terms of RE for benzene and toluene is shown in Figs. 2 and 3 respectively. It was observed that the RE for both benzene and toluene were reduced in comparison to the individual study (Rene, et al., 2005). At low inlet loading rates (ILR, g/m3.h) for both the substrates (total ILR < 7 g/m3.h), benzene removal was inhibited by around 19.2%, and toluene removal by 14.6%. At the same flow rate, when the benzene concentration was low and toluene concentration was high, the inhibition of toluene removal by benzene was less (4%) compared to that of benzene removal by toluene (23.4%). At high concentrations of both the substrate and at a total loading rate of 65 g/m3.hr, benzene degradation was inhibited by 60.1%. At the highest ILR of 200 g/m3.hr, the inhibition of benzene removal by toluene was 80.6%, while toluene removal inhibited by only 37%. Overall it is very obvious that the biofilter was much more effective in eliminating toluene than benzene.
138
ELDON R. RENE AND T. SWAMINATHAN
Figure 2. Removal profile of benzene under the influence of different toluene concentrations in biofilter.
Figure 3. Removal profile of toluene under the influence of different benzene concentrations in biofilter.
CO-TREATMENT OF BENZENE AND TOLUENE IN VAPOURS A BIOFILTER
139
This general feature of the biofilter for the treatment of BT mixtures is shown in Figure 4. The elimination capacities of the biofilter for both the pollutants were similar at lower ILRs. At high ILRs, the EC due to toluene removal was much higher than that of benzene. It may also be observed that at higher ILRs, the total EC of the biofilter decreased significantly due to inhibitory effects of the pollutants. The total maximum EC observed in this study is 91.2 g/m3.h at a total inlet loading rate of 150.2 g/m3.h. This EC value is nearly 25% higher than the EC values observed in the same biofilter when benzene was treated individually. It should be noticed that at this ILR, nearly 94% of the elimination was contributed due to toluene degradation at higher loading rates. Pressure drop and outlet temperatures were monitored during the operating period. Pressure drop values were nearly constant and varied between 3.5 – 7 cms of H2O, while the outlet temperature varied between 27 – 35 oC.
Figure 4. Effect of total inlet load on the elimination capacity of biofilter treating mixtures of benzene and toluene.
It is also well recognized that in the case of biodegradation of two or more pollutants the metabolic activity may involve the mechanism of induction, inhibition or co metabolism depending on the substrate and microbial species present. In biofiltration studies, Veir et al. (1996) studied the interaction of DCM and toluene in a compost biofilter that was earlier used to treat DCM. The presence of toluene caused
140
ELDON R. RENE AND T. SWAMINATHAN
an immediate decline in the DCM removal, but however the RE recovered considerably in the following weeks of operation. On the other hand toluene removal showed no inhibition during the recovery phase of DCM removal. Arulneyam and Swaminathan (2004) in their study on biofiltration of methanol and ethanol vapours observed that the removal efficiencies in mixtures were comparatively less than those for individual compounds. They further conclude that the effect on ethanol removal was due to inhibitory effect of methanol on ethanol utilizing organisms, while the effect on methanol removal was due to preferential utilization of ethanol by methanol utilizing microorganisms. The main effect plot illustrating the effect of co-substrate on the removal of the primary substrate for benzene and toluene are shown in Figures 5 and 6. It was observed that the removal of the primary substrate decreased with increasing concentrations of both the primary and co-substrates. However, the effect of cosubstrate was little less than that of the primary substrate. This indicates mutual inhibition between benzene and toluene on each others removal. Table 3 show the analysis of variance (ANOVA) for RE of benzene and toluene as binary mixtures under the conditions used in this biofiltration study. The main effects were significant for all the profiles as observed from the low P values (< 0.05). Among the main effects on toluene removal, toluene concentration (T = -39.25, P = 0.016) appears to play a major role than benzene concentration (-22.26, 0.029) and flow rate (-28.17, 0.023). The conclusion that could possibly be drawn from this study is that presence of toluene inhibits the degradation of benzene strongly at higher loading rates, while the presence of benzene has little effect on toluene degradation.
Figure 5. The main effects plot of variables for benzene removal in biofilter treating mixtures of benzene and toluene.
CO-TREATMENT OF BENZENE AND TOLUENE IN VAPOURS A BIOFILTER
141
Figure 6. The main effects plot of variables for toluene removal in biofilter treating mixtures of benzene and toluene.
Table 3. ANOVA for benzene and toluene removal from BT mixtures in biofilter. Source Main effects 2-Way Interactions 3-Way Interactions
Benzene removal F P 284.92 0.044 24.11 0.148 0.00 0.975
Toluene removal F P 943.15 0.024 177.18 0.055 28.91 0.117
5 CONCLUSIONS Biofiltration of benzene and toluene vapours was investigated in a laboratory scale biofilter at varying operating conditions. Experiments were carried out according to the runs specified by the 2k full factorial design. It was observed that at total ILRs less than 7 g/m3.h, a maximum removal of 72.7% and 81.1% were achieved for benzene and toluene respectively. Moreover, an increase in the concentration of both benzene and toluene from low to high levels showed a significant decrease in the removal of both the VOCs. Furthermore, the results showed that the EC of the biofilter for both the pollutants were similar at low ILRs. At high ILRs, EC for toluene was much higher than that of benzene. The results from statistical analysis reveal that the main effects of the variables were significant with low P values (<0.05) than the interaction effects for the removal of both the compounds.
142
ELDON R. RENE AND T. SWAMINATHAN
REFERENCES Arulneyam, D. and Swaminathan, T. (2004) Biodegradation of mixture of VOCs in a biofilter. J. Environ. Sci. 16(1): 30-33. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for Air Pollution Control. Lewis Publisher, Boca Raton, Florida. Jorio, H., Kiared, K., Brzezinski, R., Leroux, A., Viel, R., and Heitz, M. (1998) Treatment of air polluted with high concentrations of toluene and xylene in a pilot scale biofilter. J. Chem. Technol. Biotechnol. 73: 183-196. Martin, M.A., Keuning, S. and Janssen, D.B. (1998) Handbook on Biodegradation and Biological Treatment of Hazardous Organic Compounds, 2nd ed., Academic Press, Dordrecht. McNevin, D. and Barford, J. (2000) Biofiltration as an odour abatement strategy. Biochem. Eng. J. 5: 231-242. Murata, M., Tsujikawa, M. and Kawanishi, S. (1999) Oxidative DNA damage by minor metabolites of toluene may lead to carcinogenesis and reproductive dysfunction. Biochem. Biophys. Res. Comm. 261: 478-483. Rene, E.R, Murthy, D.V.S, and Swaminathan, T. (2005) Performance evaluation of a compost biofilter treating toluene vapours, Proc. Biochem. 40: 2771-2779. Veir, J.K., Schroeder, E.D., Chang, D.P.Y. and Scow, K.M. (1996) Interaction between toluene and dichloromethane degrading populations in a compost biofilter. Proceedings of the 89th Annual Meeting and Exhibition of the Air and Waste Manage. Assoc., Nashville, Tennessee.
143
Mathematical modeling and simulation of volatile reduced sulfur compounds oxidation in biotrickling filters G. AROCA, M. CÁCERES, S. PRADO, C. SÁNCHEZ AND R. SAN MARTÍN School of Biochemical Engineering, P. Universidad Católica de Valparaíso, Av. Brasil 2147, Valparaíso, Chile
ABSTRACT The odour generated by industrial gaseous emissions causing nuisances generally is due to the presence of volatile reduced sulfur compounds (VRSC) Although a number of microorganisms are known for degrading VRSC, the treatment of a mixture of reduced sulfur compounds remains challenging for several reasons. To resolve these problems two-stage systems have been proposed, in the first reactor H2S is bio-oxidized and in the second the rest of the VRSC mixture, avoiding the inhibition effects of H2S over the bio-oxidation of these compounds. In the systems described the complete oxidation of H2S must be performed in the first reactor, if some H2S pass though out the first reactor it would have an effect on the bio-oxidation of the other VRSC present in the mixture in the second bioreactor. This situation was modelled and simulated, and is presented in this article. The bio-oxidation of H2S and DMS in a biotrickling filter is described through a model of the mass transfer and chemical reaction processes. The biotrickling filter is modeled as a fixed bed of packing material which supports the growth of micro-organisms as biofilms. When air flows in the bed, H2S and DMS are continuously transferred from the gas phase to the biofilm, where they diffuse and are oxidized by aerobic microbial activity. A summary of the equations, results of the simulation and sensibility to the inhibition constants are reported.
1 INTRODUCTION The odour generated by industrial gaseous emissions is one of the most important environmental problems when the installations are near to urban areas or where the urban areas have grown until surround industrial areas. Depending on the type of
143
144
G. AROCA, M. CÁCERES, S. PRADO, C. SÁNCHEZ AND R. SAN MARTÍN
industry the origin of this problem is the presence of odorous volatile organic compounds (VOC) and particularly volatile reduced sulfur compounds (VRSC) also called TRS (Total Reduced Sulfur), like hydrogen sulfide, methylmercaptane, dimethylsulfur and dimethyldisulfur and other sulfur volatile compounds in the emissions. These compounds can be found in the gaseous emissions of several industrial operations, like Kraft pulp mills, petroleum refineries, tanneries, some food industries; particularly fish canning and animal rendering operations, and also in waste water treatment plants, landfills, composting and solid waste treatment plants. Although a number of microorganisms are known for degrading VRSC, the treatment of a mixture of reduced sulfur compounds remains challenging for several reasons. Firstly, H2S is preferentially degraded over dimethyl sulfide or other organic sulfur compounds (Cho et al., 1992; Wani et al., 1999; Zhang et al., 1991) because H2S oxidation is the energy yielding process (Smet et al., 1998). Secondly, the degradation of MM, DMS and DMDS is carried out with high efficiency at neutral pH, but decreases at low pH (Smet et al., 1996). Thirdly, the degradation rates decrease in the order H2S > MT > DMDS > DMS (Cho et al., 1991; Smet et al., 1998). To resolve these problems, a few two-stage systems have been proposed (Park et al., 1993; Ruokojarvi et al., 2001). The most recent one, developed by Ruokojarvi et al. (2001) consisted out of two biotrickling filters, connected in series, inoculated with a microbial consortium enriched from sludge water from a refinery, with H2S or DMS, respectively. The reactors were operated at different pH levels, to allow efficient removal of organic sulfur compounds at neutral pH in the second reactor. HBr:2S and DMS elimination capacities as high as 47.9 and 36.6 g S m-3 h-1, respectively, were obtained for the whole two-stage biotrickling filter. Sercu et al. (2005) shown that the two-stage biofiltration system is an efficient system for the treatment of waste gases containing a mixture of reduced sulfur compounds, because it optimizes the potential of DMS degradation. Maximum elimination capacities obtained for DMS in this study were 57 g m-3 h-1 (120 s. EBRT, D = 92%) and 58 g m-3 h-1 (60 s. EBRT, D = 89%). In the systems described the complete oxidation of H2S must be performed in the first reactor, if some H2S pass through out the first reactor without been oxidized it would have an effect on the bio-oxidation of the other VRSC present in the mixture in the second bioreactor. This situation was modelled and simulated, and is presented in this article.
2 MATHEMATICAL MODELING Steady-state and dynamic models have been developed to provide a description of degradation mechanisms and biomass accumulation in biofiltration processes (Dehusses et al., 1998; Ottengraf and van den Oever, 1995; Song and Kinney, 2002;
MATHEMATICAL MODELING AND SIMULATION OF VOLATILE REDUCED SULFUR
145
Zarook et al., 1993, 1998). Simplified first- and zero-order kinetic expressions were initially introduced to model the degradation process (Ottengraf and van den Oever, 1995). More recently, Monod-type kinetic models (Monod, 1942), including substrate inhibition (Andrews, 1968), were applied (Zarook and Shaik, 1997). Zarook et al (1993) have also included expressions for the reaction rate which explicitly take into account the potential limiting effects of oxygen. The bio-oxidation of H2S and DMS in a biotrickling filter is described through a model of the mass transfer and chemical reaction processes. The biotrickling filter is modeled as a fixed bed of packing material which supports the growth of microorganisms as biofilms. When air flows in the bed, H2S and DMS are continuously transferred from the gas phase to the biofilm, where they diffuse and are oxidized by aerobic microbial activity. ASSUMPTIONS All mass transfer and reaction processes occurring in the biofilter are considered to be at steady state. This assumption is justified by considering that that the time scale of biofilm growth (δ2/Db) is up to two orders of magnitude less than the residence time, and, as a consequence, time variations in the biofilm can be neglected (Zarook and Shaikh, 1997). In particular, the rate of biomass accumulation in the reactor is small compared to the overall H2S and DMS degradation rate, allowing to neglect biomass mass balances in the model. The biofilm is fully developed and biomass accumulation does not occur. At steady state, the sorption of H2S and DMS on the packing material (adsorption onto the solid plus absorption in the water retained in the pores of solid) is in equilibrium and need not to be taken into consideration in the mass balances. Other assumptions made to derive the model equations are as follows: – Oxygen is present in excess in relation to H2S and DMS. – The biofilm forms on the external surface of the packing material and no reaction occurs in the pores. – The biofilm forms as patches on the support. The extent of the patches is much larger than depth, and H2S and DMS are transported into the biofilm perpendicularly to the biofilm–gas interface. – Gas–solid mass transfer is fast compared to diffusion and reaction in the biofilm. As a consequence, H2S and DMS concentration at the biofilm gas interface are calculated through Henry’s law, assuming the same distribution coefficient as in water. – The effective diffusivities in the biofilm is calculated from the corresponding values in water through a correction factor which depends on biofilm porosity.
146
G. AROCA, M. CÁCERES, S. PRADO, C. SÁNCHEZ AND R. SAN MARTÍN
– The biofilm properties (i.e. thickness, density, specific surface area) are uniform along the bed height, and constant under different operating conditions. – The thickness of the biofilm is small relative to the main curvature of the solid particles and can be assumed as flat. As a consequence, model equations are derived for a planar geometry. – Gas mixing in vapour phase bioreactors is described using a dispersion model (Levenspiel, 1999). – The rate of oxidation of H2S can be described by Monod’s expression. – The rate of oxidation of DMS can be described using modified Monod’s expression with competitive inhibition due to the presence of H2S. EQUATIONS General Mass balance for the compound i in the gaseous phase:
General Mass balance for the compound i in the liquid phase:
Kinetic equations for the rate of bioxidation of H2S
Kinetic equations for the rate of bioxidation of DMS
BOUNDARY CONDITIONS Boundary conditions for gaseous phase:
MATHEMATICAL MODELING AND SIMULATION OF VOLATILE REDUCED SULFUR
147
Where λ = biofilm width [m]. Boundary conditions for the liquid phase:
H = biofilter height [m]
3 SIMULATION The system of equations was solved using the method of divided finite differences using Mathlab 7.0. The values of the parameters included in the model are shown in Table 1. These values were collected from different references or estimated. Table 1. Model parameters. Parameter Kinetic parameters Monod constant H2S (KS) Maximun oxidation rate (Vm) H2S Monod constant DMS (KS) Maximun oxidation rate (Vm) DMS Yield coefficient (Yx/H2S) Yield coefficient (Yx/DMS) Inhibition constant (KI) Mixing, transport and equilibrium characteristics Diffusion coefficient of H2S in air (DH2S) Diffusion coefficient of DMS in air (DDMS) Air–water H2S partition coefficient (m) Air–water DMS partition coefficient (m) Packing material Specific surface area (a,) Porosity (ε) Fraction of surface covered with biofilm (α) Biofilm properties Density (Xb,) Porosity (εb) Thickness (δ)
Value
Units
1 0.04 5 0.025 0.03 0.007 0.01 – 0.0001
g.m–3 h-1 g.m–3 h-1 g biomass g–1 H2S g biomass g–1 DMS g.m–3
1.35 × 10–9 1.93 x 10–9 0.41 0.07
m2 s–1 m2 s–1 dimensionless dimensionless
300 0.3 0.4
m–1 dimensionless dimensionless
30 0.8 30
kg.m–3 dimensionless μm
148
G. AROCA, M. CÁCERES, S. PRADO, C. SÁNCHEZ AND R. SAN MARTÍN
Figure 1 shows the profile of DMS through out the column without the presence on H2S. Figures 2 and 3 shown the results of the simulations considering different ratio of inlet concentration of DMS and H2S. From these graphs it is possible to observe the effect of H2S in the bio-oxidation of DMS.
Figure 1. Dimensionless concentration of DMS ( ) along the column. Simulation for inlet concentrations of DMS 1,5 g/m3, and inlet concentration of H2S : 0 g/m3. EBRT: 60 s.
Figure 2. Dimensionless concentration of DMS ( ) and H2S ( ) along the column. Simulation for inlet concentrations of DMS 1,5 g/m3, and inlet concentration H2S 0.1 g/m3. EBRT: 60 s.
MATHEMATICAL MODELING AND SIMULATION OF VOLATILE REDUCED SULFUR
149
Figure 3. Dimensionless concentration of DMS ( ) and H2S ( ) along the column. Simulation for inlet concentrations of DMS 1,5 g/m3, and inlet concentration H2S 1 g/m3. EBRT 60 s.
Figure 4. Effect of the values of the inhibition constant on the removal capacity of DMS for an inlet concentration of H2S 0.1 g/m3.
150
G. AROCA, M. CÁCERES, S. PRADO, C. SÁNCHEZ AND R. SAN MARTÍN
According to the results obtained from the simulations, the model represents the expected situation of inhibition in the bio-oxidation of DMS due to the presence of H2S. The results are sensible to the value of the inhibition constant as is also expected. Figure 4 shows the effect of the inhibition constant in the removal capacity of DMS when H2S is loaded at an inlet concentration of 0.1 g/m3 at an EBRT of 60 s. The validation of these simulations will be done with experimental results.
REFERENCES Andrews, J.F. (1968) A mathematical model for the continuous culture of microorganisms utilizing inhibitory substrates. Biotech. Bioeng. 10: 707-714. Cho K.S., Hirai M. and Shoda M. (1991) Degradation characteristics of hydrogen sulfide, methanethiol, dimethyl sulfide and dimethyl disulfide by Thiobacillus thioparus DW44 isolated from peat biofilter. J. Ferment. Bioeng. 71(6): 384-389. Cho K.S., Hirai M. and Shoda M. (1992) Enhanced removability of odorous sulfur-containing gases by mixed cultures of purified bacteria from peat biofilters. J. Ferment. Bioeng. 73: 219-224. Deshusses, M.A., Hamer, G. and Dunn, I.J. (1995) Behavior of biofilters forwaste air and biotreatment. 1. Dynamic model development. Environ. Sci. Technol. 29: 1048-1058. Levenspiel, O. (1999) Chemical Reaction Engineering, 3rd Ed., Wiley, New York, 665 pp. Monod, J. (1942) Recherches sur la croissance des cultures bacteriennes. Hermann et Cie, Paris. Ottengraf, S.P.P. and van den Oever, A.H.C. (1983) Kinetics of organic compound removal from waste gases with a biological filter. Biotechnol. Bioeng. 25: 3089-3102. Park, S.-J., Hirai, M. and Shoda, M. (1993) Treatment of exhaust gases from a night soil treatment plant by a combined deodorization system of activated carbon fabric reactor and peat biofilter inoculated with Thiobacillus thioparus DW4. J. Ferment. Bioeng. 76: 423-426. Ruokojärvi, A., Ruuskanen, J., Martikainen, P.J. and Olkkonen, M. (2001) Oxidation of gas mixtures containing dimethyl sulfide, hydrogen sulfide, and methanethiol using a two-stage biotrickling filter. J. Air Waste Manage. Assoc. 51: 11-16. Sercu, B., Nuñez, D., van Langenhove, H., Aroca, G. and Verstraete, W. (2005) Operational and microbiological aspects of a bioaugmented two-stage biotrickling filter removing hydrogen sulphide and dimethylsulfide. Biotechnol. Bioeng. 90: 259-269. Smet, E., Chasaya, G., van Langenhove, H. and Verstraete, W. (1996) The effect of inoculation and the type of carrier material used on the biofiltration of methyl sulphides. Appl. Microbiol. Biotechnol. 45 (1-2): 293-298. Smet, E., Lens, P. and van Langenhove, H. (1998) Treatment of waste gases contaminated with odorous sulphur compounds. Crit. Rev. Environ. Sci. Technol. 28: 89-117.
MATHEMATICAL MODELING AND SIMULATION OF VOLATILE REDUCED SULFUR
151
Song, J. and Kinney, K.A. (2002) A model to predict long term performance of vapor-phase bioreactors: a cellular automation approach. Environ. Sci. Technol. 36: 2498-2507. Wani, A.H., Lau, A.K. and Branion, R.M.R. (1999) Biofiltration control of pulping odors – hydrogen sulfide: performance, macrokinetics and coexistence effects of organo-sulfur species. J. Chem. Technol. Biotechnol. 74: 9-16. Zhang, L., Hirai, M. and Shoda, M. (1991) Removal characteristics of dimethyl sulfide, methanethiol and hydrogen sulfide by Hyphomicrobium sp. I55 isolated from peat biofilter J. Ferment. Bioeng. 72 (5): 392-396. Zarook, S., Baltzis, B.C., Oh, Y.S. and Bartha, R. (1993) Biofiltration of methanol vapor, Biotechnol. Bioeng. 41: 512-524. Zarook, S. and Baltzis B.C. (1994) Biofiltration of toluene vapor under steady-state and transient conditions: theory and experimental results. Chem. Eng. Sci. 49(24A): 4347-4360. Zarook, S., Shaikh, A.A. and Azam, S.M. (1998) Axial dispersion in Biofilters. Biochem. Eng. J. 1: 77-84. Zarook, S. and Shaikh, A.A. (1997) Analysis and comparison of biofilter models. Chem. Eng. J. 65: 55-61.
This page intentionally left blank
153
Artificial neural network based model for evaluating performance of immobilized cell biofilter ELDON R. RENE, JUNG HOON KIM AND HUNG SUCK PARK Department of Civil and Environmental Engineering, University of Ulsan, P.O. Box 18, Ulsan 680-749, South Korea
ABSTRACT Artificial neural networks (ANNs) are powerful data driven modelling tools which has the potential to approximate and interpret complex input/output relationships based on the given sets of data matrix. In this paper, a predictive computerised approach has been proposed to predict the performance of an immobilized cell biofilter treating NH3 vapours in terms of its removal efficiency (RE) and elimination capacity (EC). The input parameters to the ANN model were inlet concentration, loading rate, flow rate and pressure drop, while the output parameters were RE and EC respectively. The data set was divided into two parts, training matrix consisting of 51 data points, while the test matrix had 16 data points representing each parameter considered in this study. Earlier, experiments from continuous operation in the biofilter showed removal efficiencies from 60 to 100% at inlet loading rates varying between 0.5 to 5.5 g NH3/m3.h. The internal network parameters of the ANN model during simulation was selected using the 2k factorial design and the best network topology for the model was thus estimated. The predictions were evaluated based on their determination coefficient values (R2). The results showed that a multilayer network (4-4-2) with a back propagation algorithm was able to predict biofilter performance effectively with R2 values of 0.9825 and 0.9982. The proposed ANN model for biofilter operation could be used as a potential alternative for knowledge based models through proper training and testing of the state variables.
153
154
ELDON R. RENE, JUNG HOON KIM AND HUNG SUCK PARK
1 INTRODUCTION Ammonia is used extensively in the semiconductor industry, as the starting material for the manufacture of nitric acid and as a refrigerating fluid instead of chlorofluorocarbons. Malodors containing NH3 are released from pulp and paper industry, wastewater treatment plants, night soil treatment plants and aerobic composting of low C/N material. Hence there arises a potential need to adapt suitable control techniques for the effective removal of these emissions from related process industries. Biofiltration is a cost effective technology for treatment of waste gases containing low concentrations of VOCs at large flow rates. The high removal efficiencies (REs) achieved along with uncomplicated flexible design, low operational and maintenance costs edges biofilters over other biological treatment techniques such as biotrickling filters and bioscrubbers (Kennes and Veiga, 2001). Biofilters have proved to remove NH3 emissions effectively from gas streams using a bed of biologically active material such as compost, peat, wood bark, etc. In recent years, immobilization of microbes in support matrix such as alginate beads or suitable polymeric materials has gained popularity in the field of biofiltration. The main advantages of adopting immobilization techniques in biofiltration is to provide high cell concentrations, improve genetic stability, protection from shear damage and to enhance favorable microenvironment for microbes (nutrient gradients and pH). Chung et al. (1996) evaluated the effects of operational factors such as retention time, temperature and inlet concentration on the performance of a biofilter packed with Thiobacillus thioparus immobilized with Ca-alginate pellets and found an optimal S-loading of 25 g m–3 h–1. Traditionally the performance of biofilters has been modeled/predicted using process based models that are based on mass balance principles, simple reaction kinetics and a plug flow of air stream (Ottengraf and van Den Oever, 1983; Shareefdeen et al., 1993; Deshusses et al., 1995; Jin et al., 2006). The main advantages of these process models are that they are based on the underlying physical process and the results obtained generally provide a good understanding of the system. However this depends on numerous model parameters and obligates information on specific growth rate of microbes, biofilm thickness and density, values of diffusivity, partition, yield and distribution coefficient, intrinsic adsorption etc. The accurate estimation of some of these parameters requires elaborate technical facilities and expertise, the absence of which hinders the preciseness of the model and limits the application and reliability of the model. An alternate modelling procedure consists of a data driven approach wherein the principles of artificial intelligence is applied with the help of neural networks. It has been shown earlier that the performance of biofilters and biotrickling filters can be predicted from prior estimation of easily measurable operational parameters such
ARTIFICIAL NEURAL NETWORK BASED MODEL FOR EVALUATING PERFORMANCE
155
as flow rate, unit flow, inlet loading rate, pressure drop and inlet concentration (Rene et al., 2006; Elias et al., 2006).
2 THE ANN BASED MODELING APPROACH A multi layer perceptron (MLP) using the back propagation algorithm (Rumelhart et al., 1986) is the most widely used neural network for forecasting/prediction purposes (Maier and Dandy, 2000). Neural networks acquire their name from the simple processing units in the brain called neurons which are interconnected by a network that transmits signals between them. These can be thought of as a black box device that accepts inputs and produces a desired output. MLP generally consists of three layers; an input layer, a hidden layer and an output layer. Each layer consists of neurons which are connected to the neurons in the previous and flowing layers by connection weights (Wij). These weights are adjusted according to the mapping capability of the trained network. An additional bias term (θj) is provided to introduce a threshold for the activation of neurons. The input data (Xi) is presented to the network through the input layer, which is then passed to the hidden layer along with the weights. The weighted output (XiWij) is then summed and added to a threshold to produce the neuron input (Ij) in the output layer. This is given by: (1) This neuron input passes through an activation function f (Ij) to produce the desired output Yj. The most commonly used activation function is the logistic sigmoid function which takes the form; (2)
3 MATERIALS AND METHODS Experimental details pertaining to cultivation of micro organisms, media composition, preparation of immobilized packing media, experimental setup, biofilter operation and analytical techniques for data collection are given in our previously published work (Kim et al., 2007).
156
ELDON R. RENE, JUNG HOON KIM AND HUNG SUCK PARK
4 MODELING METHODOLOGY 4.1 MODEL INPUT-OUTPUTS A neural network based predictive model was developed with flow rate, inlet loading rate, pressure drop and inlet concentration as the model inputs and elimination capacity and removal efficiency as the outputs. 4.2 DATA DIVISION The experimental data was divided into training (NTr, 75%) and test data (NTe, 25%). The test data was set aside during network training and was only used for evaluating the predictive potentiality of the trained network. 4.3 ERROR EVALUATION The closeness of prediction between the experimental and model predicted outputs were evaluated by computing the determination coefficient values computed by the following formulae (Elias et al., 2006). (3)
4.4 DATA PRE-PROCESSING AND RANDOMIZATION Experimental data collected from the biofilter during the 67 days of continuous operation was randomized to obtain a spatial distribution of the data, which accounts for both steady state and transient steady state operation. The data was also normalized and scaled to the range of 0 to 1 using equation 4, so as to suit the transfer function in the hidden (sigmoid) and output layer (linear). (4) Where, is the normalized value, Xmin and Xmax are the minimum and maximum values of X respectively. 4.5 NETWORK PARAMETERS The internal parameters of the back propagation network namely epoch size, error function, learning rate (η), momentum term (μ), training count (Tc) and transfer function are to be appropriately selected to obtain the best network architecture that gives high predictions for the performance variables. In this study the number of neurons in the input layer (NI=4) and output layer (N0=2) were chosen based on the number of input and output variables to the network.
ARTIFICIAL NEURAL NETWORK BASED MODEL FOR EVALUATING PERFORMANCE
157
A detailed study on the effect of internal network parameters on the performance of back propagation networks and the procedure involved in selecting the best network topology has been described elsewhere (Maier and Dandy, 1998). However in most instances, literature suggests the use of a trial and error approach where the performance goal is set by the user. In this study the best values of the network parameters were chosen by carrying out simulations performed using the 2k full factorial design (Montgomery, 1991). The 2k design is of particular significance in exploring the effect of many factors on the response variable for a particular system. It provides the smallest number of runs with which ‘k’ factors can be studied in a complete factorial design (In this study, k=4, Hence 16 simulations were done – data not shown). Determination coefficient (R2) values were taken as the response variable and the setting that yielded the maximum R2 value in the test data was taken as the best network parameter. 4.6 SOFTWARES USED ANN based predictive modelling was carried out using the shareware version of the neural network and multivariable statistical modelling software, NNMODEL (Version 1.4, Neural Fusion, NY) and full factorial design was carried out by the statistical software MINITAB.
5 RESULTS AND DISCUSSIONS 5.1 EXPERIMENTAL The performance of the immobilized cell biofilter was monitored by varying the flow rate and inlet concentration. A step increase from low to high loading rates to the biofilter caused a few days to adapt to the new concentration and reach a new steady state value shortly. The results from this study are shown in Figure 1 as a function of the operating time, loading rate, EBRT and RE. These removal profiles indicated that the immobilized cells possessed good activity with steady and consistent removal even during the beginning of the experiments. The loading rate of NH3 was gradually increased to 2.5 g m–3 h–1 on the 14th day of continuous operation. The response was a sudden decline in the RE from 100% to 96% followed by a new steady state at the end of the 16th day where the RE was 98%. Hence, the loading rate was decreased to 1.7 g m–3 h–1 and subsequently increased in small time steps to a maximum of 4.5 g m–3 h–1. The biofilter RE profiles displayed minor ameliorating fluctuations due to step increase in loading rate between 1 and 4.5 g NH3 m–3 h–1. It is also evident that the RE was nearly 100% (>95%) up to a loading rate of 4.5 g m–3 h–1. However, after 60 days, when the ILR to the biofilter was increased significantly by varying both the concentration and flow rate to values as high as 7.5 g NH3 m–3 h–1, a noticeable decrease in the RE values from 100% to nearly 60% was observed. The critical NH3
158
ELDON R. RENE, JUNG HOON KIM
AND
HUNG SUCK PARK
loading rate to the biofilter was considered as 4.5 g NH3 m–3 h–1. Pressure drop values were sufficiently low during the operational time (0.1 - 1.7 cms of H2O) and did not cause any significant operational problem.
Figure 1. Time course profile of inlet loading rate and removal efficiency in the immobilized cell biofilter.
5.2 ANN BASED MODELING To model the performance of the biofilter, neural based simulations were carried out using the standard back error propagation network. The ranges of input and output parameters for the ANN model are given in Table 1. The experimental data collected from the biofilter was suitably divided into the training and test data set, pre-processed and randomized before carrying out simulations. The model was evaluated with the test data and the effect of network parameters on the R2 value was used as a measure to choose the best network architecture. Table 1. Range of input and output parameters used for training and testing ANN model developed to represent biofiltration of NH3 vapours Parameter Input Inlet concentration, ppm Flow rate, m3/h Inlet loading rate, g/m3.h Pressure drop, cms of H2O Output RE, % EC, g/m3.h
Training data, NTr-51 Min Max Mean
Testing data, NTe-16 Min Max Mean
10 6 0.3 0.1
150 16 7.5 1.5
63.3 9.25 3.08 1.1
20 6 0.6 0.2
150 16 7.5 1.5
74.1 9.13 3.54 1.16
60 0.3
100 5.3
97.2 2.93
66.8 0.5
100 5
93.2 3.18
159
ARTIFICIAL NEURAL NETWORK BASED MODEL FOR EVALUATING PERFORMANCE
5.2.1 EFFECT OF NETWORK INTERNAL PARAMETERS The different values of network internal parameters used to train the network are given in Table 2. During simulations with different combinations of settings as given by the experimental design, the following interpretations were made; (i) increasing the number of neurons in the hidden layer decreases the R2 value significantly (ii) an increase in the training count from low to high levels displays high R2 value for the model (iii) the effect of learning rate did not play a major role in increasing the R2 value, but it played a complementary role in speeding up the error convergence and (iv) the momentum term increased the R2 value when increased from lower to high levels. The best network architecture was then selected by observing high R2 value in the test data set (Table 3, For RE predictions, R2 value – 0.9825, for EC, R2 value – 0.9982). Table 2. Full 24 - factorial design for estimating the best network architecture. Parameters Neurons, NH Training count, Tc Learning rate, ηih Momentum term, μ Best R2 Error tolerance
Values 4 – 12 1000 – 16000 0.1 – 0.9 0.1 – 0.9 1 0.0001
Table 3. Best architecture obtained with different values of network internal parameters. NI 4
NH 4
NO 2
TC 16000
η 0.9
μ 0.9
5.2.2 PREDICTIVE CAPABILITY OF THE MODEL The RE and EC values predicted by the ANN model is illustrated in Figure 2 and 3 for the training data. It is quite apparent that, while predicting the RE and EC, the network was able to exactly map the data points. However, two or three data points were not adequately mapped by the network during training. This might have been caused by the step increase in loading rates where the microbes were reacclimatizing itself to attain new steady states. After training, the network was provided with the separate set of data for testing the developed model. The results
160
ELDON R. RENE, JUNG HOON KIM AND HUNG SUCK PARK
presented as EC and RE is illustrated in Figure 4 and 5 respectively. A comparison between the EC and RE values predicted by the model with the experimental values reveals the predictive capability of the model. The model was able to adequately identify the low and high peaks in the EC and RE values. The R2 values obtained during training and testing were greater than 0.98, which indicated that the predictions are accurate with best network architecture of 4-4-2.
Figure 2. Comparison of experimental and predicted values of removal efficiency during model training (NTr -51).
Figure 3. Comparison of experimental and predicted values of elimination capacity during model training (NTr -51).
ARTIFICIAL NEURAL NETWORK BASED MODEL FOR EVALUATING PERFORMANCE
Figure 4. Comparison of experimental and predicted values of removal efficiency during model testing (NTr -16).
Figure 5. Comparison of experimental and predicted values of elimination capacity during model testing (NTr -16).
161
162
ELDON R. RENE, JUNG HOON KIM AND HUNG SUCK PARK
6 CONCLUSIONS A laboratory scale immobilized cell biofilter evaluated to remove NH3 vapours showed RE higher than 90% at loading rates less than 4.5 g NH3/m3 h. This study explores the application of ANN as a performance prediction tool for a biofiltration process. The ANN model showed the ability to predict the extreme operating conditions and address the performance with R2 values greater than 0.98 for the training and test data set. The best network architecture (4-4-2) during effective training of the model was determined by 2k factorial design. The results from this study suggest that neural networks can capture and extract complex relations among the easily measurable parameters in a biofiltration process and predict the performance.
7 ACKNOWLEDGEMENTS The authors would like to acknowledge the research grants received from the Ulsan Regional Environmental Technology Research Centre in South Korea.
REFERENCES Chung, Y.-C., Huang, C. and Tseng, C.-P. (1996) Operation optimization of Thiobacillus thioparus CH11 biofilter for hydrogen sulfide removal. J. Biotechnol. 52: 31-38. Deshusses, M.A., Hamer, G. and Dunn, I.J.(1995) Behavior of biofilters for waste air biotreatment. 1. Dynamic-model development. Environ. Sci. Technol. 29: 1048-1058. Elías, A., Ibarra-Berastegi, G., Arias, R. and Barona, A. (2006) Neural networks as a tool for control and management of a biological reactor for treating hydrogen sulphide. Bioproc. Biosys. Eng. 29: 129-136. Jin, Y., Veiga, M.C. and Kennes, C. (2006) Performance optimization of the fungal biodegradation of α-pinene in gas-phase biofilter. Proc. Biochem. 41: 1722-1728 Kennes, C. and Veiga, M.C. (2001) Conventional Biofilters, In: Bioreactors for waste gas treatment, (Kennes, C. and Veiga, M.C, Eds.), Kluwer academic publishers, Netherlands, 47-98. Kim, J.H., Rene, E.R. and Park, H.S. (2007) Performance of an immobilized cell biofilter for ammonia removal from contaminated air stream. Chemosphere. 68: 274-280 Maier, H.R. and Dandy, G.C. (1998) The effect of internal parameters and geometry on the performance of back-propagation neural networks: an empirical study. Env. Mod. Soft. 13: 193-209. Maier, H.R. and Dandy, G.C. (2000) Neural networks for the prediction and forecasting of water resource variables: A review of modelling issues and applications. Env. Mod. Soft. 15: 101-124.
ARTIFICIAL NEURAL NETWORK BASED MODEL FOR EVALUATING PERFORMANCE
163
Montgomery, D C. (1991) Design and Analysis of Experiments, 3rd ed., Wiley, New York. Ottengraf. S.P.P. and van Den Oever, A.H.C. (1983) Kinetics of organic compound removal from waste gases with a biological filter. Biotechnol. Bioeng. 25: 3089-3102. Rene, E.R., Maliyekkal, S.M., Philip, L. and Swaminathan, T. (2006). Back-propagation neural network for performance prediction in trickling bed air biofilter. Int. J. Env. Poll, 28: 382-401. Rumelhart, D.E., Hinton, G.E. and Williams, R.J. (1986) Learning Internal Representations by Error Propagation. In: Rumelhart, D.E., McClelland (eds.): Parallel Distributed Processing. MIT Press, Cambridge. Shareefdeen, Z., Baltzis, B.C., Oh, Y.S. and Bartha, R. (1993) Biofiltration of methanol vapour. Biotechnol. Bioeng. 41: 512-524.
This page intentionally left blank
165
BIODESULPHURISATION AND BIOGAS TREATMENT
This page intentionally left blank
167
Development of a family of large-scale biotechnological processes to desulphurise industrial gasses ALBERT J.H. JANSSEN1,3, ROBIN VAN LEERDAM1, PIM VAN DEN BOSCH1, ERIK VAN ZESSEN2, GIJS VAN HEERINGEN3 AND CEES BUISMAN1 1 Sub-department of Environmental Technology, Wageningen University, Bomenweg 2, P.O. Box 8129, 6700 EV Wageningen, The Netherlands 2 Paques B.V., T. de Boerstraat 24, 8501 AB Balk, The Netherlands 3 Shell Global Solutions Int. B.V., Badhuisweg 3, 1031 CM Amsterdam, The Netherlands
ABSTRACT In this paper an overview is given of a new biotechnological process to remove hydrogen sulphide from gas streams. This process is jointly developed by Wageningen University, Delft University of Technology, Paques B.V. and Shell Global Solutions International B.V. In 1992, the first full-scale installation for H2S removal from biogas was taken into operation whilst in 2002 the first unit for high pressure natural gas desulphurisation was started-up. The removal of sulphur dioxide from flue gasses is feasible as well and in 2006 the first unit went on-stream in China. Currently, more than 75 full-scale plants are in operation worldwide. The formed bio-sulphur has a hydrophilic nature which enables its re-use, e.g. as a fertilizer or fungicide.
1 RELEASE OF SULPHUR-CONTAINING GASSES INTO THE ATMOSPHERE The increase in global population from 6 billion people in 2005 towards an expected number of 9 billion in 2050 is inevitably associated with a continued industrialization, urbanization and motorization. Without doubt, this will lead to a strong growth of the need for energy, especially in countries such as India and China that are on the eve of an industrial revolution. According to the International Energy Agency the energy consumption will increase with some 60 per cent in 2030 whilst in 2050 the global
167
168
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
energy consumption is expected to be three times higher than today. In the past, increased pollution effects were recorded in both industrial regions and on a global scale due to an increase in industrial activities (Earthwatch United Nations Environment Program, 1992). Some of these environmental problems were related to the emission of sulphur compounds, mainly as sulphur dioxide (SO2) that is formed from the combustion of sulphur containing fuel sources. For example, during a period of dense fog in the winter of 1952 more than 12,000 casualties were recorded in London as a result of the emission of SO2-containing flue gasses. Even today fossil fuel combustion accounts for approximately 90 per cent of the global man-made emission of SO2. Other major sources of SO2 are the processing of sulphide ores, oil refining and sulphuric acid production (Brimblecombe et al., 1989). In 1989, the global anthropogenic emission of sulphur to the atmosphere was estimated at 93 to over 200 million tons per year whereas over 90 per cent of all anthropogenic emissions of SO2 occur in the northern hemisphere (Aneja, 1990; Brimblecombe et al., 1989; Dignon and Hameed, 1989). In tropical regions natural emissions from soils, plants, the burning of biomass and volcanoes are believed to be the predominant sources of SO2 (UNEP, 1991). When released to the atmosphere, sulphur dioxide is a severe acidifying compound causing acid-rain. In case hydrogen sulphide is present in hydrocarbon containing gas streams, such as landfill-/biogas or natural gas, it has to be removed before incineration of the hydrocarbon stream because of its toxicity, corrosivity and bad smell. Emission of hydrogen sulphide into the atmosphere is therefore mainly a result of volcanic activities and evaporation from oceanic waters (Brimblecombe et al., 1989). In the atmosphere hydrogen sulphide causes acid rain due to its reaction with ozone to sulphuric acid. Fortunately, since the 1970s in many industrialized nations SO2 levels declined as a result of various emission control strategies such as selection of fuels with a low sulphur content, specialized combustion processes, coal desulphurisation and the development of new gas treatment processes. The main end-products from these processes are elemental sulphur and gypsum for respectively H2S and SO2 removal.
2 EXISTING REMOVAL METHODS FOR H2S REMOVAL 2.1 EXISTING REMOVAL METHODS FOR H2S Hydrogen sulphide (H2S) is a weak acid which dissociates into HS- (pK1=7.04) 2and S (pK2=11.96) with pK values at 18°C (Lide, 1995). The term «sulphide» is applied for all three entities. Removal of hydrogen sulphide from gases or wastewaters is required for reasons of health, safety and corrosion. The toxicity of hydrogen sulphide gas is well documented (Table 1). In the USA the worker exposure limits are 10 ppm (14 mg/m3) TWA (time weighted average) and 15 ppm (21 mg/m3) SEL (single exposure
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
169
limit). Hydrogen sulphide becomes progressively more dangerous as the H2S level incurs above toxic limits (70 ppm), becoming lethal at 600 ppm. Because of its detrimental characteristics, it is forbidden to release sulphide containing effluents and gaseous streams into the environment. Table 1. Hazard levels associated with releases of H2S. H2S concentration 1 ppm 10 ppm 20 ppm 100 ppm 200 ppm 500 ppm 700 ppm
Effects Rotten egg smell, odor complaints Occupational exposure limit for 8 hours Self-contained breathing apparatus required May cause headaches/nausea, sense of smell lost in 2-15 minutes Rapid loss of smell, burning eyes and throat Loss of reasoning and balance, respiratory distress in 200 minutes Immediate unconsciousness, seizures Without immediate resuscitation breathing will stop, leading to death
For the removal of H2S from gas streams, various well-established physicochemical techniques are available. Many of the processes that are in use at present may be grouped into the categories listed in Table 2. Liquid phase processes are used to remove H2S from sour gas streams, e.g. natural gas or refinery gas, via absorption down to levels that are typically around 410 ppm(v). As a result of steam addition to the loaded solvent, the extracted sulphide is released resulting in concentrated H2S streams (sometimes called acid gas). Obviously a second step is required to convert this into harmless elemental sulphur (S8). For H2S quantities above 20 tonnes per day, the Claus process is commonly applied whilst at low H2S loads (< 1 tonne per day) the acid gas is often sent to a flare to burn the H2S to sulphur dioxide. In the Claus process, 1/3 of the H2S is burned to SO2 whereafter the remainder reacts with the formed SO2 to elemental sulphur and water, according to: 2 H2S + SO2 ⇔ 3 S° + 2 H2O. As this is an equilibrium reaction, it is not possible to reach a full (i.e. 100%) H2S removal efficiency but in a line-up that comprises 3 consecutive process steps an overall efficiency of 96% is possible. In order to reach more than 99% H2S conversion a post treatment step, such as the SCOT or Superclaus process is required. Obviously, this will lead to a significant increase in the treatment costs (both capital and operating costs). In the range of 0.1-20 tonnes H2S per day, direct conversion with chelated iron (ferric) is carried out as in the Lo-Cat and SulFerox process. However, many of these units are plagued by plugging and foaming problems. Therefore the Gas Technology Institute (GTI, Des Plaines USA) concluded that no
170
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
existing technology was suitable for treating high pressure natural gas in the mediumscale range (O’Brien et al., 2007). Also the popularity of the Stretford process vanished as it became difficult to dispose of sulphur containing vanadium salts, a heavy metal that makes the sulphur product a hazardous waste. For low H2S loads, i.e. less than 0.1 tonnes per day, caustic scrubbing is still applied resulting in the formation of sulphide rich caustic streams that are quite difficult to dispose. These streams are regularly fed to biological wastewater treatment installations where they stimulate the growth of poor settling filamentous bacteria. Alternatively, iron sponge beds are used for H2S adsorption. However, due to the heat of reaction and the pyrophoric nature of the formed FeS considerable safety measures have to be taken. Table 2. Some examples of physico-chemical processes available for treating sour gases (Jensen and Webb, 1995). Category Liquid phase chemical reaction Liquid phase physical absorption
Example Amines Alkaline salts Sulfinol Selexol
Dry bed adsorption
Iron sponge Molecular sieve
Direct conversion
Stretford
Lo-Cat/SulFerox Claus SCOT
Superclaus Incineration
Reagents Alkanolamine Potassium carbonate Sulfolane and diisopropanolamine Dimethyl ether of polyethylene glycol Iron oxide Crystalline alkali-metal aluminosilicates Sodium carbonate, sodium vanadate, anthraquinone Iron complexes H2S and SO2, catalyst Cobalt/molybdenum catalyst and alkanolamine solvent catalyst Air-oxygen
Products H2S and CO2 H2S and CO2 H2S and CO2 H2S and CO2 S° S° S°
S° S° H2 S
S° SO2
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
171
3 NEW PROCESS FOR THE TREATMENT OF H 2S CONTAINING GAS STREAMS 3.1 INTRODUCTION Most important disadvantages attached to existing processes for H2S and SO2 removal are: • • • •
High investment costs make the existing process for H2S conversion only feasible for large-scale applications (Claus process). High maintenance costs and downtime are encountered due to foaming and plugging problems (Lo-Cat, SulFerox). An expensive post treatment method is needed for more than 99+ H2S removal efficiencies. High energy costs for pumping of limestone-gypsum slurry (SO2 removal).
As an alternative to the above mentioned existing technologies also microbiological processes for gas treatment are being considered. As these proceed around ambient temperatures and atmospheric pressure, the need for heat, cooling and pressurization power are not needed and thereby cut the energy costs to a minimum. One of the oldest and most commonly found applications of biological gas treatment involves the vent-air treatment at wastewater treatment plants. Worldwide more than 15,000 of these systems are in operation of which most belong to the ‘biotrickling’ and ‘biofilter’ type (Van Groenestijn, 2005). 3.2 THE BIOLOGICAL SULPHUR CYCLE Besides the Carbon- and Nitrogen cycle, the Sulphur cycle is important in nature. It has an oxidative and a reductive side which in a natural ecosystem should be in balance. On the reductive side, sulphate and sulphur function as an electron acceptor in the metabolic pathways, used by a wide range of anaerobic bacteria (Widdel, 1980). On the oxidative side of the cycle, reduced sulphur compounds serve as electron donors for anaerobic phototrophic bacteria or provide growth energy for the colorless sulphur bacteria (Robertson and Kuenen, 1991). The biological sulphur cycle is presented in Figure 1. On the oxidative side two different biotechnological processes can be distinguished for the removal of hydrogen sulphide. Firstly, genera of the family Chlorobiaceae and Chromatiaceae catalyze under anaerobic conditions, the photosynthetic Van Niel reaction (Niel van, 1932):
172
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
Figure 1. Biological sulphur cycle.
Cork et al. (1986) proposed a process, using the green sulphur bacteria Chlorobium limicola based on this equation. The major disadvantages in using photosynthetic bacteria on a large scale lie in their anaerobic nature and their requirement for radiant energy and hence extremely transparent solutions. Moreover, many of these organisms accumulate elemental sulphur internally, which would make the separation of sulphur and biomass difficult. Therefore, Kim et al. (1993) immobilized cells of Chlorobium limicola in strontium alginate beads in order to entrap the formed sulphur. Several researchers described the oxidation of sulphide to elemental sulphur or sulphate using chemolithoautotrophic bacteria belonging to the genus Thiobacillus (Beudeker et al., 1982; Buisman, 1990; Cadenhead and Sublete, 1990; Cho et al., 1992, Gadre, 1989; Ongcharit et al., 1991; Sublette and Gwozdz, 1991). Under sulphide limitation in the reactor, Thiobacilli can successfully compete with the chemical oxidation of sulphide because of their high affinity for this compound. Members of the genus Thiobacillus are classified as Gram-negative, rod-shaped, colorless sulphur bacteria which utilize reduced inorganic sulphur compounds as their energy source and CO2 as their main source of carbon. Some species are able to use organic carbon as a supplementary carbon source (Kelly, 1989; Robertson and Kuenen, 1991). Due to their simple nutritional requirements the use of chemolithotrophs for the removal of H2S is advantageous. Cho et al. (1992) extensively reported on H2S removal by the heterotrophic bacterium Xanthomonas sp. strain DY44. The specific H2S removal rate of this bacterium is however lower than those of purified Thiobacillus spp. (Cho et al., 1991). Furthermore, application of a heterotrophic organism is not
173
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
favorable if organic compounds are not readily available, e.g. in gas-purification. Table 3 provides examples of the many types found among the colorless sulphur bacteria, together with some of their environmental requirements. Table 3. Sulphur oxidizing bacteria (After: Schlegel, 1992). Species Thiobacillus thiooxidans Thiobacillus ferrooxidans Thiobacillus thioparus Thiobacillus denitrificans Thiobacillus intermedius Thiobacillus novellus Thiomicrospira pelophila Sulfolubus acidocaldarius a
pH of growth 2-5 2-6 6-8 6-8 2-6 6-8 6-8 2-3
Electron donor S2-, S2O32-, S Fe2+, S2O32-, S CNS-, S2O32-, S CNS-, S2O32-, S S2O32-, S, glutamate S2O32-, S, glutamate S2-, S2O32-, S S, glutamate, peptone
Typea o f o o f f o f
o, Obligately autotrophic; f, facultatively autotrophic
According to Kuenen (1982) the following two biological overall reactions may occur in an aerobic sulphide removal system:
Since the formation of sulphate yields most energy, this reaction is preferred by the micro-organisms. The formation of sulphur will only proceed under oxygenlimiting circumstances whereas sulphate is the main product in the presence of an excess amount of oxygen.(Buisman et al., 1991; Janssen et al., 1995). High sulphate levels are undesirable as they unbalance the natural sulphur cycle and can cause taste and alimentary problems in drinking water. Since the non-soluble sulphur can be removed, this process enables a reduction of the total sulphur content from the wastewater. Moreover, the sulphur can be re-used as a valuable raw-material in for instance soil-bioleaching processes (Tichy et al., 1994) or it can be purified. For this reason, the research carried out at the Sub-Department Environmental Technology at Wageningen University focusses on the development of ‘sulphur technologies’ that lead to elemental (bio)sulphur as the main end-product. The latest development is the use of halo-alkaliphilic sulphide oxidizing bacteria to limit the water intake. As these micro-organisms thrive at salt concentrations above 2 mol/L less dilution water is needed (Van den Bosch et al. 2007; Sorokin et al., 2005).
174
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
The reductive side of the biological sulphur cycle has been described extensively (Visser et al., 1993; Van Houten et al., 1994; Weijma et al., 2000; Lens et al., 2002; Lomans et al., 1999). In the presence of a suitable electron donor oxidized sulphur species such as thiosulphate and sulphate are reduced to hydrogen sulphide. Biological sulphide production is also commercially applied to remove metals from polluted groundwater or from aqueous (process) streams in the mine and metal industry by the formation of the corresponding metalsulphide precipitates (Dijkman et al., 2002). An example hereof is the sulphate-reducing bioreactor built by Paques B.V. at Pasminco (Budel, The Netherlands). The unit operates with hydrogen gas produced in a steam reformer (CH4 + 2H2O → CO2 + 4H2). 3.3 PROPERTIES OF BIOLOGICALLY PRODUCED SULPHUR In 1887, Winogradsky described the build-up and disappearance of sulphur inclusions by Beggiatoa, depending on whether or not the aqueous medium contained H2S (Trüper and Schlegel, 1964). Many authors have since described the formation and the properties of this «elemental» sulphur for both phototrophic bacteria (Guerrero et al., 1984; Hageage et al., 1970; Schmidt et al., 1971; Steudel et al., 1990; Strohl et al., 1981; Trüper and Hathaway, 1967) and aerobic Thiobacilli. (Jones and Benson, 1965; Steudel et al., 1989; Javor et al., 1990; Janssen et al., 1994). According to these reports, S° forms transparent droplets (globules) which are deposited inside or outside the cells. The droplets reach diameters of up to 1 mm, are of spherical or ellipsoidal shape and dissolve at least partly in various organic solvents like acetone, chloroform, ethanol and carbon disulfide. Biologically produced sulphur is hydrophilic and has a white or pale-yellow colour. The buoyant density of S° produced by Chromatium has been determined at 1.22 g·cm-3 (Guerrero et al., 1984). When allowed to stand in the liquid state or on drying, the sulphur globules eventually convert to crystalline S8. However, it has never been demonstrated that the sulphur globules consist of 100 percent sulphur. Surprisingly, many of the properties reported above do not match the properties of any known chemical allotrope of elemental sulphur, indicating that biologically produced sulphur is a not a standard sulphur form. The biologically produced sulphur (or ‘biosulfur’) is of oxidation state zero and is therefore often described as S0, although it should not be mistaken for atomic sulphur. In this paper we will use the term biologically produced sulphur to refer to sulphur produced in biotechnological H2S removal processes. The exact nature of the biologically produced sulphur particles produced in the process is not completely clear but it is likely that sulphur in the particles is present as S8 rings. Studies by X-ray absorption near edge spectroscopy (XANES) indicated that sulphur globules produced by chemotrophic bacteria, such as the dominating organism Thiobacillus sp. W5 (Visser et al., 1997), consist of either S8 rings or polythionates (Prange et al., 2002). Because polythionates are not stable at the slightly alkaline reactor conditions of the process, their presence
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
175
in the sulphur particles is not likely. X-ray diffraction studies showed that sulphur rings are present in the sulfur particles produced in this process and it is therefore most likely that the sulfur atoms are present as S8 rings (Janssen et al., 1994, 1999). Biologically produced sulphur is hydrophilic and dispersible in water, contrary to ‘inorganic’ elemental sulfur which is very poorly soluble in water (5 μg L-1) and strongly hydrophobic (Boulègue, 1978). The bio-sulphur particles consist of a core of sulphur, on the surface of which polymeric organic compounds such as proteins are adsorbed. These organic polymers give sterical and electrical stabilization of the colloidal particles. The hydrophilic properties of bio-sulphur become clear from Figure 2.
Figure 2. Hexadecane-water partition-test. Standard hydrophobic yellow ‘sulphur flower’ remains in the upper hexadecane-phase (left) whereas biologically produced sulphur remains in the lower water phase (right). (Taken from Janssen et al., 1999).
3.4 NEW PROCESS FOR H2S REMOVAL FROM GAS STREAMS In 1985 laboratory research was initiated at Wageningen University into applications of the biological sulphur cycle to mitigate environmental problems (Rinzema and Lettinga 1988; Buisman et al., 1990). After performing extensive pilot plant research Paques B.V. commercialised the first units for biogas desulphurisation and sulphate removal from groundwater. In the mid-nineties, the technology was further developed together with Shell Global Solutions Int. B.V. for a.o. high pressure natural gas and refinery gas desulphurisation.
176
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
This new process is based on the bioscrubber principle. First, the sulphur containing compounds are absorbed into the liquid phase whilst in the second process step the microbiological conversion to elemental sulphur and hydroxyl ions takes place, according to the following reaction equation:
It can be seen that the hydroxyl ions that are consumed in the first step are regenerated in the second process step. At excess oxygen conditions a complete oxidation to sulphate takes place, thereby leading to an acidification of the medium:
For this reason the process should be operated at oxygen-limiting conditions as described earlier (Janssen et al., 1995; Kleinjan et al., 2006). In the figure below a schematic representation of this new process for the desulphurisation of a.o. biogas, landfill gas and natural gas is given:
Figure 3. Biotechnological process for gas desulphurisation.
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
177
The gas enters a scrubber column and is desulphurized with a slightly alkaline fluid. The cleaned gas leaves the scrubber at the top section. The spent scrubber liquid is collected in the bottom of the scrubber and directed to the bioreactor. Here, air is dispersed at the bottom in order to enable the biomass to convert the dissolved sulphide into elemental bio-sulphur thereby regenerating caustic soda. The sulphur particles are separated by gravity settlement whereafter dewatering takes place, e.g. in a decanter centrifuge whilst the clear filtrate is returned to the reactor. The bioreactor effluent is recycled to the scrubber for renewed removal of H2S. The small sulfate production necessitates a continuous bleed from the unit that is taken from the bioreactor. The first full scale unit for high pressure natural gas treatment is located in Bantry (Alberta, Canada) near the town of Brooks and is owned and operated by EnCana Resources, a major Canadian and global gas producer. The natural gas is extracted from well sites that are on, or adjacent to the properties of over forty Canadian landowners around the Bantry North facility. The Shell-Paques biological technology was selected because it was the best available technology for this application. The alternative was acid gas re-injection, which was too expensive and therefore not attractive. The Shell-Paques unit is designed to meet a H S specification of less than 4 2 ppm volume on the treated natural gas.
Figure 4. First commercial full-scale unit for biotechnological H2S removal from high pressure natural gas. The Shell-Paques unit is designed to meet a H S specification 2 of <4 ppm volume on the treated natural gas.
178
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
The license agreement between EnCana Resources and New Paradigm was signed in the last week of November 2001. Basic engineering started directly in December and detailed engineering was completed by April 2002. Late April the constructor was selected and start was made with the construction of the facility on two skids. These skids were transported to site at the end of July. The construction and the tie-ins to the existing Bantry North infrastructure were done in August. The unit was handed over to the start-up team on 11 September and taken on stream on 12 September 2002. A photograph of the unit is shown below. As the installation is located in a remote area, dedicated tanks for make-up water and bleed water storage were required. Also a caustic storage tank is present and a compost filter for bioreactor vent-air treatment in case H2S is present in the vent-air during upset conditions. Monitoring of the unit provided a wealth of information from which we have prepared some graphs to show the most important parameters. The most important observations are shown in Figures 5 and 6.
Figure 5. H2S removal efficiency of first commercial full-scale unit for biotechnological H2S removal from high pressure natural gas.
From Figure 5, it follows that during a 1 month period the H2S removal efficiency was always above 99.5%. This makes this process very competitive to conventional physico-chemical processes. During this period the natural gas processing capacity fluctuated (data not shown). Despite these fluctuations, the H2S content in the treated gas was always less than 4 ppmv (Figure 6).
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
179
Figure 6. H2S concentration in untreated (sour) and treated (sweet) gas.
4 CONCLUSIONS Naturally occurring bacteria of the genus Thiobacillus can be used to remove H2S from gaseous streams whilst producing re-usable elemental sulphur. At present 5 full scale installation are in operation to treat high pressure natural gas and refinery gas. For the treatment of biogas and landfill gas around 80 installations are in operation, worldwide. Important advantages over existing technologies are: • • • • • • • • • • •
High removal efficiency for hydrogen sulfide from sour gas. High biological activity, so that peak load and other variables in the production processes can be handled effectively. Short system start-up time. Easily controlled process. Operation at ambient temperature. Operation at wide pressure range (0 to 80 bar). Very low operational costs. No sulfide containing waste stream. No use of chemical chelating agents. No hazardous bleed streams. Beneficial use of produced elemental sulfur (agricultural).
180
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
REFERENCES Aneja, V.P. (1990) Natural sulfur emissions into the atmosphere. J. Air Poll. Control Assoc. 40: 469-476. Beudeker, R.F., Gottschal, J.C. and Kuenen, J.G. (1982) Reactivity versus flexibility in thiobacilli. Antonie Leeuwenhoek. 48: 39-51. Boulègue, J. (1978). Solubility of elemental sulphur in water at 298 °K. Phosphorus and Sulphur 5: 127-128. Brimblecombe, P., Hammer, C., Rohde, H., Ryaboshapko, A. and Boutron, C.F. (1989). Human influence on the sulphur cycle. In: P. Brimblecombe and A. Yu Lein (Eds.) Evolution of the global biogeochemical sulphur cycle. Scope 39, pp 77-121. Wiley, New York. Buisman, C.J.N., Geraats, B.G., Ijspeert, P. and Lettinga G. (1990) Optimization of sulphur production in a biotechnological sulphide removing reactor. Biotechnol. Bioeng. 35: 50-56. Buisman, C.J.N., Ijspeert, P., Hof, A., Janssen, A.J.H., Ten Hagen, R. and Lettinga G. (1991) Kinetic parameters of a mixed culture oxidizing sulfide and sulfur with oxygen. Biotechnol. Bioeng. 38: 813-820. Cadenhead, P. and Sublette, K.L. (1990) Oxidation of hydrogen sulfide by Thiobacilli. Biotechnol. Bioeng. 35: 1150-1154. Cho, K.S., Zang, L., Hirai, M. and Soda M. (1991) Removal characteristics of hydrogen sulfide and methanethiol by Thiobacillus sp. isolated from peat in biological deodorization. J. Ferment. Bioeng. 71: 44-49. Cho, K.S., Hirai, M. and Shoda, M. (1992). Enhanced removal efficiency of malodorous gases in a pilot-scale peat biofilter inoculated with Thiobacillus thioparus DW44. J. Ferment. Bioeng. 71: 46-50. Cork, D.J., Jerger, D.E. and Maka, A. (1986) A biocatalytic production of sulfur from process waste streams. Biotechnol. Bioeng. Symp. Ser. 16: 149-162. Dignon, J. and Hameed, S. (1989) Global emissions of nitrogen and sulphur oxides. J. Air Poll. Control Assoc. 39: 180-186. Dijkman, H., Boonstra, J., Lawrence, R. and Buisman, C.J.N. (2002) Optimization of metallurgical processes using high rate biotechnology. Paper presented at the TMS 2002 annual meeting, Seattle, USA. Earthwatch United Nations Environment Program (1992) Chemical Pollution: A global Overview. The International Register of Potentially Toxic Chemicals and the Global Environment Monitoring System’s Monitoring and Assessment Research Centre, Geneva. Gadre, R.V. (1989) Removal of hydrogen sulfide from biogas by chemoautotrophic fixed-film bioreactor. Biotechnol. Bioeng. 34: 410-414. Groenestijn van, J.W. (2005) Biotechniques for air pollution control: past, present and future trends. Proceedings of the International Congress Biotechniques for Air Pollution Control. La Coruña, Spain, 5-7 October 2005, pp. 3-12.
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
181
Guerrero, R., Mas, J. and Pedros-Alio C. (1984) Buoyant density changes due to intracellular content of sulfur in Chromatium warmingii and Chromatium vinosum. Arch. Microbiol. 137: 350-356. Hageage, G.J. Jr., Eanes, E.D. and Gherna, R.L. (1970) X-Ray diffraction studies of the sulfur globules accumulated by Chromatium species. J. Bacteriol. 101: 464-469. Houten Van, R.T., Hulshoff Pol, L.W. and Lettinga, G. (1994) Biological sulphate reduction using gas-lift reactors fed with hydrogen and carbon dioxide as energy and carbon source. Biotechnol. Bioeng. 44(5): 586-594. Janssen, A.J.H., De Keizer, A. and Lettinga, G. (1994) Colloidal properties of a microbiologically produced sulphur suspension in comparison to a LaMer sulphur sol. Colloids Surfaces B: Biointerfaces 3: 111-117. Janssen, A.J.H., Sleyster, R., Van der Kaa, C., Jochemsen, A., Bontsema, J. and Lettinga, G. (1995) Biological sulphide oxidation in a fed-batch reactor. Biotechnol. Bioeng. 47: 327-333. Janssen, A.J.H., Lettinga, G. and de Keizer, A. (1999) Removal of hydrogen sulphide from wastewater and waste gases by biological conversion to elemental sulphur. Colloidal and interfacial aspects of biologically produced sulphur particles. Colloid. Surf. A-Physicochem. Eng. Asp., 151, 389. Janssen, A.J.H., Buisman, C.J., Kijlstra, W.S. and Grinsven, P.F.A. van (1999) The Shell-Paques desulfurisation process for H2S removal from high pressure natural gas, synthesis gas and Claus tail gas. In: Proceedings Ninth Gas Research Institute Sulfur Recovery Conference, October 24-27, 1999, San Antonio, Texas. Javor, B.J., Wilmot, D.B. and Vetter, R.D. (1990) pH-Dependent metabolism of thiosulfate and sulfur globules in the chemolithotrophic marine bacterium Thiomcrospira crunogena. Arch. Microbiol. 154: 231-238. Jensen, A.B. and Webb, C. (1995). Treatment of H2S-containing gases: A review of microbiological alternatives. Enzyme Microb. Technol. 17: 2-10. Jones, G.E. and Benson, A.A. (1965) Phosphatidyl glycerol in Thiobacillus thiooxidans. J. Bacteriol. 89: 260-261. Kelly, D.P. (1989) Physiology and biochemistry of unicellular sulfur bacteria, In: Autotrophic Bacteria. H.G. Schlegel and B. Bowien (Ed.), 193-213. Springer-Verlag, Berlin. Kleinjan, W.E., De Keizer, A. and Janssen, A.J.H. (2003) Biologically produced sulfur. Top. Curr. Chem. 230: 167-188. Kleinjan, W.E., Lammers, J.N.J.E., De Keizer, A. and Janssen A.J.H. (2006) Effect of biologically produced sulfur on gas absorption in a biotechnological hydrogen sulfide removal process. Biotechnol. Bioeng. 94(4): 633-644. Kuenen, J.G. (1975) Colourless sulphur bacteria and their role in the sulphur cycle. Plant Soil 43: 49-76. Lens, P., Vallero, M., Esposito, G. and Zandvoort, M. (2002) Perspectives of sulfate reducing bioreactors in environmental biotechnology. Reviews in Environmental Science and Bio/ Technol. 1(4): 311-325.
182
A. J. H. JANSSEN, R. VAN LEERDAM, P. VAN DEN BOSCH, E. VAN ZESSEN, G. VAN HEERINGEN AND C. BUISMAN
Lide, D.R. (Ed.) (1995) Handbook of Chemistry and Physics, 75th edition, CRC Press Boca Raton, M.A. Lomans B.P., Camp, H.J.M. op den, Pol, A. and Vogels, G.D. (1999) Anaerobic versus aerobic degradation of dimethyl sulfide and methanethiol in anoxic freshwater sediments. Appl. Environ. Microbiol. 65: 438-443. Niel van, C.B. (1932) On the morphology and physiology of the purple and green sulphur bacteria. Arch. Microbiol. 3: 1-112. O’Brien, M., Wentworth, C., Lanning, A. and Engert, T. (2007) Shell-Paques® Bio-Desulfurization process directly and selectiveley removes H2S from high pressure natural gas – start up report. In: Proceedings of the 57th Annual Laurance Reid Gas Conditioning Conference, University of Oklahoma, Norman, Oklahoma. Ongcharit, C., Sublette, K.L. and Shah, Y.T. (1991) Oxidation of hydrogen sulfide by flocculated Thiobacillus denitrificans in a continuous culture. Biotechnol. Bioeng. 37: 497-504. Prange, A., Chauvistre, R., Modrow, H., Hormes, J., Trüper, H.G. and Dahl, C. (2002) Quantitative speciation of sulfur in bacterial sulfur globules: X-ray absorption spectroscopy reveals at least three different species of sulfur. Microbiology. 148: 267. Rinzema, A. and Lettinga, G. (1988) Anaerobic treatment of sulphate containing waste water, pp. 65-109. In D.L. Wise (Ed.)., Biotreatment systems, 3. CRC Press, Boca Raton, Fl. Robertson, L.A. and Kuenen, J.G. (1991) The colorless sulfur bacteria, 385-413. In: A. Balows, H. Trüper, M. Dworkin, W. Harder, K.-H. Schleifer (Eds.), The prokaryotes, 2 nd edition, Springer-Verlag, New York. Schlegel, H.G. (1992), General Microbiology, 7th ed., Cambridge University Press, U.K. Schmidt, G.L., Nicolson, G.L. and Kamen, M.D. (1971) Composition of the sulfur particle of Chromatium vinosum strain D. J. Bacteriol. 105: 1137-1141. Sorokin, D.Y., Banciu, H., van Loosdrecht, M. C. M. and Kuenen, J.G. (2003) Growth physiology and competitive interaction of obligately chemolithoautotrophic, haloalkaliphilic, sulfuroxidizing bacteria from soda lakes. Extremophiles. 7: 195-203. Steudel, R. (1989). On the nature of the «Elemental Sulfur» (S°) produced by sulfur oxidizing bacteria- a model for S° globules. in: Autotrophic Bacteria. H.G. Schlegel and B. Bowien (Eds.), Science Tech Publishers, Madosin, WI. Steudel, R., Holdt, G., Visscher, P.T. and Van Gemerden, H. (1990) Search for polythionates in cultures of Chromatium vinosum after sulfide incubation. Arch. Microbiol. 153: 432-437. Steudel, R., Göbel, T. and Holdt, G. (1988) The molecular composition of hydrophilic sulfur sols prepared by acid decomposition of thiosulfate. Z. Naturforsch. 43b: 203-218. Strohl, W.R., Geffers, I. and Larkin, J.M. (1981) Structure of the sulfur inclusion envelopes from four Beggiatoas. Current Microbiology. 6: 75-79. Sublette, K.L. and Gwozdz, K.J. (1991) An economic analysis of microbial reduction of sulfur dioxide as a means of byproduct recovery from regenerable processes for flue gas desulfurization. Appl. Biochem. Biotech. 28/29: 635-646.
DEVELOPMENT OF A FAMILY OF LARGE-SCALE BIOTECHNOLOGICAL
183
Tichý, R., Janssen, A., Grotenhuis, J.T.C., Lettinga, G. and Rulkens, W. (1994) Possibilities for using biologically-produced sulphur for cultivation of Thiobacilli with respect to bioleaching processes. Biores. Technol. 48: 221-227. Trüper, H.G. and Schlegel, H.G. (1964) Sulphur metabolism in Thiorhodaceae. I. Quantitative measurements on growing cells of Chromatium okenii. Antonie Leeuwenhoek. 30: 225-238. Trüper, H.G. and Hathaway, J.C. (1967) Orthorhombic sulphur formed by photosynthetic sulphur bacteria. Nature. 215: 435-436. UNEP (1991) United Nations Environment. Programme Environmental Data Report, Third Edition 1991/92, Basil Blackwell, Oxford. Van den Bosch, P.L.F., Van Beusekom, O.C., Buisman, C.J.N. and Janssen, A.J.H. (2007) Sulfide oxidation under halo-alkaline conditions in a fed-batch bioreactor. Biotechnol. Bioeng. (accepted for publication). Visser, A. Gao, Y. and Lettinga, G. (1993) Effects of pH on methanogenesis and sulphate reduction in thermophilic (55 C) UASB reactors. Biores. Technol. 44: 113-121. Visser, J.M., Stefess, G.C., Robertson, L.A. and Kuenen, J.G. (1997) Thiobacillus sp. W5, the dominant autotroph oxidizing sulfide to sulfur in a reactor for aerobic treatment of sulfidic wastes. Antonie van Leeuwenhoek. 72: 127. Weijma, J., Haerkens, J.-P., Stams, A.J., Hulshoff Pol, L.W. and Lettinga, G. (2000) Thermophilic sulfate and sulfite reduction with methanol in a high rate anaerobic reactor. Water Sci. Technol. 42(5-6): 251-258.
This page intentionally left blank
185
Study of a desulfurization process to convert dibenzothiophene to 2-hydroxybiphenyl by Rhodococcus rhodochrous NRRL (B-2149) ALDO B. SOARES JÚNIOR, YANNE K. P. GURGEL, BRUNA M. E. CHAGAS, THYRONE B. DOMINGOS, GORETE R. MACEDO AND EVERALDO S. SANTOS Biochemical Engineering Laboratory, Chemical Engineering Department, Federal University of Rio Grande do Norte, Campus Universitário, 59072 – 970, Natal, Brazil
ABSTRACT Actually, a great effort is being given to research on biodesulfurization processes, i.e., processes in which sulfur can be removed selectively from sulfur-compound moities without altering its British thermal unit. This effort relies on the fact that fossil fuels (coal and oil) contain organic sulfur compounds that are released after combustion to the environment mainly as SO2 that causes acid rain becoming a potential pollutant. In this work we investigate the biodesulfurization of a model molecule that represents the main class of a group of recalcitrant compounds found in petroleum, Dibenzothiophene (DBT), to produce 2-Hydroxybiphenyl (2-HBP), a sulfur-free compound, by Rhodococcus rhodochrous (NRRL B-2149) using the 4S pathway. Experiments in which R. rhodochrous (NRRL B- 2149) was cultived during exponential growth phase using glucose and DBT as carbon and energy and sulfur sources, respectively, showed that the microorganism follows the 4S metabolic pathway in which DBT is converted to 2-HBP and sulfite. It was also showed that R. rhodochrous (NRRL B-2149) has cell-bounding surface active agents that that facilitates the emulsification of the apolar – water immiscible DBT.
1 INTRODUCTION Biodesulfurization is a process in which microorganisms are used, under control conditions, to remove sulfur containing compounds from oil and coal. Biological and chemical-physical factors affect considerably heavy oil biodegradation as well as biotransformation such biodesulfurization (Setti et al., 1995). A variety of
185
186
A. B. SOARES JUNIOR, Y. K. P. GURGEL, B. M. E. CHAGAS, T. B. DOMINGOS, G. R. MACEDO AND E. S. SANTOS
microorganisms can use sulfur from aromatic hydrocarbons sources mainly bacteria such as Mycobacterium sp. (Srinivasaraghavan et al., 2006), Rhodococcus erythropolis (Oda and Otha, 2002) as well as Microbacterium Strain ZD-M2 (Li et al., 2005). Dibenzothiophene (DBT) has been used as the model compound (Figure 1) for sulfur hetero-cycles present in hydrodesulfurization treated fuel, i.e., as a model that represents the so called recalcitrant compounds such as alkyled-thiophenes, benzonthiophenes, etc.
Figure 1. Dibenzothiophene.
Even though a significant number of microorganisms have been found to remove sulfur from DBT via a hydrocarbon degradative pathway like Kodama pathway (Kilbane and Jackowsky, 1992) such a means of sulfur removal involves the break of carbon-carbon bonds in the ring thus resulting in a reduction of fuel value. Therefore this route it is not interesting for reducing sulfur levels of compounds without loss its British termal unit. Among the microorganisms able to use sulfur from DBT, a small number have been shown to remove it via a specific-pathway so called «4S» pathway. Rhodococcus erythropolis IGTS8 as already shown follows «4S» pathway once this microorganism has three catabolic genes (dszA, dszB and dszC) that are responsible for DBT desulfurization and that are clustered on a 120-kb linear plasmid. In this pathway (Figure 2), enzyme DszC catalyzes two consecutive mono-oxigenation reactions converting DBT to DBTO2 (DBT-sulfone), in this case DszC uses NADH and FMNH2 as cofactors. Following, a flavomonooxygenase (DszA) converts DBTO2 to Hydroxyphenil benzene sulfonate (HPBS) that is converted to 2-Hydroxybiphenyl (2-HBP) and sulfite by a HPBS-desulfinase (DszB). 2-HBP is not further metabolized and accumulates in the medium. Inorganic sulfur released to the medium can be used by the strain as the sole source of sulfur. It is known that expression of the dsz gene cluster is strongly repressed by sulfate and sulfur-containing aminoacids. On the other hand, microorganisms are known to produce a special classes of a molecules of high- and low-molecular weight called biosurfactants. Biosurfactants are molecules that have a hydrophilic portion, which may consist of mono-, oligo- or polissacharydes, amino acids or peptides or carboxylated or phosphate groups, and a hydrophobic portion composed, mainly, by saturated or unsaturated (hydroxyl) fatty
STUDY OF A DESULFURIZATION PROCESS TO CONVERT DIBENZOTHIOPHENE
187
acids or fatty alcohols (Lang and Wullbrandt, 1999). These molecules reduce surface tension and Critical Micelle Dilution (CMD) in both aqueous solution and hydrocarbon mixtures. Investigation on influence of surface-active agent in cultivation medium using strain 1awq (further identified as R. erythropolis) has shown that there were no surfactant-like molecules produced in the culture medium (Feng et al., 2006).
Figure 2. «4S» pathway for DBT desulfurization.
In this work we investigate the biodesulfurization of Dibenzothiophene (DBT), to produce 2-Hydroxybiphenyl (2-HBP), a sulfur-free compound, by Rhodococcus rhodochrous (NRRL B-2149) using the «4S» pathway.
2 MATERIALS AND METHODS 2.1 MICROORGANISM AND CULTIVATION The microorganism used was Rhodococcus rhodochrous (NRRL B - 2149) supplied by the National Center for Agricultural Utilization Research (USA). The strain was maintained at -20 °C in a vial containing a 10% glycerol solution. This strain is able to degrade cholesterol.
188
A. B. SOARES JUNIOR, Y. K. P. GURGEL, B. M. E. CHAGAS, T. B. DOMINGOS, G. R. MACEDO AND E. S. SANTOS
Pre-inoculum was prepared in LB medium (containing 1% tryptone, 0.5% yeast extract, and 1% NaCl) in which cells from the vial were incubated in a 50 mL Erlenmeyer flask in an orbital shaker during 12h at 150 rpm and 32 °C. For cultivation 10% (vol/vol) of R. rhodochrous (at exponential growth) was incubated in 50 mL of a 250 mL Erlenmeyer flask in an orbital shaker during 60 at 150 rpm and 32 °C using the following medium: NaH2PO4·H2O, 4 g/L; K2HPO4, 4 g/L; MgCl2·6H2O, 0.0245 g/ L; CaCl2·2H2O, 0.001 g/L; FeCl3·6H2O, 0.001 g/L; NH4Cl, 2 g/L; glucose, 20 g/L and the model compound DBT (0.2%). Due to its low water solubility DBT was diluted in ethanol. During cultivation samples were drawn for determination of cell concentration and after centrifugation supernatant was used to determine carbohydrate content, total protein, DBT and 2-HBP contents, pH, surface tension as well as critical micelle dilutions (CMD-1 and CMD-2). 2.2 CARBOHYDRATE DETERMINATION The reducing sugars formed during the cultivation were estimated spectrophotometrically at 600 nm by using the Dinitrosalicyclic acid (DNS) method with glucose as the standard. 2.3 CELL CONCENTRATION Biomass concentration (g/L) was estimated by optical density measurement at 600 nm using a standard curve. 2.4 PROTEIN CONTENTS Determination of total protein was carried out according to the Sedmak and Grossberg modified method as described by Santos (2001). 2.5 DBT AND 2-HBF ASSAYS The concentration of DBT and 2-HBP were analyzed by HPLC (Shimadzu, Japan), using a UV detector at 236 nm. A Shim-Pack C-18 column (150 mm × 4.6 mm, with 5 μm particles and 100Å pore size, Shimadzu) was used. The mobile phase was 50% of acetonitrile/water (ultra pure) with a flow rate of 1.0 mL/min and 25 °C temperature. The retention time of HBP was 7.4 min and 32 min, respectively. 2.6 PH MEASUREMENT pH measurement was carried out using a potenciometer (Model Digimed DM21, Brazil).
STUDY OF A DESULFURIZATION PROCESS TO CONVERT DIBENZOTHIOPHENE
189
2.7 SURFACE TENSION Surface tension was determined with a Du Noüy Tensiometer (Central Scientific Company, USA) using the ring method. 2.8 CRITICAL MICELLE DILUTIONS (DCM-1 AND DCM-2) Critical micelle dilutions (CMD-1 and CMD-2) were determined by measuring the surface tension of 10-times and 100-times diluted broth in distilled water at room temperature, respectively.
3 RESULTS AND DISCUSSION 3.1 BIODESULFURIZATION OF DBT BY R. RHODOCHROUS (NRRL B- 2149) The capacity of biodesulfurization of DBT by R. rhodochrous (NRRL B- 2149) in a cultivation in which DBT was used as sulfur source has been investigated.
Figure 3. Biodesulfurization of DBT by R. rhodochrous (NRRL B- 2149).
According to Figure 3, it is possible to observe that R. rhodochrous (NRRL B2149) is able to use the «4S» pathway in order to desulfurize DBT to produce 2-HBP. It can be seen that R. rhodochrous (NRRL B- 2149) grew during cultivation reaching about 1.7 g/L at 60 h without any significant change at cultivation pH. Corybacterium sp Strain SY1, a strain that can use the «4S» pathway, showed a 2-HBP (0.18 mM)
190
A. B. SOARES JUNIOR, Y. K. P. GURGEL, B. M. E. CHAGAS, T. B. DOMINGOS, G. R. MACEDO AND E. S. SANTOS
concentration maximum at 65 h (Omori et al., 1992). In our work maximum 2-HBP production occurred at 36 h (0.11 mM) suggesting that it can be used as an intermediary compound to another reaction. It is known that 2-HBP inhibits cells growth in the biodesulfurization process, for instance, 0.5 mM 2-HBP inhibited growth of R. erythropolis D-1 (Oshiro et al., 1996) while for R. rhodochrous (NRRL B- 2149) biomass stationary phase occurred at 0.1 mM. Reducing sugars measurement showed a reduction at glucose concentration reaching about 14.0 g/L at the cultivation end. Protein assay at supernatant (cell-free) cultivation showed a slightly increase in extracellular protein during cultivation time. In this case, about 0.013 g/L protein content was found at 60h. 3.2 SURFACE TENSION DURING DBT BIODESULFURIZATION In order to observe any presence of surface-active agent during cultivation of DBT as a sulfur source by R. rhodochrous (NRRL B- 2149) surface tension of cell free culture medium as well as CMD-1 and CMD-2 were assayed.
Figure 4. Surface tension during desulfurization of DBT by R. rhodochrous (NRRL B- 2149).
Figure 4 shows that R. rhodochrous (NRRL B- 2149) is not able to produce any surface-active agent in order to reduce the surface tension at cell free cultivation medium. However, a slightly increase on surface tension is observed for CMD-1 during
STUDY OF A DESULFURIZATION PROCESS TO CONVERT DIBENZOTHIOPHENE
191
the period of increase of 2-HBP concentration in the cultivation. It is known that 2HBP has a good solubility in water if compared to DBT. A significant surface tension occurred for CDM-2 during the initial time of cultivation after that surface tension approaches to the water surface tension, i.e., about 72 dyn/cm. Van Hamme and Ward (2001) found that Rhodococcus sp. Strain F9-D79 could not produce any biosurfactants at cultivation medium when growing on oil/water interface. However, Strain F9-D79 showed a reduction at surface tension for a non-free cell medium. Therefore it has been supposed that R. rhodochrous (NRRL B-2149) has cell-bounding surface active agents that that facilitate the emulsification of the apolar – water immiscible DBT since none significant surface-activity was observed at cell-free cultivation medium.
4 CONCLUSIONS Experiments in which R. rhodochrous (NRRL B- 2149) was cultivated during exponential growth phase using glucose and DBT as carbon and energy and sulfur sources, respectively, showed that the microorganism follows the 4S metabolic pathway in which DBT is converted to 2-HBP. Maximum 2-HBP production occurred at 36 h (0.11 mM) suggesting that it can be used as an intermediary compound to another reaction. 2-HBP inhibited cells growth at 0.1 mM. It was also showed that R. rhodochrous (NRRL B-2149) has cell-bounding surface active agents that that facilitates the emulsification of the apolar – water immiscible DBT.
5 ACKNOWLEDGEMENTS Authors thank CNPq-CTPETRO (Brazil) for financial supporting (Project 50485904/8).
REFERENCES Feng, J., Zeng, Y., Ma, C., Cai, X., Zhang, Q., Tong, M., Yu, B. and Xu, P. (2006) The surfactant tween 80 enhances biodesulfurization. Appl. Environ. Microbiol. 72(11): 7390-7393. Kilbane, J.J. and Jackoswki, F. (1992) Biodesulfurization of water soluble coal-derived material by Rhodococcus rhodochrous IGTS8. Biotechnol. Bioeng. 40(9): 1107-1114. Lang, S. and Wullbrandt, D. (1999) Rhamnose lipids – biosynthesis microbial production and application potential. Appl. Microbiol. Biotechnol. 51(1): 22-32.
192
A. B. SOARES JUNIOR, Y. K. P. GURGEL, B. M. E. CHAGAS, T. B. DOMINGOS, G. R. MACEDO AND E. S. SANTOS
Li, W., Zhang, Y., Wang, M.D. and Shi, Y. (2005) Biodesulfurization of dibenzothiophene and other organic sulfur compounds by a newly isolated Microbacterium Strain ZD-M2. FEMS Microb. Lett. 247: 45-50. Oda, S. and Ohta, H. (2002) Biodesulfurization of dibenzothiophene with Rhodococcus erythropolis ATCC 53968 and its mutant in an interface bioreactor. Jour. Biosc. Bioeng. 94(5): 474-477. Omori, T., Mona, L., Saiki, Y. and Kodama, T. (1992) Biodesulfurization of dibenzothiophene by Corynebacterium Strain SY1. Appl. Environ. Microbiol. 58(3): 911-915. Oshiro, T., Suzuki, K. and Izumi, Y. (1996) Regulation of dibenzothiophene degrading enzyme activity of Rhodococcus erythropolis D-1, J. Ferment. Bioeng. 81(2): 121-124. Santos, E. S., Guiradello, R. and Franco, T.T. (2002) Preparative chromatography of xylanase using expanded bed adsorption. J. Chromat. 944(1): 217-224. Setti, L., Lanzatini, G. and Pifferi, P.G. (1995) Dibenzothiophene biodegradation by a Pseudomonas sp. in model solutions. Proc. Biochem. 30(8): 721-728. Srinivasaraghavan, K., Sarma, P.M. and Lal, B. (2006) Comparative analysis of phenotypic and genotypic characteristic of two desulfurizing bacterial strains, Mycobacterium phlei SM1201-1 and Mycobacterium phlei SM1201-1 GTIS10. Lett. Appl.. Microbiol. 42(5): 483-489. Van Hamme, J.D. and Ward, O.P. (2001) Physical and metabolic interactions of Pseudomonas sp. Strain JA5-B45 and Rhodococcus sp. strain F9-D79 during growth on crude oil and effect of a chemical surfactant on them. Appl. Environ. Microbiol. 67 (10): 4874-4879.
193
Control of methane emissions issuing from landfills: the Canadian case JOSIANE NIKIEMA AND MICHÈLE HEITZ Chemical Engineering Department, Faculty of Engineering, Université de Sherbrooke; 2500, Boulevard de l’Université, Sherbrooke, (Québec), J1K 2R1, Canada
ABSTRACT During their storage in landfills, wastes are biodegraded, which results in the production of biogas and leachate. Over recent years, the handling of the leachate product has become one of major concern. However, in the case of biogas product, elimination or valorization processes are applied in a smaller proportion, even if the methane emissions, directly related to landfills, are some 25 % of the total anthropogenic methane emissions. Indeed, many older or smaller landfills are deprived of gas collection systems, thereby making impossible the application of gas combustion and/or valorization methods. Therefore, other processes have to be considered, e.g., the biofiltration of methane. In this paper, the results of an experiment, undertaken to confirm the stability of the biofiltration system that has been developed at Université de Sherbrooke by the Biocom group, are presented. At a methane inlet concentration of around 7500 ppmv and a gas flow rate of 0.25 m3/h, the conversion of the biofilter can be maintained at 22 % unchanged for a period of 150 days or more. Even after the cessation of methane feeding and biofilter irrigation for some 2 weeks, the biofilter performance was able to be restored, in only one week, to the same operating level as it was maintained before the deliberate shutdown.
1 INTRODUCTION 1.1 WASTES IN CANADA During the last decade, it is to be noted that a continuous increase in the total amount of wastes generated in Canada took place. The total waste generated during the year 2004 was > 33 million metric tons (that is to say 790 metric tons per inhabitant, being 3 % higher than in 2002), of which 10 million metric tons originated from
193
194
JOSIANE NIKIEMA AND MICHÈLE HEITZ
residential sources. The remaining part arose from commercial and public institutions (50 %), and other sources. These wastes are mainly composed of organic materials (40 % m/m), papers (26 % m/m), plastics (9 %), glass, metals and others (Buchanan et al., 2007). Various reasons are advanced to explain this situation: the population increase, the increases in goods’ consumption, caused in part by the higher levels of incomes, the changes in society that have resulted, for example, in increased need for nonreusable products, and the continuous evolution of technologies (Cameron et al., 2005). Wastes can be substantially eliminated through recycling efforts (papers, plastics, metals and glass materials), composting and anaerobic digestion (organic wastes), thermal treatments (part of energy generation), and finally, by landfilling. Increasingly, wastes are now being valorized through various recycling processes (e.g. 40-50 % of residential wastes may be recycled) but in Canada, the main way of disposing of wastes is still through landfilling, which affected, in 2005, around 75 % of the wastes arising during the same period. Indeed, about 10000 landfills (active and inactive, all of them requiring attention) presently exist in Canada (Nikiema et al., 2007). The majority of them (83 %) are public institutions but private landfills also exist. Public landfills receive around 56 % of the wastes, private landfills receive the remainder. It must also to be mentioned that most landfills are now rather old. Indeed, some 30 % of landfills had, in 2002, a remaining useful life of less than 10 years. However, these landfills were still receiving more than half of the currently generated wastes (Cameron et al., 2005). 1.2 THE BIOGAS AND THE LEACHATE During the degradation of wastes stored in landfills, a leachate and a biogas, both of which need to be handled, are generated. The quantities and compositions of these materials are influenced by various factors such as the types of wastes, their ages, etc. (Trebouet et al., 2001). The leachate is mainly composed of water, in which soluble and solid pollutant particules, are present, including minerals, e.g. iron, and organic matters (COD, up to 70000 mg/L; BOD, up to 56000 mg/L and TKN, up to 2000 mg/L) (Sanphoti et al., 2006; Tränkler et al., 2005; Visvanathan et al., 2007). The leachate is generated during dehydratation of the biodegrading wastes or when rain water passes through the wastes; this explains its wide variations in composition. In Canada, in 2000, 46 % of municipal landfills were equipped with membranes, limiting the infiltration of external water into the site, while 18 % of landfills had collection systems installed for the leachate. These two kinds of landfills handled some 75 % of the total of landfilled wastes arising in the same year (Cameron et al., 2005). The other landfill product, the biogas, is a mixture of gases, composed principally of methane (30-70 % V/V), carbon dioxide (20-50 % V/V) (both greenhouse gases),
CONTROL OF METHANE EMISSIONS ISSUING FROM LANDFILLS
195
along with various sulphur compounds, volatile organic compounds, and others. Because of its high heating value during the landfill’s early years (usually half that of natural gas), the biogas can be valorized, if collected, as an energy source (Nikiema et al., 2007). In as much as is feasible, Canada Government policy encourages the energy recovery of biogas through combustion. There are some 50 landfills in Canada that collect their biogas and at least 30000 metric tons/day of methane are burned on each site. However, for the older or smaller landfills (i.e. < 200000 m3 of capacity), these are frequently deprived of gas collection systems, and thus valorization methods cannot be reasonably applied. Therefore, in order to avoid important methane emissions to atmosphere, biological processes, such as biofiltration, may be applied. 1.3 THE METHANE BIOFILTRATION Biofiltration is performed within a triphasic reactor, packed with stationary filter material, in which growth of the microorganisms is favoured (Delhoménie and Heitz, 2005). Indeed, methanotrophs are able to biodegrade the methane pollutant, and then generate, as in all biological processes, new biomass, salts, water and carbon dioxide, the latter product to a lesser extent than occurs in chemical oxidation processes, as presented in Equation 1. (1) Experiments conducted to date have confirmed that biofiltration is deemed suitable for the direct elimination of methane on landfill sites. The Biomet group, located at Université de Sherbrooke (Sherbrooke, Canada), has for more than 5 years, conducted research on the problematic of the methane issuing from landfills. The main interest was to determine the relationship between the measured biofilter performance and some operating parameters, including the concentrations of inlet methane and the nitrogen present in the nutrient solution. For example, during previous experiments, the nitrogen concentration, required for the proper operation of the biofilter, has been optimized and appears to be 0.75 g/L for a methane inlet concentration of between 7000 and 7500 ppmv (Nikiema et al., 2005). The objective of this present study has been to confirm the stability of the biofiltration system, and its capacity to remain as efficient as it was initially, even after a substantial period of time (nearly one year).
2 MATERIALS AND METHODS The lab-scale biofiltration system employed in this study is presented in Figure 1. It is mainly composed of 3 sections: 1) the polluted air generation section (humidified air
196
JOSIANE NIKIEMA AND MICHÈLE HEITZ
and pure methane (99 % V/V of purity); gas inlet temperature: ~ 20°C; inlet concentration of methane: 7500 ppmv; total gas flow rate: 0.25 m3/h); 2) the biofiltration system, which is composed of an up-flow biofilter (composed of two stages, each containing 33 cm of filtering material) and an irrigation system, and 3) the disposal of the exit gases and liquids.
Figure 1. Methane biofiltration set-up.
It must be mentioned that the periodic bed irrigation (1 L/day for each biofilter) was performed using a nitrogen mineral salt solution, as described elsewhere (Delhoménie et al., 2007), containing 0.5 g/L of nitrogen. The filter bed used in the present study consists of an inorganic material having particles of around 5 cm mean diameter. The following experimental results are expressed in terms of the inlet load, elimination capacity, carbon dioxide production and conversion, as described in Table 1.
CONTROL OF METHANE EMISSIONS ISSUING FROM LANDFILLS
197
Table 1. Parameters used to express the results. Description
Calculation
Inlet load (g/m3.h)
Elimination capacity (g/m3.h) Carbon dioxide production (g/m3.h)
Conversion (%)
with: C: Concentration in g/m3
Q: Gas flow rate in m3/h
V: Bed volume in m3.
3 RESULTS Figure 2 presents the conversion (%) and inlet load (g/m3.h) in the inorganic-based biofilter, as a function of time (days). On day 0, the biofilter was inoculated with the lixiviate, being taken from another biofilter already treating methane (under similar operating conditions). At day 160, the biofiltration processing was stopped for a 2week break period, during which the methane feed and biofilter irrigation ceased. However, the air feed was maintained for this period. During this study, the inlet load remained at around 120 g/m3.h. It was noted that, after the 6 months of operation, the biofilter was still as efficient as it was at the start-up. After the imposed 2-week interruption period, the biofilter restart procedure (this time without inoculation) was achieved within a week, the biofilter reaching its maximum conversion rate soon after. In addition, no difference was detected in the performance of the biofilter, either before or after the imposed interruption (conversion = 22 %). During the whole of the experimental period, the 2 stages of the biofilter exhibited similar methane conversion. Indeed, around 48 % and 52 % of the methane elimination were achieved in stage 1 (lower) and stage 2 (upper) respectively. Figure 3 presents data on the elimination capacity (g/m3.h) for, and the carbon dioxide production (g/m3.h) within, the biofilter, as a function of time (days). During the first six months of operation, the average elimination capacity was estimated to be ~30 g/m3.h, while the average carbon dioxide production was ~ 60 g/m3.h. However, after the planned interruption period, the average carbon dioxide production rate was
198
JOSIANE NIKIEMA AND MICHÈLE HEITZ
Figure 2. Conversion (%) and inlet load (g/m3.h) in the biofilter as a function of time (days).
Figure 3. Elimination capacity (g/m3.h) and carbon dioxide production (g/m3.h) in the biofilter, as a function of time (days).
decreased, by ~15 %, indicating that the biomass growth rate was now at a higher value than that previously observed (Nikiema et al., 2005). As reported for other methane treating biofilters, a good correlation can be observed between the methane elimination capacity and the carbon dioxide production, in the biofilter, except for the biofilter initial start-up, from day 0 to day 35. A possible explanation could be the multiplication rate of the microorganisms; it is at its highest
CONTROL OF METHANE EMISSIONS ISSUING FROM LANDFILLS
199
level at the start-up and then decreases with time until it becomes quite constant, when an equilibrium level is reached (consequently, CO2 production keeps increasing until it reaches its maximum value, this being the phenomenon observed in the present biofilter). Another possible reason is the adsorption of methane on the biofilter’s packing material. However, this possibility was later excluded because such phenomena would be reduced in intensity when the filter material is of an inorganic origin (Devinny and Ramesh, 2005). It is to be mentioned that higher methane conversions (compared to the average value of 22 % reported for this study) can be obtained, e.g. when higher nitrogen concentration is used (data not shown).
4 CONCLUSIONS Wastes result principally from the expansion of the needs of humans living and their activities on the Earth. Thus, for many years in Canada, the total of human generated wastes has never stopped increasing. Along with this increase are created the problems associated with the environment maintenance and safety during the disposal of these wastes. In the case of methane emissions control, from older and/or smaller landfills, biofiltration could be a technically reliable solution. During the lab-scale experiments some of the parameters involved were identified as being important, some of them already having been optimized; e.g. the concentration of the input nitrogen. Therefore, the particular aim of this present study has been to confirm the stability of the biofiltration system undergoing development, i.e. its capacity to give the same performance, continuously, even after several operating months including a substantial process interruption period. Following the experimental operating period (6 months), the biofilter was observed to be still as efficient as it was at the test period commencement. In addition, the restarting of the biofilter, following the 2-week interruption period was readily performed. Indeed, the biofilter reached its maximum conversion performance after only one week. There was essentially no difference between the filter performance, before and after this deliberate interruption, except for the carbon dioxide production level which was at a slightly lower level in the post interruption operating period. Finally, as reported for other methane treating biofilter systems, it was observed that there was a good correlation between the methane elimination capacity and the carbon dioxide production.
5 ACKNOWLEDGEMENTS The authors are indebted to the Natural Science and Engineering Research Council of Canada (NSERC) for their financial support to this project (Discovery Grant). In
200
JOSIANE NIKIEMA AND MICHÈLE HEITZ
particular, co-author J. Nikiema would also like to thank NSERC for providing the supporting Scholarship for her doctoral studies (Canada Graduate Scholarships Program). The authors also express their gratitude to Dr. Peter Lanigan for the text revision.
REFERENCES Buchanan, B., Dewis, G., Lavergne, M., Mitchell, B. and Trépanier, H. (2007) Enquête sur l’industrie de la gestion et des administrations publiques des déchets - secteurs des entreprises: 2004. In (Statistique Canada, Division des comptes et de la statistique de l’environnement), Système de comptabilité nationale, Ottawa, Ontario, pp. 48. Cameron, M., Elliott, A., Marshall, J. and Wang, J. (2005) L’activité humaine et l’environnement, Statistiques annuelles 2005; Article de fond: Les déchets solides au Canada. In (Statistique Canada, Division des comptes et de la statistique de l’environnement), Système de comptabilité nationale, Ottawa, Ontario, pp. 110. Delhoménie, M.-C. and Heitz, M. (2005) Biofiltration of air: A review. Crit. Rev. Biotechnol. 25(12): 53-72. Delhoménie, M.-C., Nikiema, J., Bibeau, L. and Heitz, M. (2007) A new method to determine the microbial kinetic parameters in biological air filters. Chem. Eng. Sci. submitted in 2007. Devinny, J.S. and Ramesh, J. (2005) A phenomenological review of biofilter models. Chem. Eng. J. 113(2-3): 187-196. Nikiema, J., Bibeau, L., Lavoie, J., Brzezinski, R., Vigneux, J. and Heitz, M. (2005) Biofiltration of methane: an experimental study. Chem. Eng. J. 113 (2-3): 111-117. Nikiema, J., Brzezinski, R. and Heitz, M. (2007) Elimination of methane generated from landfills by biofiltration: a review. Rev. Environ. Sci. Bio/Technol. DOI: 10.1007/s11157-006-9114-z. Sanphoti, N., Towprayoon, S., Chaiprasert, P. and Nopharatana, A. (2006) The effects of leachate recirculation, with supplemental water addition, on methane production and waste decomposition in a simulated tropical landfill. J. Env. Manage. 81(1): 27-35. Tränkler, J., Visvanathan, C., Kuruparan, P. and Tubtimthai, O. (2005) Influence of tropical seasonal variations on landfill leachate characteristics–Results from lysimeter studies. Waste Manage. 25(10): 1013-1020. Trebouet, D., Schlumpf, J.P., Jaouen, P. and Quemeneur, F. (2001) Stabilized landfill leachate treatment by combined physicochemical–nanofiltration processes. Water Res. 35(12): 2935-2942. Visvanathan, C., Choudhary, M.K., Montalbo, M.T. and Jegatheesan, V. (2007) Landfill leachate treatment using thermophilic membrane bioreactor. Desalination 204 (1-3): 8-16.
201
Desulfurisation of biogas by biofiltration DIANA RAMÍREZ-SÁENZ AND E. INÉS GARCÍA PEÑA Bioprocesses Department, Unidad Profesional Interdisciplinaria de Biotecnología, Instituto Politécnico Nacional. Av. Acueducto s/n, 07340, Mexico D.F.
ABSTRACT A biofiltration system to eliminate volatile fatty acids (VFAs) and hydrogen sulfide (H2S), using a microbial consortia, was established. The characterization of lava rock, vermiculite and glass rings, as potential packing materials, was previously performed. Vermiculite showed higher VFAs degradation rates than the ones obtained using lava rock; however, some compaction was noticed when this material was used as support. In the lava rock biofilter (1.7 L packed volume) acetic (AA) and propionic (PA) acids were completely removed from the gas and liquid phase, reaching 100% of removal efficiency (RE). Maximum elimination capacities (ECmax) for AA and PA in gaseous phase were 15.2 and 24.5 g/m3bioifilterh, respectively; while in the liquid phase an ECmax of 76.3 and 122.5 g/m3bioifilterh were attained. Empty bed resident times (EBRT) of 85 and 31 s were assessed for increasing H2S inlet loads (36 - 396 g/m3h) into the biofilter. At 85 s, an ECmax of 142 g/m3bioifilterh was determined, with RE of around 99% for all the evaluated inlet loads. At 31 s, a maximum EC of 232 g/m3bioifilterh and a RE of 95 % were found. Complete removal of VFA and H2S from the biogas was obtained in the lava rock biofilter.
1 INTRODUCTION Recent data showed that in 2005 around 12,500 ton/day of waste were produced in Mexico City (INEGI, 2005), the limited landfill capacity and the demand for sustainability technologies for the reduction and treatment of the solid waste are becoming increasingly necessary. Anaerobic digestion systems (ADS) are an effective technology for the reduction of the organic matter and the simultaneous production of energy when they are efficiently applied for treating municipal solid waste and sewage of water treatment plants. The process is environmentally friendly and cost effective because the heat required is generated by bacterial action, few additions of chemicals
201
202
DIANA RAMÍREZ-SÁENZ
AND
INÉS GARCÍA PEÑA
are needed, and the final product can be readily applied. The main products of the process are carbon dioxide (CO2) and methane (CH4), but minor quantities of nitrogen, hydrogen, ammonia, VFAs and H2S (1000- 3000 ppm) are also generated (Angelidaki et al., 2002). CH4 can be utilized to generate different forms of energy (heat and electricity) or be processed for automotive fuel, but VFAs and H2S are toxic and odorous. Many different types of control for these compounds have been successfully used, including chemical scrubbers, activated carbon, and biofilters. Some studies have demonstrated that biofilters can readily handle the odorous and toxic air coming from ADS in a cost-effective and low-maintenance manner (Pride, 2002). Biofiltration is one of the most promising clean technologies for reducing emissions of pollutants into the atmosphere (van Groenestijn and Hesselink, 1993). This technology based on microbial degradation of compounds from a gas stream is considered an attractive alternative when compared to chemical and physical treatments. It is economic and environmental friendly since generates less residues that others technologies, due to total biological oxidation of the pollutants. Many studies have been conducted to study the design and operational parameters, as well as the microbial process involved in biofiltration systems, showing that the system effectively controls and removes odors in diluted gas streams contaminated basically with sulphur compounds (Yang and Allen, 1994; Smer et al., 1998; Ergas et al., 1995; Morton and Caballero 1998). In the biofiltration process the type of packing materials is essential for a proper performance. The support material acts not only as a surface for microbial growth, it also provides the water required to promote the metabolic activity and in some cases can also act as buffer in shielding microorganisms from inhibitory substances while adsorbing high initial concentration of substrate and progressively releasing it for microbial degradation of some sulphur compounds (Ng et al., 2004). In some literature reports, various materials have been used as the support media for microbial growth and significant differences have been reported in the performance of biofilters for H2S removal packed with different materials. Van Langenhove et al. (1986) used wark woods, Hirai et al. (1990) used peat, Yang and Allen (1994) used compost, Chung et al. (1996) used calcium alginate pellets, Morton and Caballero (1998) used lava rock and Wani et al. (1998) used various mixtures of compost-perlite hog fuel. Cedar Rapids reports and discusses the issues of the selection of the packing materials and provides insight on some potential beneficial properties of lava rock (Martin et al., 2002). H2S elimination, as the main odor compound, has been extensively studied in biofiltration systems (Cho et al., 2000; Oyarzún et al., 2003; Ng et al., 2004; Duan et al., 2006). Less is known about the aerobic degradation of VFAs, the anaerobic elimination of VFAs from waste water has been proved in biofilters and biotrickling filters. Yun and Ohta (1997) reported the characteristics of the microorganisms capable of assimilating VFAs, as their sole source of carbon and energy, and more recently
DESULFURISATION OF BIOGAS BY BIOFILTRATION
203
describe the feasibility of the removal of VFAs by an immobilized strain of Rhodococcus sp (Yun and Ohta, 2005). The main goal of this work was to couple a biofilter to an ADS while treating vegetable waste in order to eliminate the malodors pollution and the H2S produced during the anaerobic process. Potential packing materials for the biofiltration systems were characterized and the biofilter performance to degrade VFAs as well as high concentrations of H2S was determined. Different substrate loads and empty bed resident times were also evaluated to characterize the biofiltration system.
2 MATERIALS AND METHODS 2.1 MICROBIAL CULTURE AND PACKING MATERIAL A microbial consortium was cultivated in 2 L Erlenmeyer flask with 1 L of mineral media and enriched with VFAs by injecting a gas stream saturated with these acids. The same consortia were adapted to sulfur compounds by adition of sodium thiosulphate (Na2S2O3) in the mineral media. The mineral media contains (g/L): (NH4)2SO4, 3; KH2PO4, 0.6; K2HPO4, 2.4; MgSO4·7H2O, 1.5; CaSO4, 0.15; FeSO4, 0.03. The supports evaluated as potential packing materials for the biofiltration systems were: vermiculte, glass rings and lava rock. Characterization of these packing materials was performed by using standard methods. The three different support materials were packed in the biofilter. In situ immobilization was facilitated by recirculation of the microbial culture previously adapted to VFAs and Na2S2O3 using a peristaltic pump, at a flux of 0.11 L/min. 2.2 MICROCOSM 4 g of the packed material with the immobilized biomass were introduced in serum bottles of 125 ml of total volume. Initial concentrations of 40 mg/L of acetic, propionic, butyric and valeric acids were injected separately to determine its consumption. Headspace samples were periodically taken to evaluate the CO2 production due to the VFAs assimilation. 2.3 EXPERIMENTAL SETUP 2.3.1 VFA CONSUMPTION Lava rock biofiltration system (Figure 1a) consists in a glass column of 0.94 m length and 5.5 cm of internal diameter (1.7 L of packed volume). An empty bed residence time EBRT of 120 s was used for VFAs elimination tests. The gas stream was humidified and fed at the top the biofilter using a mass flow controller. Different loads of the VFAs were introduced into the biofilter to evaluate its performance (from
204
DIANA RAMÍREZ-SÁENZ
AND
INÉS GARCÍA PEÑA
6.1 to 24.5 g/m3bioifilterh). Sample ports were located in the output and input of the gas stream. Gas samples were taken directly using gas tight syringes.
a.
b.
Figure 1. Experimental systems packed whit lava rock: a. Configuration used for VFAs elimination test, b. biofiltration system coupled to the ADS used to evaluate H2S consumption.
DESULFURISATION OF BIOGAS BY BIOFILTRATION
205
2.3.2 H2S CONSUMPTION H2S removals were determined in the biofiltration system coupled to the ADS. The scheme is shown in Figure 1b. The biofilter (1.7 L volume) was packed whit lava rock. Different amounts of humidified air were mixed with the biogas stream and fed into the biofilter to obtain increasing H2S inlet loads, from 36 to 144 gH2S/m3h for an EBRT of 85 s and from 99 to 396 gH2S/m3h when an EBRT of 31 s was established. 2.4 ANALYTICAL METHODS CO 2 productions were measured by injecting headspace samples of the microcosms in a gas chromatograph (GowMac) equipped with a thermal conductivity detector. A CTR1 packed column was used for the analysis. Temperatures in the column, injector and detector were 30, 75 and 120ºC, respectively. Helium was used as the carrier gas at a flow of 65 mL/min. VFA in the biofilter samples were measured using a gas chromatograph (Buck Scientific) equipped with a flame ionization detector and a 4 - 6.9 feet packed column (Hayesep R 80/100 Mesh, Chromatography Research Supplies, Inc, Louisville, USA). The column temperature was 190ºC and the injector/ detector temperature was 200ºC. Nitrogen was used as the carrier gas at a flow of 30 mL/min. Data integration was accomplished by Peak Sample software. Inlet and outlet H2S concentrations were determined using a H2S analyzer Jerome 631-X (Arizona Instruments LLC, USA).
3 RESULTS AND DISCUSSION 3.1 PACKING MATERIAL CHARACTERIZATION Density, water retention capacities (WRC), pH and particle size distribution of the different packing materials were determinated. Moisture is one of the most important parameters for the development of the biofilm and the microbial activity. The WRC of the packing materials were different, vermiculite exhibited the higher WRC (65%) compared with lava rock and glass rings. Vermiculite has been reported as a good packing material; its water retention capacity was high compared to those obtained with the other materials, however, when mineral medium was circulated in the reactor to favor the biofilm establishment, some disintegration of the particles was noticed, reducing the original volume and provoking some compaction. The glass rings showed the lower value of water retention; however this material was evaluated because its good mechanic characteristics that could allow water addition by irrigation in order to maintain the moisture in the biofilter. Lava rock showed a WRC of 15%, which was lower to the one reported by Cho et al. (2000). These authors evaluated the physical characteristics of lava rock, used as support for malodorous gases (H2S, methandiol, dimethyl sulfide and ammonia) removal, WRC between 25 and 47% were determined in this study.
206
DIANA RAMÍREZ-SÁENZ
AND
INÉS GARCÍA PEÑA
Similar initial pH was obtained for vermiculite and for glass rings, 6.7 and 6.9, respectively, while a pH of 8.1 was measured for lava rock. The pH in the biofiltration systems is an important factor for its application, lower pH inhibits the activity of the deodorizing microorganisms, it was reported that the deodorization efficiency significantly decreases at lower pH values. The VFA and other intermediate products could provoke a drop of the pH in the biofilter. Thus, buffering capacity of a carrier, to resist pH change, is important to maintain biological activity for a long term biofilter operation (Cho et al., 2000). Yun and Ohta (2005) reported the necessity of controlling de initial pH between 8 and 9 for the effective removal of VFA by Rhodococcus sp. B216. The results of the physic properties (WRC and initial pH) showed that vermiculite and lava rock were of potential interest to be used as packing materials for eliminating VFAs. However, for effective removal of the pollutant’s odors within the biofiltration system is essential to evaluate the microorganisms adhesion and the microbial activity in these supports. 3.2 VFA CONSUMPTION EXPERIMENTS IN MICROCOSMS Once the different packing materials were physically characterized, they were packed in the biofilter divided into two separate sections each one packed with one of the supports and divided by using a Teflon mesh. The biofilm in the packed material was obtained by a periodic circulation of mineral medium containing the microbial community, previously adapted with acetic acid for one month. Once a visible growing was evident, after approximately 2 weeks, initial VFAs concentrations of approximately 40 mg/L were injected, as only carbon and energy sources in closed systems. The evolution of the CO2 production was periodically determined, as a measure of the substrate assimilation. Data obtained with the microbial community developed in vermiculite are presented in Figure 2a, the degradation started after the first hours of culture, showing a short adaptation period. The substrates were oxidized in different extends as it is presented in Table 1, acetic, butyric and valeric acids were almost completely mineralized to CO2, obtaining 84, 85 and 88 % of conversion, respectively, in approximately 40 hours of cultivation. Meanwhile, less extent of transformation was determined for propionic acid. Using the theoretical estoichiometric equation a degradation rate could be calculated, the degradation rates are summarized in Table 1. A higher degradation rate was attained with acetic acid compared to those rates obtained with the others VFAs evaluated. A similar analysis was performed using the microbial community attached to lava rock, the CO2 productions are presented in Figure 3b, VFA degradation started during the 10 hours of cultivation, acetic, propionic and valeric acids were 79, 68 and 77% mineralized to CO2 (Table 1), butyric acid was only partially degraded obtaining for it a 55% of mineralization.
DESULFURISATION OF BIOGAS BY BIOFILTRATION
Figure 2. C02 produced by the assimilation of VFA; acetic acid, propionic acid, butyric acid, valeric acid, respectively in microcosm’s. Data obtained from the biofilm developed in vermiculite (a) and lava rock (b).
207
208
DIANA RAMÍREZ-SÁENZ
AND
INÉS GARCÍA PEÑA
Higher amount of biomass was established in vermiculite (8.5 x 105 UFC/g) compared to the biomass determined for lava rock (6.5 x 104 UFC/g). This was a possible explanation for the better activity (higher VFAs elimination rates) determined in the vermiculite samples. This fact was also correlated with the WRC, vermiculite showed higher WRC (65%) compared to the one of lava rock WRC (15%), which allowed both, better growth and metabolic activity. Table 1. Removal efficiency and degradation rates obtained with the microbial consortium developed in three packing materials evaluated with four VFA. Acid
Packing Material Glass Rings
Acetic Propionic Butyric Valeric
Vermiculite
Lava rock
Removal Efficiency (%)
Removal Rates (mg/Lh)
Removal Efficiency (%)
Removal Rates (mg/Lh)
Removal Efficiency (%)
Removal Rates (mg/Lh)
-
-
84 64 85 88
1.46 0.94 0.42 0.71
79 68 55 77
0.41 0.37 0.16 0.24
Less biofilm development was obtained using the glass rings, no biological activity with the VFA as substrate was detected. These results could be due to the low water retention capacity of this support, probably the lower water content inhibit the metabolic activity of the microbial community. Yun and Ohta (2005) reported the elimination of high concentrations of acetic, propionic, butyric and valeric acids from a waste-food solution by immobilized cells of Rhodococcus sp. B261, the consumption of the VFA was initiated after 48 hours, then acetic and propionic acids were removed in 64 hours. Butyric and valeryc acids were depleted in 72 hours. During the present study, acetic, propionic and valeric acids were degraded in low concentration by the microbial consortium fixed in vermiculite and lava rock, while the butyric acid was partially removed. Higher degradation rates were obtained with the microbial biofilm developed in vermiculite. However, under biofiltration conditions, in the section packed with vermiculite some compaction and fungal growth was noticed due to the drought of the support and the fast decrease of the moisture. 3.3 VFA CONSUMPTION UNDER BIOFILTRATION CONDITIONS Based on the preliminary results, lava rock was used as packing material in the biofilter. Since the main goal of the present study was to eliminate the malodors
DESULFURISATION OF BIOGAS BY BIOFILTRATION
209
Figure 3. Relationship between inlet load and elimination capacity of lava rock biofilters to acetic acid (a) and propionic acid (b).
210
DIANA RAMÍREZ-SÁENZ
AND
INÉS GARCÍA PEÑA
pollutants emitted in the ADS by using a biofiltration system, and considering that during the initial phase of the anaerobic digestion process high amounts of VFA (mainly acetic and propionic acids) was produced. Some experiments were performed to evaluate the VFA elimination under biofiltration conditions. Increasing inlet loads of acetic and propionic acid were fed into the biofilter. In Figure 3a and 3b the elimination capacities (ECs) vs the inlet loads obtained with acetic and propionic acid, respectively, are presented. All the inlet concentrations of the pollutant evaluated were 100% removed. Considering that VFA are highly soluble in the liquid phase, and the relation of water/gas in the biofilter was around 5 (288g of water, for an initial moisture of 55%, and 1.12L of air), an EC in the liquid phase could be calculated and it is shown in Figure 3. The EC obtained for acetic and propionic acids are in the range of the EC reported for the elimination of other compounds in biofiltration systems. The system reached an EC of around 120 g/m3h of the propionic acid with 100% of removal efficiency. Higher inlet loads were not evaluated to avoid saturation and accumulation of the VFA in the liquid phase. An important decrease in the pH, to values around 4.3, was determined in the biofiltration when the experiments were performed, enhanced acetic and propionic acids removals could be expected by controlling and adjusting the pH, which increases the immobilized cells number. Yun and Ohta (2005) demonstrated that higher valeric acid removal rate was obtained when the pH was between 8 and 9 and more cells were immobilized and developed in the ceramic support, attaining 1.4 x 109CFU/g-ceramic beads. 3.4 H2S ELIMINATION IN THE BIOFILTRATION SYSTEM The evolution of the elimination capacity with different H2S inlet loads is shown in Figure 4. At EBRT of 85 sec. (Figure 4a) increasing H2S inlet loads of 50, 80, 120 and 150 g/m3h were 99% removed by the biofiltration system, reaching a maximum EC of 148 g/m3h. When the EBRT was reduced to 31 sec. (Figure 4b), H2S inlet loads of 100 and 200 g/m3h were complete degraded (100% removal efficiency), an increment in the inlet load to 300 g/m3h reduced the removal efficiency in the system to 90%, inlet load of 400 g/m3h inhibited the biological activity and the removal efficiency dropped to 50%, this load corresponds to a very high concentration of pollutant, as high as 1500 ppm of H2S. A maximum EC of 232 g/m3h was reached by the biofiltration system, which is three times higher that those reported by Cho et al. (1991) and Ootani et al. (1991), using a Thiobacillus sp. HA43 and an activated sludge, respectively. Similar to the ones obtained by Cho et al. (2000), these authors reported EC of around 342 and 428 g/m3h (Table 2).
DESULFURISATION OF BIOGAS BY BIOFILTRATION
211
Figure 4. Relationship between inlet load and elimination capacity of lava rock biofilters to H2S at an EBRT of 85 s (a) and 31 s (b).
212
DIANA RAMÍREZ-SÁENZ
AND
INÉS GARCÍA PEÑA
Table 2. Reference Wani et al., 1998 Cho et al., 2000 Park et al., 2001 Oyarzún et al., 2003 Ng et al., 2004 Duan et al., 2006 This work
ECMAX (g/m3h) 120 392 7 55 52 181 232
4 CONCLUSIONS The microbial consortium, previously adapted to VFA and Na2S2O3, showed higher biological activity for the degradation of propionic, butyric and valeric acids when it was immobilized in vermiculite, compared to lava rock and glass rings. No biological activity was detected in glass rings due to its low water retention capacity. However, humidity losses which in turn allowed fungal growth in vermiculite provoked compaction of this support under biofiltration condition. Analyses conducted in the lava rock biofilter demonstrated that the acetic and propionic acids were completely removed from the gas and liquid phase at increasing concentration, reaching 100% of removal efficiency in all the evaluated concentrations. Tests with the ADS stream showed that all the components of the gas phase are almost completely removed in the biofilter. The results strongly suggest the feasibility of the biofiltration system to degrade and eliminate VFA and H2S emitted from the ADS. This will allow both, to eliminate the odors problems and the removal of the H2S in order to use the biogas as an alternative source of energy.
5 ACKNOWLEDGEMENTS Research and Diana Ramirez-Saenz´s fellowship were partially supported by the Instituto Politecnico Nacional, grant SIP20071100. Authors want to tank the assistantship and collaboration of Ing. Federico Muñoz.
DESULFURISATION OF BIOGAS BY BIOFILTRATION
213
REFERENCES Cho, K.S, Ryu, H.W. and Lee, N.Y. (2000) Biological deodorization of hydrogen sulfide using porous lava as carrier of Thiobacillus thioxidans. J. Biosc. Bioeng. 1: 25-31. Chung, Y.C., Huang, C. and Tseng, C.P. (1996) Operation and optimization of Thiobacillus thioparus CH11. Biofilter for hydrogen sulphide removal. J. Biotechnol. 52: 31-39. Ergas, S.J., Schroeder, E.D., Chang, D.P.Y. and Morton, R.L. (1995) Control of volatile organic compounds emissions using a compost biofilter. Water Environ. Res. 67: 816-821. Hirai, M., Ohtake, M, and Shoda, M. (1990) Removal kinetics of hydrogen sulfide, methanethiol and dimethyl sulfide by peat biofilters. J. Ferment. Bioeng. 70: 334-359. Instituto Nacional de Estadística, Geografía e Informática (INEGI). http://www.inegi.gob.mx/est/ default.aspx?c=5911 (pub. 2005; cons. june, 2007). Martin, R.W., Li, H., Mihelcic, J.R., Crittenden, J.C., Lueking, D.R., Hatch, C.R. and Ball, P. (2002) Optimization of biofiltration for odor control: model calibration, validation, and applications. Water Environ. Res. 74(1): 17-27. Morton, R.L. and Caballero, R.C. (1998) Using full scale biotrickling for the removal of hydrogen sulfide and odor from waste water facilities air streams. Proceedings of the USC-TRG 1998 Conference on Biofiltration. Los Angeles, California, 107-114. Ng, Y.L., Ran, X.G., Chen, A.L., Gen, W.D., Gould, W.D., Liang, D.T. and Koe, L.C.C. (2004) Use of activated carbon as a support medium for H2S biofiltration and effect of bacterial immobilization on available pore surface. Appl. Microbiol. Biotechnol. 66: 259-265. Pride, C. (2002) ATADs, Odors, and Biofilters Florida Water Res J. 18-26. Van Groenestijn, J.W. and Hesselink, P. (1993) Biotechniques for air pollution control. Biodegradation. 4: 283-301. Wani, A.H., Branion, R.M.R. and Lua, A.K. (1998) Efects of periods of starvation and fluctuating of hydrogen sulfide concentration on biofilters dynamics and performance. J. Hazard. Mat. 60: 287-296. Yang, Y. and Allen, E.R. (1994) Pollution control of hydrogen sulfide 1. Desing and operation parameters. J. Air Waste Manage. Assoc. 4: 863-869. Yang, Y. and Allen, E.R. (1994) Biofiltration control of hydrogen sulphide. 2. Kinetics, biofilter performance, and maintenance. J. Air Waste Manage. Assoc. 44: 1315-1321. Yun, S.I. and Ohta, Y. (2005) Removal of volatile fatty acids with inmovilized Rhodococcus sp. B261. Biores. Technol. 96: 41-46.
This page intentionally left blank
215
An innovative biotrickling filter for H2S removal from biogas LAURA BAILÓN ALLEGUE Innovative Energy Systems, Profactor Produktionsforschungs GmbH, Im Stadtgut A2, A-4407 Steyr/Gleink, Austria
ABSTRACT A novel biotrickling filter system was developed to remove H2S from biogas. The aim was to remove 2000 ppmv of H2S to less than 3 ppmv in order to use biogas in combination with fuel cells, as vehicle fuel or inject it in the natural gas grid. It was found that for H2S inlet concentrations up to 1000 ppmv the H2S outlet concentration was < 3 ppm with efficiencies > 99%. The maximum practical elimination capacity was 32.5 g-H2S. m-3filter. h-1. For 2000 ppmv H2S inlet concentrations the outlet concentrations were up to 75 ppmv and the elimination capacity was 55 g-H2S. m-3filter. h-1. The H2S removal efficiency never dropped below 93.5% in all the concentration range tested.
1 INTRODUCTION Biogas, produced by the digestion of organic materials in landfills and sewage treatment plants, is a potentially important renewable energy source. It has a high content of methane and its use is highly encouraged by the need of reducing greenhouse gas emissions. Nowadays it is mainly utilized to obtain electrical and thermal energy in combustion engines, but a more efficient and widespread use can be reach by introducing it in the natural gas grid or using it in fuel cells or as automobile fuel. Hydrogen sulphide, which is always present in the biogas, normally at concentrations between 80 - 4000 ppmv, is one of the most problematic contaminants in order to use digested gas as energy source because is toxic and corrosive to most equipment. Moreover, its combustion leads to sulphur dioxide emissions. In the case of the fuel cells, very low concentration of H2S can damage the FCs’ catalyst. The requirements are also rather strict when injecting upgraded biogas into the natural gas grids or using it as vehicle fuel (Table 1).
215
216
L. BAILÓN ALLEGUE
In general, H2S removal methods can be classified in two big groups, the physical-chemicals, which are the traditional ones and currently still dominate the market, and the biotechnological. In the past two decades increasing attention has been paid to the biotechnological methods because having the same or even more efficiency than the physical-chemical ones are generally cheaper, avoid the use of catalysts, and do not generally originate secondary contaminant streams (Sublette et al., 1988; Gadre, 1989; Deshusses and Cox, 2000). Table 1. H2S requirements for different biogas utilization technologies (Wellinger and Lindberg, 2000; Trogisch et al., 2004). Technology Boilers and stirling engines Kitchen stoves Internal Combustion Engines
H2S tolerance (ppmv) < 1000 < 10 < 1000
Turbines Microturbines Fuel Cells: MCFC Natural Gas Upgrade/Vehicle fuel
< 10000 < 70000 < 10 in fuel <4
Remarks
Otto engines more susceptible than diesel
<0.1-0.5 at the anode Depends on the country
A large variety of aerobic and aerobic bacteria are capable of H2S oxidation and hence serve as potential candidates for gas desulphurisation technology (Syed et al., 2006). Biogas has a very low content of oxygen so, the use of anaerobic bacteria seems a logical option. Several studies with phototrophic bacteria and anaerobic chemotrophic bacteria have been done (Cork et al., 1985; Sublette and Sylvester, 1987; Kim and Chang, 1991). Important disadvantages for the application of photosynthetic bacteria is their requirement of radiant energy and their very slow growth rate (Syed et al., 2006). In the case of facultative anaerobic chemotrophic bacteria, like Thiobacillus denitrificans, there are two main drawbacks. The lower H2S degradation capacity under anaerobic conditions and the fact that air can be supplied to bioreactors more economically than nitrate (Sublette, 1987). Aerobic bacteria like Thiobacilli present several advantages as low nutritional requirements, usually high H2S affinity and slow biomass growth (Cho et al., 1995; Chung et al., 1996; Jin et al., 2005). In this respect, Profactor GmbH has patented a novel biotrickling filter for the removal of H2S from biogas using autotrophic aerobic bacteria. In this article the performance of this system is presented.
AN INNOVATIVE BIOTRICKLING FILTER FOR H2S REMOVAL FROM BIOGAS
217
2 MATERIALS AND METHODS 2.1 BACTERIA AND MEDIUM The biotrickling filter (BTF) was inoculated with a biomass suspension harvested from a BTF that was eliminating H2S for several months. Originally this BTF was inoculated with bacteria isolated from activated sludge taken from a wastewater treatment plant in Asten (Austria). A 16S-rDNA sequencing was performed on a sample from the biofilm of the new BTF after 2 months of operation. The dominant bacterial band obtained by single strand conformation polymorphism was identified as Thiobacillus denitrificans. The identification was commissioned to the Natural Science University of Graz (Austria). The medium used in the BTFs contained per litre of distilled water: 1.5 g of KH2PO4, 8.58 g of K2HPO4.3H2O, 0.1 g of MgCl2.6H2O, 0.055 g of CaCl2.2H2O and 0.8 g of NH4Cl. 2.2 NOVEL BIOTRICKLING FILTER To determinate the performance of our novel biotrickling filter and to find the optimal operational conditions a co-current laboratory system was set-up (Figure 1). The filter was divided into 3 packed beds. Oxygen saturated medium was introduced at the inlet of each bed, it flowed from that bed to the next ones and finally left the system at the bottom of the third bed. The conventional way of supply oxygen into a biofilter, when working with biogas, is injecting directly air into the gas stream. When doing this, a safety system is required to avoid explosive mixtures in case of compressor failure. Moreover, the quality of the biogas decreases when nitrogen is added and when not all the oxygen is consumed. This is a problem if the biogas is used as a vehicle fuel or injected in the natural gas grid because of the strict biogas quality requirements. When working with fuel cells oxygen is also a problem because the anode of the fuel cells is damaged with very low concentrations of this compound. By introducing oxygen through the liquid medium the use of a safety system is avoided and the quality problems minimized. The filter column, of 0.07 m of diameter and 0.6 m of total height, was made of plexiglass. Each of the beds had a height of 0.1 m. A total working volume of 1.15 dm3 was filled with glass Raschig rings. The liquid distribution system consisted of perforated plates on the top and bottom of each bed with 3 and 4.5 mm holes. The bubble column had a diameter of 0.07 m and a length of 0.5 m. It was made of plexiglass and the air was introduced in the liquid through a horizontal perforated pipe on the bottom of the column, obtaining dispersed bubbles. Instead of using biogas, due to safety reasons, a mixture of N2 (65%), CO2 (35%) and H2S (traces) was utilized. The H2S was supplied from a gas cylinder containing 2000 or 4000 ppm of H2S with N2 as a diluted gas. The inlet loads were
218
L. BAILÓN ALLEGUE
Figure 1. Scheme of the biotrickling filter system.
controlled by changing the effluent flow rates of the gas cylinders utilizing Bronkhorst HI-TEC electronic mass flow controllers. Control parameters were the temperature, kept at 30ºC through a thermal bath; the pH, adding NaOH 2N as neutralized agent of the formed sulphate (Titroline Alpha Plus, Schott) and the conductivity by adding or/and removing manually medium. In daily bases 10% of the total recirculated medium was lost by evaporation, mainly at the bubble column. The same amount of fresh medium was added, in order to keep the nutrient concentrations sufficiently high and to prevent sulphate accumulation. When conductivity was near to 30 mS.cm-2 approximately half of the recirculated medium was replaced by fresh one. The H2S concentration on the gas phase was determined at the inlet gas stream, at the gas outlets of the three beds and at the air outlet of the bubble column. Hydrogen sulphide was analysed using a gas chromatograph (Perkin-Elmer Autosystem XL) equipped with a flame photometric detector. The whole system was placed inside of a hood extractor and the outlet gas and air streams were forced to pass through a CuSO4 solution to retain as CuS the possible no degraded H2S. The operational parameters are given in Table 2.
AN INNOVATIVE BIOTRICKLING FILTER FOR H2S REMOVAL FROM BIOGAS
219
Table 2. Standard operating conditions of the biotrickling filter. Operating parameters Inlet H2S concentration (ppmv) Gas flow rate (m³.h-1) Fresh liquid inlet medium rate per bed (m³.h-1) Empty bed retention time per bed (min) Air flow rate at the bubble column (m³. h-1) pH Temperature (°C) Liquid oxygen concentration at the BTF inlets (mg.l-1)
Value 200 – 2000 0.020 7.7.10-3 1.15 0.34 7 29 – 31 ~ 6-6.5
The packed division of the biotrickling filter, as well as the general design, was performed taking into account the results obtained from two previous biotrickling filter prototypes. The first one worked with a single packed bed and it was running in a biogas plant at the University of Nitra, Slovakia, during 18 months. Good results where achieved for relatively low H2S concentrations but for concentrations higher than 450 ppmv the H2S outlet concentrations were much higher than 3 ppmv. It was observed that the oxygen was mainly consumed at the top part of the filter bed. Consequently, the biodegradation in the lower part was reduced, leading to less bacterial efficiency and a not optimal H2S reduction. To try to solve this problem a laboratory biotrickling filter divided in 3 sections was set-up. At this filter equal amounts of oxygen saturated medium was introduced at the beginning of each bed leaving the system at the end of that bed. In this way more flexibility and a higher elimination capacity was expected by optimising the oxygen transference. With this system the oxygen transference limitation was partially solved as for H2S inlet concentrations up to 900 ppmv the outlet values were inferior to 3 ppmv. Nevertheless, H2S was stripped in the air stream leaving the bubble column. This is not affecting the quality of the cleaned gas but could produce odour and healthy problems if the system is in a close environment. The H2S stripping is produced because not all the H2S that is dissolved in the liquid phase is degraded by the bacteria. Part of it remains in the recycled liquid and past to the air when reaching the equilibrium at the bubble column. The largest source of dissolved H2S come from the first and second bed liquid exists. To solve the stripping problem the new configuration here presented, where all the liquid medium is drained at the filter bottom, was tested.
220
L. BAILÓN ALLEGUE
3 RESULTS AND DISCUSSION Increasing H2S concentrations were fed to the filter at the standard operational parameters (Table 2) and the H2S concentration was measured at the different bed outlets. When reaching 2000 ppmv decreasing H2S concentrations were introduced to evaluate if there was any difference in the performance of the system due to the adaptation of the microorganisms to high H2S concentrations (Figure 2).
Figure 2. Relationship between the inlet and outlet H2S.
In standard operational conditions the system was able to remove 1,000 ppmv of H2S to less than 3 ppmv with efficiencies >99%. The maximum practical elimination capacity was 32.5 g-H2S. m-3filter. h-1. For 2000 ppmv H2S inlet concentration, outlet concentrations of 34 to 75 ppmv were found. RE of H2S never dropped below 93.5% in the tested concentration range (Figure 3). At the highest concentration applied, 2,100 ppmv, the elimination capacity was 55 g-H2S. m-3filter. h-1.
AN INNOVATIVE BIOTRICKLING FILTER FOR H2S REMOVAL FROM BIOGAS
221
Figure 3. Relationship between H2S elimination capacity, removal efficiency and H2S inlet loading rate.
There were not significant differences in the performance of the biotrickling filter for increasing or decreasing H2S inlet concentrations, which suggests that the bacteria are easily adaptable to different H2S loads. In this system the amount of H2S stripped was always lower than 1.3 mg.h-1 (3.3 ppmv) (Figure 4). The new liquid configuration was a good solution to the stripping problem. Moreover the new liquid and therefore oxygen distribution did not have negative effect in the performance of the system as the EC and RE were slightly better than in the previous one. Referring to the performance of the different beds, the contribution of each of them to the overall removal efficiency is presented in Figure 5. The removal efficiency at each bed has been calculated as the ratio of the difference between the inlet and outlet concentrations of H2S in every bed to the concentration at the entrance of the filter. Around 70% of the H2S is removed in the 1 bed, 25% in the 2 bed and 5% in the last one. For the highest loadings the percentage of H2S removed on the 1 bed decreases and on the 2 and 3 increases but very slightly. It can be possible that in the 1 bed the H2S were not totally degraded by the bacteria but just dissolved and degraded afterwards in the other beds. However, when working with just 2 beds the contribution of each of them is quite similar than when working
222
L. BAILÓN ALLEGUE
Figure 4. H2S stripped through the bubble column at different H2S inlets in the old (x) and new ( ) BTF configuration.
Figure 5. Removal efficiency profiles in the different beds.
AN INNOVATIVE BIOTRICKLING FILTER FOR H2S REMOVAL FROM BIOGAS
223
with 3 (results not shown) and with almost no H2S stripping, which means that the amount of H2S dissolved on the liquid phase is very low. According to the obtained results it seems that the BTF should have more capacity for removing H2S. The fact that for H2S inlet concentrations of 1000 ppmv more than 3 ppmv are detected in the outlet could be due to oxygen limitations or to H2S mass transfer limitations. To check this air was directly introduced in the system through the gas stream (flow rate of 50% vol. air of the gas stream). In these conditions for an inlet 2000 ppmv H2S the outlet was lower than ~3 ppmv. This indicates that the system is oxygen limited. The presence of oxygen in the gas phase specially benefits the areas in the biofilm with low water content because of the small diffusion barrier and the lower availability of oxygen through the liquid phase. This means a higher active biofilm layer. The large availability of oxygen in the liquid phase implies as well a higher microbiological activity. Oxygen concentration in the liquid phase highly influences the bacterial degradation rate (Jaworska and Urbanek, 1998). There are two basic possibilities of increasing the amount of oxygen introduced in our BTF. The simplest would be to raise the recycled liquid flow rate. In this way the oxygen liquid concentration remains constant. The other would be to use in the bubble column, instead of air, pure oxygen or a mixture of air and oxygen. Doing this higher oxygen saturation concentrations would be reached at the liquid. The effect of an increase in the liquid flow was studied by doubling it respect to standard operation conditions. After an adaptation period of 11 days with double liquid flow rate, 46.2 l.h-1, the RE of the system increased. For example, for an inlet H2S concentration of 2000 ppmv the H2S outlet was 16 ppmv, the RE 99.23 % and the stripped H2S 1 ppmv. The improvement in the performance of the system with double liquid flow rate could be due apart from a more availability of oxygen to a better liquid distribution. It is likely that the biotrickling filter was limited by liquid channelling and partial wetting of the carrier material. Other studies with a higher oxygen saturation concentration should be done.
REFERENCES Cork, D., Mather, J., Maka, A. and Srnak. A. (1985) Control of oxidative sulfur metabolism of Chlorobium limicola forma thiosulfatophilum. Appl. Environ. Microbiol. 49: 269-272. Cho, K.-S., Sublette, K.L. and Raterman, K. (1995) Oxidation of hydrogen sulfide by an enrichment from sour water coproduced with petroleum. Appl. Biochem. Biotechnol. 51/52: 761-770. Chung, Y.-C., Huang, C. and Tseng, C.-P. (1996) Biodegradation of hydrogen sulfide by a laboratoryscale immobilized Pseudomonas putida CH11 biofilter. Biotechnol. Prog. 12(6): 773-778.
224
L. BAILÓN ALLEGUE
Deshusses, M.A. and Cox, H.H.J. (2000) Biotrickling filters for air pollution Control. The Encyclopedia of Environmental Microbiology (Book). Gadre, R.V. (1989) Removal of hydrogen sulfide from biogas by chemoautotrophic fixed-film bioreactor. Biotechnol. Bioeng. 34: 410-414. Jaworska, M. and Urbanek, A. (1998) The influence of oxygen concentration in liquid medium on elemental sulphur oxidation by Thiobacillus Thiooxidans. Bioproc. Eng 18: 201-205. Jin, Y., Veiga, M.C. and Kennes, C. (2005) Effects of pH, CO2, and flow pattern on the autotrophic degradation of hydrogen sulfide in a biotrickling filter. Biotechnol. Bioeng. 92: 462-471. Kim, B.W. and Chang, H.N. (1991) Removal of hydrogen sulfide by Chlorobium thiosulfatophilum in immobilized cell and sulfur settling free-cell recycle reactors. Biotechnol. Prog. 7: 495-500. Sublette, K.L. (1988) Microbiological desulfurization of gases. Patent US 4.760.027. Sublette, K.L. (1987) Aerobic oxidation of hydrogen sulfide by Thiobacillus denitrificans. Biotechnol. Bioeng. 29: 690-695. Sublette, K.L. and Sylvester, N.D. (1987) Oxidation of hydrogen sulfide by Thiobacillus denitrificans: Desulfurization of natural gas. Biotechnol. Bioeng. 29: 249-257. Syed, M., Soreanu, G., Falleta, P. and Béland, M. (2006) Removal of hydrogen sulfide from gas streams using biological processes – A review. Canadian Biosyst. Eng. 48: 2.1 - 2.14. Trogisch, S., Baaske, W.E., Accettola, F., Bailón, L., Benito, M., Berger, P. et al. (2004) Biogas Powered Fuel Cells. Trauner Verlag, Linz. Wellinger, A. and Lindberg, A. (2000) Biogas upgrading and utilisation. IEA Bioenergy, Task 24:. Energy from biological conversion of organic waste.
225
REMOVAL OF INORGANIC POLLUTANTS
This page intentionally left blank
227
Removal of ammonia by immobilized Nitrosomonas europaea in a biotrickling filter packed with polyurethane foam MARTÍN RAMÍREZ, JOSÉ MANUEL GÓMEZ AND DOMINGO CANTERO Department of Chemical Engineering, Food Technology and Environmental Technologies, Faculty of Sciences, University of Cádiz, 11510 Puerto Real (Cádiz), Spain
ABSTRACT A chemolithoautrotrophic microorganism Nitrosomonas europaea has been utilized to remove gaseous ammonia in a biotrickling filter packed with polyurethane foam. The optimal pH for removing was 7.5 and the biological removal efficiency was zero at pH 6.5. Empty bed residence time of 150, 100, 50, 25, 20, 11 and 5 seconds were tested; the removal efficiency was of 100% in all range for a constant load of 8 gN.m-3.h-1. The critical elimination capacity was 270 gN.m-3.h-1 while the nitrite concentration was below of 100 mM (EBRT of 11 second, pH 7.5-7.6). Therefore, these results demonstrate that is possible to reach a high removal of ammonia using polyurethane foam, as solid support for Nitrosomonas europaea, and a biotrickling filter system.
1 INTRODUCTION Ammonia (NH3) is a colourless air pollutant with a strong and irritating odour. Breathing levels about 100 ppmv ammonia in air, noticeable irritation of eyes and nasal passages after few minutes exposure (Carson and Mumford, 2002). Ammonia is released into the atmosphere from various sources, such as sludge and wastewater treatment plants, composting plants, livestock farms, wastewater treatment plants (Chung et al., 1996; Smet et al., 2000; Busca and Pistarino, 2003). Ammonia emissions control is essential to protect the environmental impact and the public health (Erisman et al., 2003). The technologies involved in the treatment of waste gases containing ammonia are based on physical and chemical process, these technologies include incineration, condensation, absorption and adsorption (Busca
227
228
MARTÍN RAMÍREZ, JOSÉ M. GÓMEZ
AND
DOMINGO CANTERO
and Pistarino 2003). However, biological treatments have become an effective and inexpensive alternative to physico-chemical process (Devinny et al., 1999). Up to the present few studies on biofiltration using biotrickling filter system for treatment ammonia emissions are available (Kanagawa et al., 2004; Melse and Mol 2004; Sakuma et al., 2004; Chou and Wang, 2007). The objective of this work was to study the feasibility of a biotrickling filter packed with polyurethane foam particles inoculated with Nitrosomonas europaea. The following operations variables were tested: rate of the recirculation liquid, nitrite concentration in the recirculation liquid, pH, EBRT, ammonia load and pressure drop.
2 MATERIALS AND METHODS 2.1 ORGANISM CULTIVATION AND MEDIUM PREPARATION The original pure-culture strain of autotrophic Nitrosomonas europaea ATCC 19718 was obtained from the American Type Culture Collection. Nitrosomonas europaea is a chemolithoautrotrophic soil bacterium which obtains all its energy and reducing power from the oxidation of NH3 to NO2- (Prosser 1989; Stein and Arp 1998). This stock culture was grown using a rotary shaker at optimal temperature (30ºC) and pH 8.0 in the dark. The mineral medium was the ATCC Medium #2265: Solution 1: 4.95 g of (NH4)2SO4 (for 50 mM NH4+), 0.2 g of KH2PO4, 0.27 g of MgSO4·7H2O, 0.04 g of CaCl2, 0.5 ml of FeSO4 (30 mM in 50 mM EDTA at pH 7.0), 0.2 mg of CuSO4·5H2O in 1.2 litre of distilled water. Solution 2: 8.2 g of KH2PO4, 0.7 of NaH2PO4 in 0.3 litre of distilled water (pH 8.0 with NaOH 10N). Solution 3 (buffer): 0.6 g of Na2CO3 in 12 ml of distilled water. The three solutions were sterilized at 121ºC during 20 min and mixing at room temperature. 2.2 PACKING MATERIAL Polyurethane foam as used in this study as carrier. This material has a surface area of approximately 600 m2.m-3 and a density of 20 kg.m-3 (Devinny et al., 1999). It is an inert material with low density, large porosity (near 96%), good scaling-up possibilities and very low commercial cost (McNevin and Barford 2000; Moe and Irvine 2000). Low density provides advantage in construction and minimizes problems of compaction of packing material. High porosity permits uniform gas flow distribution needed for maximum contact between the gas stream and biofilm biomass. 2.3 IMMOBILIZATION METHOD A PVC column (63 mm of diameter) was used to build-up the biotrickling filter with a working volume of 1.0 L. This column was packed with 10 grams of polyurethane foam cubes. A culture of Nitrosomonas europaea was continuously
REMOVAL OF AMMONIA BY IMMOBILIZED NITROSOMONAS EUROPAEA
229
recirculated over the packed bed using a centrifugal pump (EHEIM 1046) at a constant volumetric flow rate of 26.7 L.h-1. The temperature was controller at 30ºC (Heildoph EKT 3001), and the culture was mixing at 200 rpm (Agimatic-N, Selecta) in the dark. When the pH decreased below 6.0 the total medium was drained and replaced with 1 L of fresh medium without inoculation. Several consecutive batches were run on a «drawn and fill» basis until steady-state biomass levels had been achieved. 2.4 EXPERIMENTAL SET-UP The experimental set up is shown in Figure 1. As biotrickling filter it was used the same column that has been used for the immobilization. The air supply used was compressed air available in the building. Pressure regulation and filtering were achieved by having four filters: silica gel, active carbon, wool glass and Millipore Filter SLG05010 (0.45 μm). Air was humidified using fine bubble diffusion. Flow rates were controlled with mass flow controller (Bronkhorst, Model F-201C). For generation of high loads it was used an ammonia generator column of PVC (63 mm of diameter, packed with glass beads of 5 mm, 25 mm height). A solution whose composition was similar to the liquid culture medium without (NH4)2SO4, the energy source, was added to supplement nutrients. The pH of medium was controller at 6.5-6.6 with addition of NaHCO3 and controller (CRISON PH28). The temperature of experiment was maintained at 30ºC.
Figure 1. Diagram of the experimental setup. 1. Ammonia gas cylinder (NH3/sintetic air); 2. Mass flow controller; 2.1 Rotameter 3. Pressure air regulator; 4. Air prefilter; 5. Humidification system; 6. Expansion deposit; 7. Air filter; 8. Ammonia generation system; 8.1 Peristaltic pump; 8.2. PVC column with glass beads; 8.3. Discharge deposit; 8.4. NH4OH deposit; 9. Biotrickling filer; 10. Recirculation tank; 11. Recirculation pumps; 12. pH control pump; 13. pH controller; 14. NaHCO3 deposit; 15. NH3 sensor.
230
MARTÍN RAMÍREZ, JOSÉ M. GÓMEZ
AND
DOMINGO CANTERO
2.3 ANALYTICAL TECHNIQUES 2.3.1 GAS Ammonia was analysed using an electrochemical sensor of Crowcon (Model GASFLAG, TXGARD-IS). 2.3.2 SUBSTRATE AND PRODUCT CONCENTRATION Ammonia in the recirculation liquid was measure using a Nessler Method and nitrite was measure using a colorimetric method (Clescerl et al., 1989). 2.3.3 IMMOBILIZED BIOMASS Immobilized biomass concentration was measured by counting of total biomass in a Neubauer chamber. A unit of carrier was removed from the reactor and squeezed lightly in order to remove the interstitial liquid. Then, it was submerged in an erlenmeyer flask containing 25 ml of sodium phosphate buffer solution (pH 7.0). In a second step, the flask was placed in an ultrasonic bath at room temperature for 15 min. These conditions led to the total desorption of adhered cells. In the last stage, the Neubauer chamber re-count method for the submerged cells was carried out on the liquid phase. The carrier was subsequently removed from the flask and dried in an oven at 80 °C during 24 h. It was then possible to calculate the number of immobilized cells per milligram of carrier (Gómez et al., 2000; de Ory et al., 2004). This technique has been previously validated by developing experiments concerned with cellular resistance to ultrasonic treatment and studying the desorption efficiency. 2.3.4 ELECTRON MICROSCOPY Scanning electron microscopy (SEM) was used for examination of immobilized bacterial in the carrier, with a microscope FEI QUANTA 200 (Philips) of 2.5 nm of resolution. Fixation with glutaraldehyde (2.5%) at 4 °C for 1 h, cacodylate salt (0.1 M, pH 7.0) for 30 min, dehydration with acetone and drying, and metallization with gold.
3 RESULTS AND DISCUSSION 3.1 IMMOBILIZED BIOMASS The total immobilized biomass at the end of batches was of 3.29±0.52·1010 cel -1 g of solid support in 10 cycles. The duration of the experiment was of 310 hours. The Figure 2 shows the total immobilized biomass in each cycle. The immobilized biomass was very inhomogeneous. In the fourth cycle the biotrickling filter was inundate to homogenize the system and to improve the immobilization. The inundation was realized during one hour before the reposition of recirculation liquid.
REMOVAL OF AMMONIA BY IMMOBILIZED NITROSOMONAS EUROPAEA
231
Figure 2. Immobilized biomass in each cycle.
In Figure 3, we can see the bacteria of Nitrosomonas europaea immobilized about surface of polyurethane foam.
Figure 3. SEM of Nitrosomonas europaea immobilized.
232
MARTÍN RAMÍREZ, JOSÉ M. GÓMEZ
AND
DOMINGO CANTERO
3.2 REMOVAL OF AMMONIA 3.2.1 EFFECT OF LIQUID RECIRCULATION VELOCITY The recirculation liquid in biotrickling filters use to have benefits on micronutrients distribution , absorb gaseous contaminants, remove metabolite, moisten the biofilm and control the thickness biofilm. Therefore the effect of liquid recirculation velocity on the ammonia removal efficiency was tested. The biotrickling filter was operated at 8.57, 3.55 and 1.59 m.h-1. Results showed that the increase of liquid recirculation velocity didn’t have a significant effect on the removal efficiency. 3.2.2 EFFECT OF NITRITE CONCENTRATION In the biological treatment of ammonia gas, ammonium and nitrite are produce and accumulate in the reactor. Therefore, the influence of ammonium and nitrite concentrations on Nitrosomonas europaea was examined. The EBRT was 30 second, the liquid recirculation velocity was 8.57 m.h-1, the load was 6.76 gN.m-3.h-1 and the pH was controlled between 7.5-7.6. Removal efficiencies of 100% were obtained in this experiment, but the biological removal efficiency was not total. The biological removal efficiency was calculated using the following equation obtained from matter of balance:
where, Rb (%)= biological removal efficiency, Q (m3.h-1)= volumetric flow medium, C0 (gN.m-3)= inlet ammonia concentration, CS (gN.m-3)= outlet ammonia concentration, t (h)= time, VL (m-3)= volume of medium recirculation. Because the biotrickling filter was inoculated with a pure culture, acclimation was unnecessary. These results show that the nitrite concentration must be kept below 100 mM N(NO2-) by replacing the recirculation liquid with fresh medium. When the nitrite concentration exceeds 100 mM, the ammonia concentration in the recirculation liquid increased rapidly (Figure 4). 3.2.3 EFFECT OF PH The growth of Nitrosomonas europaea is optimal at pH of 7.5-8.0 (Hunik et al., 1992). The effect of pH on ammonia removal efficiency was studied in the range from 6.5 to 8.2 (Figure 5). The rest of the parameters was fixed as follows: EBRT 30 second, the liquid recirculation velocity 8.57 m.h-1, the load 6.76 gN.m-3.h-1 and the nitrite concentration smaller than 150 mM. When the pH decreased from 7.5 to 6.5, the biological removal efficiency decreased to zero.
REMOVAL OF AMMONIA BY IMMOBILIZED NITROSOMONAS EUROPAEA
233
Figure 4. Effect of nitrite concentration. Ammonia concentration ( ), nitrite concentration ( ) and biological removal efficiency ( ) versus time.
Figure 5. Effect of pH. pH (—). Removal efficiency () and biological removal efficiency ( ) versus time.
234
MARTÍN RAMÍREZ, JOSÉ M. GÓMEZ
AND
DOMINGO CANTERO
Therefore the optimal pH was 7.5, the same value that obtained one for Chung and Huang (1998) using Ca-alginate beads with Nitrosomonas europaea. This result showed that the pH control in the range is important for maintaining a high removal ratio of ammonia. 3.2.4 EFFECT OF LOAD The load on the biotrickling filter was gradually increased by increasing the gas concentration from 60 to 1600 ppmv. The effect of load on ammonia removal efficiency was studied in the range from 0.89 to 21.7 gN.m-3.h-1. The EBRT was maintained constant at 150 second. The pH was controlled between 7.5-7.6 and the liquid recirculation velocity was 8.57 m.h-1. The ammonia gas was successfully treated in the whole range. Therefore, the biotrickling filter attained good operational efficiency (R=100% and Rbe>100% for each load at stationary state). The nitrite concentration of the recirculation liquid was smaller than 150 mM. 3.2.5 EFFECT OF EMPTY BED RESIDENCE TIME The effects of EBRT of 150, 100, 50, 25, 20, 11 and 5 second on the ammonia removal efficiency were tested. The load was maintained constant at 8 gN.m-3.h-1 by increasing the gas concentration from 20 to 592 ppmv. The pH was controlled between 7.5-7.6 and the liquid recirculation velocity was 8.57 m.h-1. In this study removal efficiency of 100% was reached, and the biological removal efficiency was higher than 100% (Figure 6). To avoid inhibition of ammonia removal owing to nitrite concentration, the liquid of recirculation was replaced for fresh medium at 143h of operation. Liang et al. (2000) have also observed that the EBRT can be decreased further without decreasing the ammonia removal efficiency. Chung et al. (1997) obtained a decreased of 20% when the EBRT was decreased from 70 to 12 second working with a biofilter. In a biotrickling filter Chou and Wang (2007) studied the effect of EBRT from 236 to 30 seconds reached removal efficiency of 99 and 96% for EBRT of 59 and 30 seconds respectively. To know the elimination limits, it was realized an experiment maintaining constant EBRT at 11 second and increasing the ammonia concentration from 134 to 1434 ppmv (loads from 24.7 to 270 gN.m-3.h-1). As shown in Figure 7, the biological removal efficiencies was higher than 100% when the nitrite concentration was smaller than 100 mM N(NO2-). When the nitrite concentration was higher than 100 mM the biological removal efficiency decreased and the ammonia concentration increased rapidly. The maximum elimination capacity was 270 gN.m-3.h-1. Working with a biotrickling filter the maximum elimination capacity observed for other researchers were: 59.9 gN.m-3.h-1 (99.8%) (Kanagawa et al., 2004), 33.83 gN.m-3.h-1 (90.0%) (Melse and Mol, 2004), 2.78 gN.m-3.h-1 (98.0%) (Sakuma et al., 2004) and 10.16 gN.m-3.h-1 (94.2%) (Chou and Wang, 2007).
REMOVAL OF AMMONIA BY IMMOBILIZED NITROSOMONAS EUROPAEA
Figure 6. Effect of EBRT (—). Ammonia concentration ( ), nitrite concentration ( ), removal efficiency (+) and biological removal efficiency ( ) versus time.
Figure 7. Effect of load. Ammonia concentration ( ), nitrite concentration ( ), removal efficiency (+) and biological removal efficiency ( ) and Load (—) versus time. EBRT = 11 seconds.
235
236
MARTÍN RAMÍREZ, JOSÉ M. GÓMEZ
AND
DOMINGO CANTERO
3.2.6 STUDY OF PRESSURE DROP The pressure drop in the biotrickling filter was measure before and after the immobilization. The pressure drop increased from 6.2 to 25.9 cm of water per meter of column when decreased de EBRT from 11 to 5 second after immobilization (Figure 8).
Figure 8. Pressure drop per meter of column versus superficial gas velocity. With biomass; without biomass.
4 CONCLUSIONS The biotrickling filter can be considered as a good configuration for treatment waste gases contaminated with ammonia. This configuration allows a rapid absorption and biological oxidation of ammonia in the liquid. Ammonia gas was successfully treated at a load of 270 gN.m-3.h-1 when nitrite concentration was smaller than 100 mM N(NO2-) at EBRT of 11 seconds. The more important parameters affecting the performance of this configuration are: pH (optimal 7.5) and nitrite concentration in the recirculation liquid (optimal <100mM).
5 ACKNOWLEDGEMENTS Authors wish to express their sincere gratitude to Spanish Government for financial support of this work through Project PPQ2002-0217.
REMOVAL OF AMMONIA BY IMMOBILIZED NITROSOMONAS EUROPAEA
237
REFERENCES Busca, G. and Pistarino, C. (2003) Abatement of ammonia and amines from waste gases: A summary. Journal of Loss Prevention in the Process Industries 16(2): 157-163. Carson P. and Mumford, C.J. (2002) Hazardous Chemicals Handbook: Elsevier ButterworthHeinemann. 568 p. Clescerl, L.S., Greenberg, A.E. and Eaton, A.D. (1989) Standard Methods for Examination of Water and Wastewater. APHA/AWWA/WPCF. Association APH, Editor. Chou, M.-S. and Wang, C.-H. (2007) Treatment of ammonia in air stream by biotrickling filter. Aerosol and Air Quality Research. 7(1): 17-32. Chung, Y.C. and Huang, C. (1998) Biotreatment of ammonia in air by an immobilized Nitrosomonas europaea biofilter. Environ. Prog. 17(2): 70-75. Chung, Y.C., Huang, C. and Tseng, C.-P. (1996) Reduction of H2S/NH3 production from pig feces by controlling environmental conditions. J. Environ. Sci. Health - Part A Toxic/Hazardous Substances and Environmental Engineering 31(1): 139-155. Chung, Y.C., Huang, C. and Tseng, C.P. (1997) Biotreatment of ammonia from air by an immobilized Arthrobacter oxydans CH8 biofilter. Biotechnol. Prog. 13(6): 794-798. de Ory, I., Romero, L.E. and Cantero, D. (2004) Optimization of immobilization conditions for vinegar production. Siran, wood chips and polyurethane foam as carriers for Acetobacter aceti. Proc. Biochem. 39(5): 547-555. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for Air Pollution Control: Lewis Publishers. Erisman, J.W., Grennfelt, P. and Sutton, M. (2003) The European perspective on nitrogen emission and deposition. Environ. Interna.l 29(2-3): 311-325. Gómez, J.M., Cantero, D. and Webb, C. (2000) Immobilisation of Thiobacillus ferrooxidans cells on nickel alloy fiber for ferrous sulfate oxidation. Appl. Microbiol. Biotechnol. 54(3): 335-340. Hunik, J.H., Meijer, H.J.G. and Tramper, J. (1992) Kinetics of Nitrosomonas europaea at extreme substrate, product and salt concentrations. Appl. Microbiol. Biotechnol. 37(6): 802-807. Kanagawa, T., Qi, H.W., Okubo, T. and Tokura, N. (2004) Biological treatment of ammonia gas at high loading. Wat. Sci. Technol. 50(4): 283-290. Liang, Y., Quan, X., Chen, J., Chung, J.S., Sung, J.Y., Chen, S., Xue, D. and Zhao, Y. (2000) Longterm results of ammonia removal and transformation by biofiltration. J. Hazard. Mat. 80 (1-3): 259-269. McNevin, D. and Barford, J. (2000) Biofiltration as an odour abatement strategy. Biochem. Eng. J.l 5(3): 231-242. Melse, R.W. and Mol, G. (2004) Odour and ammonia removal from pig house exhaust air using a biotrickling filter. Wat. Sci. Technol. 50(4): 275-288.
238
MARTÍN RAMÍREZ, JOSÉ M. GÓMEZ
AND
DOMINGO CANTERO
Moe, W.M. and Irvine, R.L. (2000) Polyurethane foam medium for biofiltration. I: Characterization. J. Environ. Engin. 126(9): 815-825. Prosser, J.I. (1989) Autotrophic nitrification in bacteria. Advances Microb. Physiol. 30: 125-181. Sakuma, T., Aoki, M., Hattori, T., Gabriel, D. and Deshusses, M.A. (2004) A conceptual model for the treatment of ammonia vapors in a biotrickling filter. Proc. Annual Meeting and Exhibition of the Air and Waste Management Association, p 1531-1546. Smet, E., Van Langenhove, H. and Maes, K. (2000) Abatement of high concentrated ammonia loaded waste gases in compost biofilters. Water Air Soil Poll. 119(1-4): 177-190. Stein, L.Y. and Arp, D.J. (1998) Ammonium limitation results in the loss of ammonia oxidizing activity in Nitrosomonas europaea. Appl. Environ. Microbiol. 64(4): 1514-1521.
239
Study of NH3 removal by gas-phase biofiltration: effects of shock loads and watering rate on biofilter performance GUILLERMO BAQUERIZO, JUAN PEDRO MAESTRE*, XAVIER GAMISANS, DAVID GABRIEL* AND JAVIER LAFUENTE* Department of Mining Engineering and Natural Resources, Universitat Politècnica de Catalunya, Bases de Manresa 61-73, 08240 Manresa, Spain * Department of Chemical Engineering, Universitat Autònoma de Barcelona, Edifici C, 08193 Bellaterra, Barcelona, Spain
ABSTRACT Ammonia biofiltration performance under shock loads episodes was studied in a reactor packed with coconut fiber as carrier material. Periodical gas and leachate samplings were analyzed and used to characterize the biofilter performance in terms of removal efficiency (RE) and elimination capacity (EC). Nitrogen fractions in the leachate were quantified to identify the experimental rates of nitritation and nitratation.. In a primary experiment a sudden increment of ammonia load was applied for 1 day by changing the ammonia inlet load from 5.2 to 29.1 g N.m-3.h-1. Even though stable operation was obtained (RE of 99.9%), a notable accumulation of nitrite was verified in the leachate. Experimental rates showed that nitritation increased at the same the same ratio that ammonia load was varied. However the nitratation seemed to be largely affected by high ammonia and nitrite concentration. In a subsequent experiment varying the inlet ammonia load, the system was rapidly recovered by increasing the watering rate. Since ammonia was partially removed by physicochemical process as observed in previous experiments, a final experimental was conducted to improve the nitritation capacity. The addition of inorganic carbon source demonstrated to enhance the capacity of the biofilter to degrade a higher amount of ammonia.
239
240
G. BAQUERIZO, J. P. MAESTRE, X. GAMISANS, D. GABRIEL AND J. LAFUENTE
1 INTRODUCTION Ammonia is a colorless, toxic, and strong odor gas produced in many industrial and agricultural processes (Ryer-Power, 1991). Traditionally physical-chemical treatments have been applied to remove ammonia from contaminated air streams (Busca and Pistarino, 2003). Although traditional methods provide high odor removal efficiency, some drawbacks such as elevated cost operation, important energy consumption or by-products generation are associated to these technologies. During the last years biofiltration has emerged as a reliable, environmental friendly and cost effective technology for odor control. This technology has been successfully used to treat a wide range of pollutants at relatively low concentration (Devinny et al., 1999; Kennes and Thalasso, 1998). The main advantage of biofiltration is that pollutants are oxidized into harmless products. Currently a few number of studies dealing with ammonia abatement by biofiltration are available in literature. The majority of these works have been focused in determining ammonia removal capacity using different carrier materials (Chen et al., 2005; Hartikainen et al., 1996; Hirai et al., 2001;. Kim et al., 2000; Yani et al., 1998). Indeed a broad range of operation conditions have been applied in biofilters treating ammonia. While EBRT is normally set lower than 1 minute, ammonia inlet concentrations have been varied in the range of 10 to 1000 ppmv. Consequently substantial differences have been reported regarding the maximum ammonia elimination capacity obtained by biofiltration. In early studies, capacities lower than 10 g N.m-3.h-1 at inlet concentration between 10 and 50 ppmv were reported (Joshi et al., 2000; Van Langenhove et al., 1998). Additionally some of these studies pointed out that biofilter removal efficiency are strongly affected at ammonia inlet concentration exceeding 60 ppmv (Don, 1985; Hartikainen et al., 1996). More recently elimination capacities up to 40 g N.m-3.h-1 have been reported operating at inlet concentration up to 1000 ppmv (Kanagawa et al., 2004). By the other hand studies dealing with the effects of fluctuating conditions in ammonia operation have been scarcely reported in literature, despite biofiltration operation is often exposed to varying operating conditions. Partial oxidation of ammonia during biofiltration has been frequently reported. Some works have found that 50% of ammonia is nitrified and the other is retained in the packed bed as ammonium (Chen et al., 2005; Don, 1985; Smet et al., 2000). Likewise amounts of ammonia and nitrate at a ratio of 1:1 have recurrently observed in either liquid drain or packed bed in ammonia biofilters (Baquerizo et al., 2005; Chen et al., 2005; Kanagawa et al., 2004; Kim et al., 2000; Liang et al., 2000). Kinetics analyses have been usually conducted in ammonia biofilters to describe nitrification process using a Monod expression and assuming a plug flow pattern to describe the gas phase (Chung and Huang, 1998; Yani et al., 2000). However
STUDY OF NH3 REMOVAL BY GAS-PHASE BIOFILTRATION
241
nitrification is better described as two intermediate steps: oxidation from ammonium to nitrite and oxidation from nitrite to nitrate (Baquerizo et al., 2005). Nitrification inhibition by nitrogen species have been reported but this phenomenon is usually ignored in the Monod kinetic approach. The objective of this study is to determine the influence of both shock loads and watering rate on the performance of a biofilter treating ammonia. The study was performed using data collected from both gas and leachate measurements. Nitrogen fractions in the leachate were employed to identify the experimental rates (nitritation and nitratation). The utilization of a carbon supply to improve the nitrification capacity of the biofilter was also studied.
2 MATERIALS AND METHODS 2.1 EXPERIMENTAL SETUP The biofiltration experiments were performed in a laboratory-scale plant. As main characteristics of the setup, the reactor was constructed using transparent PVC and divided into four equivalent modules, with a total bed height of 1 m and an inner diameter of 0.1 m. Each module was packed with coconut fiber to a height of 20 cm, meaning a total volume of the filter bed of 6.3 L. A schematic of the pilot-unit and a comprehensive description of biofilter automation and characterization of the organic packing material can be found elsewhere (Baquerizo et al., 2005). The inlet gaseous stream was obtained by mixing compressed air and pure ammonia from a cylinder. Both gas flows were measured and controlled by means of digital mass flow controllers (Bronkhorst, NL) which allowed adjusting accurately the EBRT and the ammonia inlet concentration. The air stream was previously humidified using a PVC humidification column. The reactor was operated in up flow mode and a nutrient solution was supplied periodically from the top of the reactor. The bottom was fitted with an automated liquid drain system. The watering load applied to the reactor was equal to 750 ml.d-1. Coconut fiber used as packing material was withdrawn from a full-scale biofilter at a municipal solid waste treatment plant facility. No inoculation was needed since the full-scale biofilter had been running for more than 2 years at an average ammonia inlet concentration of 40 ppmv (Gabriel et al., 2007). 2.2 ANALYTICAL METHODS Continuously monitored parameters included temperature and relative humidity (Testo, Hygrotest 600 PHT), besides data logging of pumps and valves actuations. Ammonia gas was measured on-line using an ammonia electrochemical sensor (Vaisala, AMT102). Leachate volume and composition analyses were done periodically.
242
G. BAQUERIZO, J. P. MAESTRE, X. GAMISANS, D. GABRIEL AND J. LAFUENTE
Conductivity and pH were measured with lab probes (Crison, microCM 2100 and MicropH 2001 respectively) prior to filtering. The concentration of NO2- and NO3- in aqueous samples (leachate) was determined by capillary electrophoresis in a Quanta 4000E unit (Waters). Ammonia and ammonium content in leachate was measured in a continuous flow analyzer (Baeza et al., 1999). 2.3 NITRIFICATION RATES AND MASS BALANCES In this study the nitrification process is described by means of two intermediate steps: oxidation from ammonium to nitrite (nitritation) and oxidation from nitrite to nitrate (nitratation). Stoichiometric equations for each reaction can be derived assuming a representative biomass composition (Baquerizo et al., 2007). Experimental nitritation and nitratation rates are calculated according to the Equations (1) and (2), which are directly derived from ammonia measurements in the gas phase and leachate analysis.
(1)
(2)
where R1 is the nitritation rate (g N-NH4+ consumed.m-3 h-1), NH4(abs) is the amount of ammonium (g N-NH4+) absorbed in the packed bed during the time interval «t over two leachate sampling events, NH4(leach) is the amount of ammonium (g NNH4+) collected in the leachate during the time interval Δt (h), and V is the reactor bed volume (m3). R2 is the nitratation rate (g N-NO3- produced.m-3 h-1), and NO3(leach) is the amount of nitrate (g N-NO3-) collected in the leachate during the time interval Δt (h).
3 RESULTS AND DISCUSSION 3.1 EFFECTS OF AMMONIA SHOCK LOAD Prior to the experiment, the biofilter had been operated under steady-sate conditions at an average inlet concentration of 90 ppmv for more than 1 year with a EBRT of 36 s (EC equal to 5.2 g N.m-3.h-1). Stable and efficient operation (RE equal to 99.9%, data non shown) was verified during the aforementioned period. Afterward a stepwise increasing of ammonia load was applied by both varying the inlet concentration from 90 to 260 ppmv and decreasing the EBRT from 36 to 19 s (Figure 1). Ammonia inlet load was raised up from 5.2 to 29.1 g N m-3 h-1 in a period of 1 day. Stable operation was verified under high load operation in which removal efficiency
STUDY OF NH3 REMOVAL BY GAS-PHASE BIOFILTRATION
243
was maintained at 100% over the whole experiment as can be seen in Figure 1. Results are in coincidence with previous ammonia biofiltration studies in which high RE were achieved under sudden variation or, at least, short acclimation periods (less than 2 days) were necessary to recover a complete removal efficiency after some operating condition variation (Chen et al., 2004; Kim et al., 2000).
Figure 1. Evolution of the main biofilter parameters: (a) removal efficiency, EBRT, EC, and inlet concentration; (b) nitrite, nitrate, ammonium, and pH; (c) experimental nitritation and nitratation rates.
244
G. BAQUERIZO, J. P. MAESTRE, X. GAMISANS, D. GABRIEL AND J. LAFUENTE
However drain analyses showed important variations in the leachate composition. Ammonium and nitrite concentration increased while nitrate decreased notably. A slight increment of pH was observed for high load operation. Higher pH values indicate that sorption processes prevail over the ammonia biodegradation (nitrification). The examination of experimental rates reveals that nitritation rate increased at the same ratio as ammonia load was augmented. It should be emphasized that nitritation and nitratation rates show similar values before the increasing of ammonia load. In that sense, nitritation rate seems to be not affected by elevated ammonium and nitrite concentrations. Conversely nitratation rate kept constant after the load increase, showing that nitrite oxidizing bacteria are largely affected by high both ammonium and nitrite concentration (higher than 800 and 300 mg N L -1, respectively). Consequently an increase of nitrite concentration was observed in the leachate. Inhibition of nitrogen species over the nitrification process have been reported previously (Baquerizo et al., 2005), but no experimental report had been provided so far specifying the species affecting either nitritation or nitratation in ammonia biofiltration. Analyses performed after shock load episode showed a rapid increment of nitrate concentration while a slight diminishing of nitrite concentration was also observed. High amount of ammonium recovered in the leachate in subsequent days are probably explained for ammonia accumulation in the packed bed during the high load operation. In overall, more than 1 week was necessary for recovering the same reactor conditions (i.e. leachate composition and experimental rates) before the shock load episode. 3.2 EFFECTS OF WATERING RATE TO RECOVER BIOFILTER CONDITIONS AFTER SHOCKS LOADS EPISODES
The influence of watering rate to recover biofilter performance under rapid changes in ammonia load was also studied. Keeping the EBRT constant (24 s), high load shocks were applied by increasing inlet ammonia concentration from 90 to 260 ppmv for 12 h, corresponding to an inlet load variation from 5.2 to 15.4 g N.m-3.h-1. Subsequently inlet concentration was varied from 260 to 170 ppmv for 12 h (load from 15.4 to 9.9 g N. m-3.h-1). Similarly to the previous experiment, RE showed a stable value of 100% during the entire experiment (data no shown), while the nitration rates showed that the system was operating under inhibition conditions (Figure 2). Again nitritation rate increased the same ratio as ammonia load was varied. Nitratation was inhibited under high concentration of nitrite and ammonium and therefore nitratation rate remained constant under the ammonia load increment.
STUDY OF NH3 REMOVAL BY GAS-PHASE BIOFILTRATION
245
Figure 2. Profiles for experiment varying the watering rate: (a) nitrite, nitrate, ammonium, and pH; (b) experimental nitritation and nitratation rates.
Leachate analyses showed an increment of ammonia concentration while a continuous reduction of nitrate was verified. In addition a constant increase of nitrite was also observed (Figure 2). Watering rate was increased 3 times (i.e. 2250 ml.d-1) in order to recover the system and to reestablish the same conditions observed before the load increase (i.e. R1 = R2). The effects of high watering rates are clearly observed in Figure 2 where the concentration of each nitrogen species was reduced. However the large amounts of water applied over the reactor promotes that ammonia is mainly removed by adsorption followed by a washing out process. Indeed the nitritation rate under high watering rate dropped off until minimum values as can be seen in Figure 2. Amounts of nitrate collected in the leachate are obviously produced by the nitrate oxidation, which is not inhibited by large quantities of ammonium and nitrite and thus nitratation rate reached a maximum value.
246
G. BAQUERIZO, J. P. MAESTRE, X. GAMISANS, D. GABRIEL AND J. LAFUENTE
High values of nitrogen species concentration observed after watering increase (day 8) are explained for the drastic reduction of water amounts applied over the reactor bed (watering was reestablished to the initial rate). Indeed the amounts of nitrogen species collected from day 8 (i.e. g N-recovered d-1) are similar to those observed before the increase of ammonia load. Furthermore, a similar value of both experimental rates (R1 and R2) was verified after day 8. Therefore, the reestablishment of the operating conditions before the shock loading episode was confirmed after 3 days by applying higher water flows over the reactor bed. 3.3 EFFECTS OF INORGANIC CARBON SUPPLY TO ENHANCE THE NITRIFICATION PROCESS Ammonium mass percentage of about 50% was encountered in the leachate in the experiment described above as well as in long operation periods at ammonia inlet concentration of 90 ppmv. This result confirms that ammonia removal by biofiltration is also achieved by sorption processes. High concentrations of ammonium in the leachate may lead to think that the nitrification capacity of the biofilter could be enhanced. In order to improve the nitritation rate (i.e. decreasing the amount of ammonium recovered in the leachate), a carbon supply was applied to the biofilter. Sodium bicarbonate was used as carbon source to improve the performance of autotrophic ammonium oxidizers bacteria. A concentration of 2 g C-NaHCO3.L-1 was supplied in each watering. Before applying the extra carbon source the reactor was operated at 260 ppmv using an EBRT of 24 s (load of 0.55 kg N m-3 d-1) for more than 100 days. Stable operation and removal efficiencies around 98-100% were obtained but a high concentration of ammonium in the leachate was also confirmed (mass percentage of 50%, data no shown). A decrease of 20% of ammonium mass percentage in the leachate after applying sodium bicarbonate (from 50% to 40%) can be observed in Figure 3. The inorganic carbon supply allowed increasing the ammonia removal capacity in the reactor by biodegradation instead of absorption. However a decrease of nitrate amounts collected in the leachate was also observed due to the increase of nitrite concentration which inhibited the nitratation process. In Figure 4, the ratio of each nitrogen species to total nitrogen amount in the leachate is depicted. Despite of the diminution of ammonia content in the leachate (i.e. increase of R1) the nitratation is negatively affected by the increase of the nitrite concentration in the packed bed. Indeed nitrate rate in the leachate remained constant along the experiment. The increment of total organic compounds in the leachate (TOC, Figure 4) is linked to the growth of ammonium oxidizing bacteria in the reactor.
STUDY OF NH3 REMOVAL BY GAS-PHASE BIOFILTRATION
Figure 3. Mass percentage distribution in the leachate for carbon supply experiment.
Figure 4. Ratio of nitrogen species and TOC content in the leachate.
247
248
G. BAQUERIZO, J. P. MAESTRE, X. GAMISANS, D. GABRIEL AND J. LAFUENTE
4 CONCLUSIONS Effects of shock load over ammonia biofiltration have been studied including both gas phase behavior and drain analyses. Results demonstrated that biofilter is capable to operate at high removal efficiencies (higher than 99%) under sudden increases of ammonia load. However substantial alterations in experimental rates describing the overall nitrification process were observed. The nitritation rate increased at the same ratio than ammonia load was augmented revealing that the first nitrification step is not affected by accumulation of ammonium and nitrite in the packed bed. On the contrary nitratation rate is largely inhibited by accumulation of nitrite and thus its value remained constant under the ammonia load increment. Despite RE kept constant, experimental rates need several days to recover their original values. Increasing of watering load allow accelerating the recovering process. However excessive amount of water may promote that the fraction of ammonia removed by sorption process increases. Experimental results showed that practically the 50% of the total nitrogen species collected in the leachate correspond to ammonium, confirming that sorption process play a fundamental role in ammonia biofiltration. The addition of the inorganic carbon source (NaHCO3) enhances the activity of the autotrophic ammonium oxidizing consortium, improving the reactor capacity to biodegrade ammonia. Nevertheless a higher nitritation rate causes larger concentration of nitrite that inhibits the oxidation of nitrite to nitrate. Higher removal capacities can be obtained avoiding high amount of nitrogen species in the filter bed by increasing the EBRT and modifying the watering rate.
REFERENCES Baquerizo, G., Maestre, J.P., Sakuma, T., Deshusses, M.A., Gamisans, X., Gabriel, D. and Lafuente, J. (2005) A detailed model of a biofilter for ammonia removal: model parameters analysis and model validation. Chem. Eng. J. 113(2-3): 205-214. Baquerizo, G., Gamisans, X., Gabriel, D. and Lafuente, J. (2007) A dynamic model for ammonia abatement by gas-phase biofiltration including pH and leachate modelling. Biosyst. Eng. In press. Baeza, J., Gabriel, D. and Lafuente, J. (1999) An expert supervisory system for a pilot WWTP. Environ. Modell. & Software. 14(5): 383-390. Busca, G. and Pistarino, C. (2003) Abatement of ammonia and amines from waste gases: a summary. J. Loss. Prev. Process. Indust. 16(2): 157-163. Chen, Y-X., Yin, J., Wang, K-X. and Fang, S. (2004) Effects of periods of nonuse and fluctuating ammonia concentration on biofilter performance. J. Environ. Sci. Health., Part A. A39(9): 2447-2463.
STUDY OF NH3 REMOVAL BY GAS-PHASE BIOFILTRATION
249
Chen, Y-X., Yin, J. and Wang, K-X. (2005) Long-term operation of biofilters for biological removal of ammonia. Chemosphere. 58(8): 1023-1030. Chung, Y-C. and Huang, C. (1998) Biotreatment of ammonia in air by an immobilized Nitrosomonas europaea biofilter. Environ. Prog. 17(2): 70-76. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. Boca Raton: Lewis Publishers. 299 p. Don, J.A. (1985) The rapid development of biofiltration for the purification of diversified waste gas streams. VDI Berichte. 561: 63-73. Gabriel, D., Maestre, J.P., Martín, L., Gamisans, X. and Lafuente, J. (2007) Characterisation and performance of coconut fibre as packing material in the removal of ammonia in gas-phase biofilters. Biosyst. Eng. In press. Hartikainen, T., Ruuskanen, J., Vanhatalo, M. and Martikainen, P.J. (1996) Removal of ammonia from air by a peat biofilter. Environ. Technol. 17(1), 45-53. Hirai, M., Kamamoto, M., Yani, M. and Shoda, M. (2001) Comparison of the biological NH3 removal characteristics among four inorganic packing materials. J. Biosci. Bioeng. 91(4): 428-430. Joshi, J.A., Hogan, J.A., Cowan, R.M., Strom, P.F. and Finstein, M.S. (2000) Biological removal of gaseous ammonia in biofilters: space travel and earth based applications. J. Air Waste Manage. Assoc. 50(9): 1647-1654. Kanagawa, T., Qi, H. W., Okubo, T. and Tokura, N. (2004) Biological treatment of ammonia gas at high loading. Water Sci.Technol. 50(4): 283-290. Kennes, C. and Thalasso, F. (1998) Waste gas biotreatment technology. J. Chem. Technol. Biotechnol. 72(4): 303-319. Kim, N-J., Hirai, M. and Shoda, M. (2000) Comparison of organic and inorganic packing materials in the removal of ammonia gas in biofilters. J. Hazard. Mater. 72(1), 77-90. Liang, Y., Quan, X., Chen, J., Chung, J.S., Sung, J.Y., Chen, S., Xue, D. and Zhao, Y. (2000) Longterm results of ammonia removal and transformation by biofiltration. J. Hazard. Mater. 80(1-3): 259 -269. Ryer-Power, J.E. (1991) Health effects of ammonia. Plant/Operations Prog. 10(2): 228-232. Smet, E., Van Langenhove, H. and Maes, K. (2000) Abatement of high concentrated ammonia loaded waste gases in compost biofilters. Water, Air, Soil Pollut. 119(1-4): 177-190. Van Langenhove, H., Lootens, A. and Schamp, N. (1988) Elimination of ammonia from pigsty ventilation air by wood bark biofiltration. Med. Fac. Landbouww. Rijksuniv. Gent. 53: 1963-1969. Yani, M., Hirai, M. and Shoda, M. (1998) Ammonia gas removal characteristics using biofilter with activated carbon fiber as a carrier. Environ. Technol. 19(7): 709-715. Yani, M., Hirai, M. and Shoda, M. (2000) Enhancement of ammonia removal in peat biofilter seeded with enriched nitrifying bacteria. Environ. Technol. 21(10): 1199-1204.
This page intentionally left blank
251
High H2S concentrations abatement in a biotrickling filter: start-up at controlled pH and effect of the EBRT and O2/H2S supply ratio MARC FORTUNY1, MARC A. DESHUSSES3, XAVIER GAMISANS2, CARLES CASAS1, DAVID GABRIEL1 AND JAVIER LAFUENTE1 1 Department of Chemical Engineering, ETSE, Universitat Autònoma de Barcelona, Barcelona, Spain 2 Department of Mining Engineering and Natural Resources, Universitat Politècnica de Catalunya, Manresa, Spain 3 Department of Chemical & Environmental Engineering, University of California, Riverside, California, USA
ABSTRACT In this study, a biotrickling filter reactor was set up and used to treat high concentrations of gaseous H2S. Inoculation was carried out at an inlet H2S concentration of 1,000 ppmv (27.8 g H2S m-3 h-1) and sludge from a municipal wastewater treatment plant (MWWTP) was used as inoculum. After 3 days, removal efficiency (RE) above 98 % was achieved even after the loading rate (LR) was increased up to 55.6 g H2S m-3 h-1 (2,000 ppmv). Operation at such LR, with an empty bed residence time (EBRT) of 180 s and controlled pH of 6.5-7 was carried out during 3 months. The start-up phase, the effect of decreasing EBRTs at constant inlet concentration and the composition of the process end-products in relation to the supplied O2/H2S ratio were studied. Also, a carbon mass balance under steady state conditions was calculated.
1 INTRODUCTION There are multiple industrial processes which produce biogas as a by-product of their main objective, but still it is a common practice only to use it as a heat-power source or just to burn it at the torch. This is mainly because secondary components present in biogas make it technically difficult and economically expensive to use it for electric
251
252
M. FORTUNY, M. A. DESHUSSES, X. CAMISANS, C. CASAS, J. LAFUENTE AND DAVID GABRIEL
power production. Among them, hydrogen sulfide (H2S), a corrosive, toxic and odorous gas, typically representing from 0.1 to 2 % (vv-1) of the biogas (Janssen et al., 1998), has a major role since concentrations below 500 ppmv (0.05 % vv-1) are usually required for biogas burning engines. However, biogas energy recovery is becoming more and more interesting due to the increasing environmental and economical problems associated to fossil fuels and because an increasing number of solid and liquid waste management facilities are being installed with biogas energy recovery as the main economic benefit. So far, most commonly applied technologies are adsorption and absorption processes, but their high operating costs represent an important drawback and, thus, other less expensive alternatives are being developed. Biological techniques have proven to be a suitable, environmental-friendly alternative for low H2S concentrations treatment (Devinny et al., 1999; Yang and Allen, 1994; Gabriel and Deshusses, 2003), although few references can be found dealing with biological treatment of high concentrations of H2S in biotrickling filters (Fortuny et al., 2006; Bailón, 2005). Biological H2S abatement is based on biological sulphur metabolism, being the main reactions involved (Kuenen, 1975):
In a previous study, H2S concentrations up to 10,000 ppmv were proven to be successfully treated in a biotrickling filter with an empty bed retention time of 180 s. However, some operational problems such as liquid pH variation, carbon limitation and sulphur accumulation due to very low O2/H2S supplied ratios hindered reactor performance and life span, and probably contributed to a very long start-up phase of about 25 days (Fortuny et al., 2006). Thus, the purpose of this study was to obtain a better knowledge of the H2S biological oxidation process, through a deeper insight in some of the operational parameters such as the EBRT and the process end-products speciation in relation to the amount of supplied oxygen. Also, a different approach for the reactor start-up was investigated.
2 MATERIALS AND METHODS 2.1 EXPERIMENTAL SETUP In this work, an experimental reactor based on a conventional biotrickling filter with a separated oxygen supply system was used. (Fig. 1). HD-QPAC® (Lantec Products
HIGH H2S CONCENTRATIONS ABATEMENT IN A BIOTRICKLING FILTER
253
Figure 1. Schematic of the lab-scale setup. 1: Main reactor; 2: Air supply reactor; 3: Gas inlet; 4: Gas outlet; 5: HCO3- supply; 6: Gas monitoring; 7: MM supply; 8: Recirculation pump; 9: pH control; 10: Liquid monitoring; 11: Air supply; 12: Level control; 13: Liquid purge.
254
M. FORTUNY, M. A. DESHUSSES, X. CAMISANS, C. CASAS, J. LAFUENTE AND DAVID GABRIEL
Inc., CA, USA) with a 4 × 4 mm (0.16" × 0.16") grid opening cut to tightly fit inside the reactor was used as packing material since its regular and open structure had proved to better suit the system requirements in previous research (Fortuny et al., 2006). Operation was continuously carried out for a period of three months at EBRT of 180 s, an average liquid retention time (LRT) of 54 ± 7 h , an inlet concentration of 2,000 ppmv (55.6 g H2S m-3 h-1) and a liquid recirculation velocity (LRV) of 3.6 m h-1 (241 ml min-1). Metered amounts of H 2S, N2 and air using digital mass flow controllers (Bronkhorst, The Netherlands) were used to simulate a controlled biogas inflow. Mineral medium (MM) containing (g L-1): NH4Cl, 1; KH2PO4, 0.12; K2HPO4, 0.15; CaCl2, 0.02; MgSO4·7H2O, 0.2; trace elements, 1 ml L-1, and NaHCO3 as inorganic carbon source were continuously fed. Liquid phase was continuously renewed by automated timing of the MM supply, bicarbonate supply and the liquid purge, using 3 different peristaltic pumps (Fig. 1). 2.2 ANALYTICAL METHODS Continuous monitoring of outlet H2S and CO2 gas phase concentrations was performed using an electrochemical H2S sensor (Sure-cell, Euro-Gas Management Services LTD, UK) and a Carbocap® Carbon Dioxide Probe GMP343 (Vaisala, Helsinki, Finland). On-line liquid phase monitoring included pH, oxidation-reduction potential (ORP) and dissolved oxygen (DO) measurements. A pH control by HCl or NaOH addition and a level control by liquid purge regulation was also installed. Also, daily samples of liquid outlet were taken for inorganic carbon (TIC) and sulphur ionic species analysis using a TOC 1020 analyzer (IO Analytical) and an ICS1000 Ion Chromatography system with an IonPac AS9-HC column (Dionex Corporation), respectively. Measurement of dissolved sulfide species (H2S, HS-, S2-) was also carried out by flow injection analysis (Delgado et al., 2006). 2.3 REACTOR INOCULATION AND START-UP Reactor inoculation was carried out using aerobic sludge from a MWWTP. A sludge volatile suspended solids (VSS) concentration of 1.9 g L-1 and an inlet H2S concentration of 1000 ppmv (27.8 g H2S m-3 h-1) were used. During the first four days no new MM was supplied excepting NaHCO3 to ensure no carbon limitation. During that time, 10 % of the liquid phase volume was twice removed (second and third days) in order to keep the reactor volume constant.
HIGH H2S CONCENTRATIONS ABATEMENT IN A BIOTRICKLING FILTER
255
3 RESULTS AND DISCUSSION 3.1 INOCULATION AND START-UP After one hour of operation at 1,000 ppmv inlet concentration, significant amounts of H2S were already detected in the outlet gas phase, therefore showing a low sorption capacity of the system, even working at constant pH = 7 (Fig. 2). However, the RE did not drop under 60 % during the first day, operating at very low DO and ORP values, and raised up to 70 % the second day after an oxygen supply increase. From then on, a progressive removal efficiency increase leading to outlet concentrations under the H2S setup detection limit (thus having RE over 97 %, Fig. 2b) was observed, whilst the DO decreased and the ORP increased (Fig. 2a). Increasing ORP and decreasing DO measurements probably indicated a progressive change in the sulphurspecies liquid-phase composition, i.e. from dominating H2S(aq) and HS- to SO42- (Lens and Hulshoff, 2000), due to an increasing biological oxygen consumption.
Figure 2. a): pH, ORP and DO during start-up. b): RE, [H2S]in and [H2S]out during start-up.
256
M. FORTUNY, M. A. DESHUSSES, X. CAMISANS, C. CASAS, J. LAFUENTE AND DAVID GABRIEL
Accordingly, an initial accumulation and subsequent depletion of thiosulfate and inorganic carbonate during the first two days and a progressive accumulation of sulfate and phosphate were detected in the liquid phase (Fig. 3). Hence, it seems that an initial sorption process could have been supported by a low biological sulfide oxidation activity from the inoculum biomass. Favourable operating conditions from the very beginning contributed to a fast start-up, since operating pH was kept constant at the same original inoculum pH. In addition, favourable room temperature and excess inorganic carbonate were ensured.
Figure 3. Liquid phase ionic composition. Arrow indicates beginning of liquid renewal.
However, the thiosulfate trend indicates that the second and third days of operation, after increasing the oxygen supply, the chemical oxidation of sulfide to thiosulfate under low biological sulfide activity (Janssen et al., 1995) was the dominating process, subsequently being substituted by an already favoured biological oxidation to sulfate from the third day on. Inorganic carbonate and phosphate measurements were also in agreement. At a constant carbonate supply (0.71 ± 0.08 g C-NaHCO3 g-1 S-H2S) the initial accumulation was consumed from the second day on, even if until the sixth day no liquid phase renewal was applied. Also, probably an initially growing biomass population would have had more phosphate requirements, thus leading to a phosphate accumulation trend up to the 15th -20th day after starting operation. Thus, a very short start-up phase of only 3 to 5 days was observed. Moreover, even after increasing the LR up to 55.6 g H2S m-3 h-1 (2,000 ppmv) the sixth day of operation, the system performance in terms of RE did not showed any appreciable drop; on the contrary, increased accordingly (Fig. 2).
HIGH H2S CONCENTRATIONS ABATEMENT IN A BIOTRICKLING FILTER
257
This is a much shorter time than the observed in similar previous studies (Fortuny et al., 2006). Such a fast start-up has been attributed to two main topics. On the one hand, the pH control ensured a constant operation at a pH between 6.5 and 7.0 (Fig. 2, a), the same original inoculum pH. Also, the pH control initially avoided the pH increase which would have been caused by the constant addition of NaHCO3 at an early stage of the system operation, with not enough sulfate production to balance the pH. Actually, the possibility to constantly add bicarbonate without altering the pH was a very positive aspect that contributed to a fast start-up since no carbon limitation could be guaranteed. On the other hand, the WWTP sludge used as inoculum may have played an important role in the process. Most probably the high biomass concentration facilitated the biofilm formation onto the new packing material (Prado et al., 2005) and enough sulfide oxidizing biomass ensured a fast adaptation to the new substrate therefore facilitating the start-up phase. Overall, this results show that it is not always worth spending time, energy and money obtaining a specific culture in order to start-up a biological treatment system, as it has usually been reported (Fortuny et al., 2006; Veiga and Kennes, 2001; Duan et al., 2006). Sludge from MWWTP can perfectly work as inoculum since its high biomass variety and concentration ensure some kind of sulfide oxidising population that will be favoured under appropriate growing conditions inside the biological reactor. 3.2 SYSTEM PERFORMANCE UNDER DIFFERENT O2/H2S LOADING RATES As it has been previously reported (Buisman et al., 1989; Janssen et al., 1995; Fortuny et al., 2006), biological oxidation of sulfide leads to sulfate or sulfur formation, depending on the oxygen availability. Janssen et al., (1995) showed that probably both reactions can be performed by the same metabolic type of micro-organism and that the oxidation end-product can be selected just by changing the amount of supplied oxygen. In this study, operating at O2/H2S supply ratios from 1.6 to 23.6 (vv-1) allowed obtaining a relationship between this parameter and the percentage of H2S oxidised to sulfate (Fig. 4). This is a useful information to take into account for a reactor scale-up or for other similar systems (as long as equal or very similar O2 gas-liquid mass transfer can be assumed). WWTP and waste management facilities usually deal with biogas effluents with H2S concentrations typically in the range of a few thousands ppmv (Syed et al., 2006) and sulphur production can be an important problem if it is not controlled. With such information, if the reactor’s design allows it, only varying the air flow rate accordingly to an on-line H2S concentration measurement, it may be possible to choose which or what amount of each possible end product should be obtained (SO42- or S0). On the other hand, if elemental sulfur generation is a problem, depending on the O2/H2S supplied ratio it will be possible to estimate the amount of produced sulfur over time and thus anticipate possible operational problems.
258
M. FORTUNY, M. A. DESHUSSES, X. CAMISANS, C. CASAS, J. LAFUENTE AND DAVID GABRIEL
Figure 4. % H2S oxidized as SO42- at different O2/H2S supplied ratios.
Also, unnecessary amounts of supplied air to the biogas flow could be saved, thus avoiding operational adjustments (for the correct burning mixture) to the burning engines and possible explosion risks at very high air supply. 3.3 SYSTEM MAXIMUM ELIMINATION CAPACITY ASSESSMENT After one month operation at a constant LR of 55.6 g H2S m-3 h-1 and EBRT of 180 s, an experiment to evaluate the effect of decreasing EBRTs at a constant inlet concentration (2,000 ppmv) and constant O2/H2S supply ratio (23.6 v v-1) was performed. As shown in Fig. 5a, the system was able to operate with RE above 95 % up to an EBRT of 120 s (LR = 83.5 g H2S m-3 h-1) and with 90 % RE at an EBRT of 90 s. Thus, up to LR of about 80 g H2S m-3 h-1 there is no mass transfer limitation at such operating conditions, which would allow reducing the EBRT almost 60 s without any effect on the system performance. According to the fast performance recovery after applying again an EBRT of 180 s, it can be concluded that only mass transfer limitation was hindering the reactor performance and no significant sulfide accumulation occurred in the liquid phase during the high loaded periods. Also, if sulfide had accumulated, an important drop on the ORP would have been observed and ORP did not drop below values of -50 mV. Thus, at that time, a maximum elimination capacity (EC) of 125.6 g H2S m-3 h-1 was achieved.
HIGH H2S CONCENTRATIONS ABATEMENT IN A BIOTRICKLING FILTER
259
Figure 5. a) RE, ORP, DO and EBRT vs. time. B) EC vs. LR
However, not only biological oxidation contributed to the sulfide oxidation during the high loading periods, since during three days after the experiment (LRT= 54 ± 7 h) small concentrations of thiosulfate (representing less than 1 % of the total amount of degraded H2S) were detected in the liquid phase (results not shown). This probably means that chemical oxidation of sulfide occurred again.
260
M. FORTUNY, M. A. DESHUSSES, X. CAMISANS, C. CASAS, J. LAFUENTE AND DAVID GABRIEL
3.4 CARBON AND SULFUR MASS BALANCES During the whole operation period, a constant ratio of 0.71 ± 0.08 g C-NaHCO3 g-1 S-H2S was supplied since previous research had shown it was necessary to guarantee a C/S supplied ratio of 0.3-0.4 g g-1 in order to avoid carbon limitation. Thus, at constant operating conditions corresponding to: inlet H2S concentration = 2000 ppmv; O2/H2S supplied ratio = 23.6 v v-1 (leading to 90 % hydrogen sulfide oxidized to sulfate) and at pH 6.55 ± 0.05, a carbon mass balance was calculated according to the following expression:
[C ]NaHCO
3
× QC + [CO2 ]G × FGin = [CO2 ]G × FGout + [C ]P × QP
Where: [C ]NaHCO3 : carbonate concentration, mg C L-1;
QC : carbonate inflow, L min-1
[CO2 ]G : gas phase carbon concentration, mg C L
-1
;
FG in : air flow in, L min-1 FG out : total gas flow out, L min-1;
QP : liquid purge flow, L min-1;
[C ]P : total dissolved C in the liquid purge, mg C L
-1
;
According to the balance, a 92 % of the total supplied carbon was detected either as dissolved carbon or CO2. The other 8 %, representing 259 mg C day-1, may be attributed to biomass growth and extracellular polymer substances.
4 CONCLUSIONS A biotrickling filter based system was used to successfully treat up to 2,000 ppmv of H2S (55.6 g H2Sm-3h-1) with steady-state RE over 99 % and complete oxidation to sulfate. Inoculation with MWWTP sludge led to a very fast start-up phase that needed only 3 days to reach RE over 98 %, therefore showing that it is not always needed to obtain a specific culture of sulfide oxidizing bacteria to inoculate an H2S degrading system. The biomass diversity of a MWWTP’s sludge, under favourable and stable growing conditions, among which a pH control may play a major role, can become an optimum and easy-to-obtain inoculum.
HIGH H2S CONCENTRATIONS ABATEMENT IN A BIOTRICKLING FILTER
261
Operation at different O2/H2S supply ratios allowed obtaining a relationship between this parameter and the percentage of degraded sulfide as sulfate, therefore being able to select the process end-product only by varying the amount of air supplied. Finally, an experiment to evaluate the effect of decreasing EBRT at constant inlet concentration showed that the system’s EBRT could be decreased up to 120 s without a notable loss of performance. Lower EBRT (higher LR) led to an important RE drop, which was caused by mass transfer limitation instead of biological limitation. No sulfide accumulation was observed although it was not completely biologically degraded. Chemical oxidation occurred since thiosulfate was afterwards detected.
REFERENCES Bailón, L. (2005) Development of a biotrickling filter for the removal of H2S from biogas. In: (Kennes C. and Veiga M.C., Eds.), Proceedings of the 2005 International Congress on Biotechniques for Air Pollution Control, La Coruña, Spain, pp 143-148. Buisman, C., Post, P., Ijspeert, S., Geraats, G. and Lettinga, G. (1989) Biotechnological process for sulphide removal with sulphur reclamation. Acta Biotechnol., 9: 255-267. Delgado, L., Masana, M., Baeza, M., Gabriel, D. and Alonso, J. (2006) Flow X, Flow Analysis 10th International Conference, September, Porto, Portugal. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for Air Pollution Control. CRC-Lewis Publishers: Boca Raton, FL. Duan, H., Koe, L.C.C., Yan, R. and Chen, X. (2006) Biological treatment of H2S using pellet activated carbon as a carrier of microorganisms in a biofilter. Water Res. 40: 2629-2636. Fortuny, M., Tomàs, M., Baeza, J.A., Deshusses, M.A., Gamisans, X., Casas, C., Gabriel, D. and Lafuente, J. (2006) Performance of a lab-scale bioreactor treating up to 10,000 ppmv of H2S. Proceedings of the 2006 USC-TRG Conference on Biofiltration for Air Pollution Control. October 18-20, Long Beach, California, USA. Gabriel, D. and Deshusses, M.A. (2003) Performance of a full-scale biotrickling filter treating H2S at a gas contact time of 1.6 to 2.2 seconds. Environ. Prog. 22: 111-118. Janssen, A.J.H., Sleyster, R., van der Kaa, C., Jochemsen, A., Bontsema, J. and Lettinga, G. (1995) Biological sulphide oxidation in a fed-batch reactor. Biotechnol. Bioeng. 47: 327-333. Janssen, A.J.H., Meijer, S., Bontsema, J. and Lettinga, G. (1998) Applicaton of the redox potential for controlling a sulfide oxidizing bioreactor. Biotechnol. Bioeng. 60: 147-155. Kuenen, J.G. (1975) Colourless sulphur bacteria and their role in the sulphur cycle. Plant Soil. 43: 49-76. Lens, P and Hulshoff, P. (2000) Environmental technologies to treat sulfur pollution. Principles and Engineering. IWA Publishing, London, UK. (pag. 3) Prado, Ó. J., Veiga, M.C. and Kennes, C. (2005) Treatment of gas-phase methanol in conventional biofilters packed with lava rock. Water Res. 39: 2385-2393.
262
M. FORTUNY, M. A. DESHUSSES, X. CAMISANS, C. CASAS, J. LAFUENTE AND DAVID GABRIEL
Syed, M., Soreanu, G., Falleta, P. and Béland, M. (2006) Removal of hydrogen sulphide from gas streams using biological processes - A review. Canadian Biosystems Engineering. 48: 2.1-2.14. Veiga, M.C. and Kennes, C. (2001) Parameters affecting performance and modelling of biofilters treating alkylbenzene polluted air. Appl. Microbiol. Biotechnol, 55:254-258. Yang, Y; Allen, E.R.(1994) Biofiltration control of hydrogen sulfide.1. Design and operational parameters. J. Air Waste Manage. Assoc. 44: 863-868.
263
Ammonia transformation in a biotrickling air filter LARS PETER NIELSEN, MARIE LOUISE NIELSEN, MATHIAS ANDERSEN, AND ANDERS M. NIELSEN Department of Biological Sciences, University of Aarhus, Ny Munkegade 1540, 8000 Aarhus C, Denmark
ABSTRACT A simple, tubular biotrickling filter was designed for optimal removal of ammonia and odour in ventilation air from a pig house. The removal and transformation of ammonia was studied in detail by analysis and modelling of chemical gradients through the filter. Good correspondence between measurements and model was obtained by using conventional substrate and inhibition kinetics of ammonium and nitrite oxidizing bacteria. Highest rates of ammonia removal were observed in the central section of the filter. Near the air outlet and water inlet the process was ammonia limited, while high nitrous acid concentrations almost excluded any biological activity near the air inlet and water outlet. Nitrous acid inhibition also stabilized pH at 6.5-7 all through the filter. Being sensitive to both ammonia and nitrous acid the nitrite oxidation process occurred mainly in the filter sections near the air outlet / water inlet, and only 8% of the nitrite was turned into nitrate. Water supply only exceeded evaporation by 20% but modelling indicated that additional watering would have limited effect on filter efficiency. The filter was also robust to varying loading, as a 4-fold increase in ammonia inlet concentration only reduced filter efficiency from 86 to 76%.
1 INTRODUCTION Ammonia emitted from animal facilities is a major contributor to acidification and eutrophication of terrestrial and aquatic environments (McCrory and Hobbs, 2001). Significant reductions of both ammonia and door emissions can be accomplished by use of biological trickling filters. The good results, however, are not sufficiently reproducible, and further optimization of design and operation is required for more widespread application.
263
264
L. PETER NIELSEN, M. L. NIELSEN, M. ANDERSEN AND A. M. NIELSEN
Ammonia is a highly water soluble gas (62 M/atm at 20 °C) that is readily protonized to ammonium (NH4+) in water. In biofilters ammonia is oxidized to nitrite (NO2-) by ammonium oxidizing bacteria (AOB) and further to nitrate (NO3-) by nitrite oxidizing bacteria (NOB). The overall removal efficiency is the result of complex interactions of ammonia load, filter surface area, air and water flow, temperature, nitrifier biomasses, and inhibition kinetics. To resolve this we analyzed chemical gradients in a simple counter-current biofilter and simulated the results with a mathematical model. Long term development of nitrifier biomass and the impact of nitrification processes on the removal of organic odorants are two important aspects that will not be addressed in this paper.
2 MATERIALS AND METHODS 2.1 TUBULAR BIOTRICKLING FILTER The first, small version of the tubular biotrickling filter optimized for studies of function is shown in Figure 1 (M. Andersen, DK Patent no. 3108_06). Contaminated air is lead into the lower end of 3 independent tubes being 5.5 meter long and with a cross-section area of 11 cm2 each. The straight airflow gives a good gas-filter contact in relation to pressure drop and minimizes the risk of clogging. The inner surface of the tubes is covered with a thin layer of a fibrous material through which the water slowly percolates driven by gravity. The biofilm developing on the surface of the conductive layer is thus moisturized with fresh water from inside and in direct contact with the air stream outside. This almost eliminates a liquid film diffusion barrier thus promoting the removal of airborne contaminants that are not easily soluble in water. Details on NH3 load and air and water dynamics during the experimental run are given in Table 1. 2.2 ANALYTICAL METHODS Ammonia in the air was sampled in acid solution and analyzed by spectrophotometer (Bower and Holm-Hansen, 1980). Water was sampled with filter paper sticks through the sampling ports, and after dilution NO3- and NO2- was analyzed by HPLC and NH4+ as above. pH was determined by pH sticks. Concentrations of HNO2 and NH3 were calculated from pH and concentrations of NO2- and NH4+ using pKa values of 3.4 and 9.4 respectively. Evaporation was calculated from the concentration gradient of a bromide tracer added to the water supply and analyzed by HPLC.
AMMONIA TRANSFORMATION IN A BIOTRICKLING AIR FILTER
265
Figure 1. The experimental tubular biotrickling filter. The 5 sample points divide the filter in 4 sections of equal length.
266
L. PETER NIELSEN, M. L. NIELSEN, M. ANDERSEN AND A. M. NIELSEN
Table 1. Parameters of the model and the experimental filter. Values apply to one filter tube. The terms AOB and NOB stands for ammonium and nitrite oxidizing bacteria, respectively. Parameter Inlet NH3 concentration Airflow Water flow Air volume per section Water volume per section Evaporation, section 3 Evaporation, section 4 Air/water mass transfer coefficient per section Substrate limitation factors NH3 inhibition factor for AOB Other inhibition factors Km of NH3 for AOB Km of HNO2 for NOB Ki of NH3 for AOB Ki of NH3 for NOB Ki of HNO2 for AOB & NOB AOB capacity NOB capacity Model iteration frequency Model runtime
Value 8.6 ppm 19400 L/h 63 mL/h 4.65 L 23 mL 23 mL/h 30 mL/h 5 mL/s C/(C+Km) e(-C/Ki) 1/(1+(Ci/Ki)2) 786 μM 114 μM 2000 μM 168 μM 1.14 μM 50 x NH3 load 20 x NH3 load 100/s 20 h
2.3 MODEL The mathematical model was programmed in JAVA and followed in most aspects a new, more general biofilter model (Nielsen et al., in preparation). The model filter was divided in 4 sections according to Figure 1 with fully mixed air and water phases and fixed biomass. In each iteration the changes of gaseous NH3, total NH4+, total NO2-, and total NO3- were calculated from air/water NH3 mass transfer, nitrification rates and water and air transport between sections. Subsequently the pH was calculated by solving a charge balance equation using a two-step Newton-Raphson approximation as described by Volcke et al. (2005). Inlet ammonia concentrations, air and water flow and evaporation were set to observed values in the experimental filter (Table 1). Compared to NH3 the concentration of organic compounds in the ventilation air was an order of magnitude lower and transformations of organics were ignored in the present model version. Model parameters, kinetic equations, and initial values are
AMMONIA TRANSFORMATION IN A BIOTRICKLING AIR FILTER
267
Figure 2. Concentration gradients from the water inlet at the top of the filter to the air inlet at the bottom showing NH3 in the air (a) and NO3-, NH4+, NO2-, HNO2, NH3, and pH in the water (b-g). Lines are model simulation results and symbols are measured averages with standard errors (n=3).
268
L. PETER NIELSEN, M. L. NIELSEN, M. ANDERSEN AND A. M. NIELSEN
listed in Table 1. Many different kinetic constants are found in the literature, but the general patterns of the model output were quite robust to variations in kinetic constants. The chosen values were mainly derived from Anthonisen et al. (1976). Notice that the uncharged species NH3 and HNO2 and not NH4+ and NO2- were considered the real species for substrate uptake and inhibition of nitrifiers (Anthonisen et al., 1976). Chemical equilibrium constants for the inorganic nitrogen and carbon species at 20 oC were obtained from table values.
3 RESULTS 3.1 CHEMICAL GRADIENTS Of the water supplied 86% evaporated in the two lower sections (Table 1). Good correspondence between measured and modelled chemical gradients was obtained after proper adjustment of ammonium oxidation capacity and the results are shown together in Figure 2. Ammonia content of the air declined throughout the filter to an outlet concentration of 1 ppm corresponding to 85% overall removal. Most of the removal, 78%, occurred in the two central sections while the upper and lower section only accounted for 21 and 1% respectively. Concentrations of NO2- and NH4+ increased down the filter with final steep increases up to about 300 mM in the evaporation zone (Figure 2c-d). Concentrations of NO3- were much lower and both measured and simulated data showed that virtually all nitrite oxidation occurred in the upper two sections (Figure 2b). The discrepancy between simulated and measured NO3- concentrations could be ascribed to under estimation of the nitrite oxidation capacity. Despite the absence of any chemical buffer or pH regulation, the pH values remained between 6.5 and 7 (Figure 2g). By comparison with the Km and Ki values the concentrations of free NH3 and HNO2 indicated strong substrate limitation of AOB in the upper sections and strong HNO2 inhibition of both AOB and NOB in the lower section (Figure 2e-f, Table 1).
4 DISCUSSION There are many ways to examine the regulation of ammonia transformations in a biological airfilter, and some interesting points can actually be derived without running any models or experiments. One point is that if the maximum capacity of the ammonium oxidizer biomass exceeds the NH3 load, the process will have to be restricted accordingly by NH3 limitation and/or inhibitors. As long as the overall NH3 removal efficiency is good, a high inhibitor level is therefore not an indicator of a critical filter situation but rather
AMMONIA TRANSFORMATION IN A BIOTRICKLING AIR FILTER
269
Figure 3. Modelled NH3 removal efficiency as a function of inlet NH3 concentration (a) and water supply (b). All other parameters are as shown in Table 1.
an indicator of excess nitrifier biomass. In the present case the maximum capacity in the model was set to 50 times the NH3 load, and therefore the kinetics somehow had to reduce ammonium oxidation to less than 2% of the capacity. The results showed that this was accomplished mainly by HNO2 inhibition in the lower half of the filter with concentrations up to 170 times the inhibition constant, and by NH3 limitation in the upper end of the filter with concentrations 15-80 times lower than the Km value (Figure 2e-f, Table 1). The point that inhibitor level and substrate limitation is essentially determined by the biomass/loading ratio has other perhaps counter-intuitive implications; including that inhibitor level must decrease significantly with higher NH3 loading but not with enhanced watering. Model perturbations indeed showed that a 6 times increase in NH3 inlet concentration only reduced filter efficiency from 86 to 76% (Figure 3a). Another general point is that in time only half of the NH3 taken up in a biofilter will be oxidized, while the other half will remain in solution as NH4+ (Smet et al., 2000). This is simply because NH4+ in reality is the only cation available to balance the produced anions NO2- and NO3- when concentrations are up in tens to hundreds of mM N. The measurements confirmed this by perfect charge balances, i.e. 310 mM NH4+ versus 320 mM NO2- + NO3- in the outlet (Figure 2). Another way to address the general point of half-way nitrification is to consider pH: If the nitrifiers tried to generate just 1 mM NO2- + NO3- more in excess of NH4+ the anion excess would have to be balanced by H+, thus implying a pH drop to around 3 which would stop the bacteria long before. In the present filter with significant NO2- accumulation, the immediate effect of lowering pH was the protonization of NO2- to form the highly inhibitory HNO2. In that way kinetics of HNO2 inhibition of ammonium oxidation served as a
270
L. PETER NIELSEN, M. L. NIELSEN, M. ANDERSEN AND A. M. NIELSEN
biological pH buffer ensuring that pH nowhere dropped below 6.5, despite the continuous uptake of a strong base, NH3, and conversion to a moderately strong acid, HNO2 (Figure 2). In other filters with well-established nitrite oxidation and therefore accumulation of NO3- in place of NO2- pH may drop below 5, as the HNO2 block is not operating (L.B. Guldberg, unpublished). In biological airfilters there must be some water run off removing the generated nitrogen salts. Not only in order to sustain NH3 removal but also to avoid emission of the detrimental gasses NO and N2O as observed in poorly watered biofilters (Trimborn et al., 2003). The cost of handling the wastewater, however, makes it relevant to consider how the run off can be minimized. Of the 63 ml supplied every hour to each filter tube 53 ml evaporated in the lower part of the filter and only 10 ml drained off. This means that microorganisms in the upper half of the filter were blessed with more than 6 times the flow of water that eventually was discharged wastewater. The lowest section essentially served as a waste condenser with poor biological conditions due to nitrous acid accumulation. Osmotic stress might have been another important limiting factor at these high salinities (Smet et al., 2000). Model perturbations (Figure 3b) showed that the overall removal efficiency would only increased from 85% to 95% following a doubling of the water supply to 126 mL/h and thereby the generation of about 7 times more wastewater ((63 mL + 10 mL)/10mL). On the other hand a reduction of the water supply by 16% to 53 mL or less would leave no wastewater and the filter would stop working. The major ambition with the modelling studies partly presented here is actually to develop better algorithms for optimization of water supply and biomass management in response to varying ammonia and door loading, air flow, temperature, humidity, etc.
5 ACKNOWLEDGEMENTS The Danish Ministry of Food, Agriculture and Fisheries funded this work.
REFERENCES Anthonisen, A.C., Loehr, R.C., Prakasam, T.B.S. and Srinath, E.G. (1976) Inhibition of nitrification by ammonia and nitrous-acid. J. Wat. Pollut. Cont. Fed. 48(5): 835-852. Bower, C.E. and Holm-Hansen, T. (1980) A salicylate-hypochlorite method for determining ammonia in seawater. Can. J. Fish. Aquat. Sci. 37(5): 794-798.
AMMONIA TRANSFORMATION IN A BIOTRICKLING AIR FILTER
271
Smet, E., Langenhoven, H.V. and Maes, K. (2000) Abatement of high concentrated ammonia loaded waste gases in compost biofilters. Wat. Air Soil Poll. 119: 177-190. Trimborn, M., Goldbach, H., Clemens, J., Cuhls, C. and Breeger, A. (2003) Endbericht zum DBUForschungsvorhaben Reduktion von klimawirksamen Spurengasenin der Abluft von Biofiltern auf Bioabfallbehandlungsanlagen (AZ: 15052). ISBN 3-933865-30-1, Bonner Agrikulturchemische Reihe 14. Volcke, E.I.P., Van Hulle, S., Deksissa, T., Zaher, U. and Vanrolleghem, P.A. (2005) Calculation of pH and concentrations of equilibrium components during dynamic simulation by means of a charge balance. BIOMATH Technical Report. Ghent University, Belgium. pp. 63.
This page intentionally left blank
273
Removal of hydrogen sulfide using upflow and downflow biofilters WONGPUN LIMPASENI
AND
NATTAPOL RATTANAMUK
Department of Environmental Engineering, Chulalongkorn University, Thailand 10330
ABSTRACT The objective of this study was to determine the efficiency of different types of biofilters for hydrogen sulfide removal. The study consisted of two parts. The first part compared the hydrogen sulfide removal efficiency among four different biofilter media comprising compost, bamboo fluff soil, lava rock and activated carbon. The test columns were filled with each of the filter media together with coconut shell fiber to increase void as well as nutrient and microorganism seed. All four biofilters were subjected to empty bed residence time of 45, 60 and 75 seconds and hydrogen sulfide concentration ranged from 50 – 300 ppm. The experiment was run at room temperature (2633 °C), while the filter moisture was controlled between 60-70%. The measured parameters included gas flowrate, inlet and outlet gas concentration, pressure drop, temperature, moisture, pH, microorganism count, organic matter contents and sulfate contents. The results showed that the compost media could achieve the removal efficiency of 100% with gas inlet concentration of 300 ppm when the height of filter was 1 meter and empty bed residence time was 45 seconds. However, other filter media needed height of filter more than 1.25 meter. Compost, bamboo fluff soil, lava rock and activated carbon achieved the maximum elimination capacity of 122, 111, 72 and 108 g/m3-hr, respectively. The compost was found to be the best biofilter medium and was used in the second part of the study to examine the influence of gas flow direction by comparing the removal efficiencies between upflow and downflow biofilters. This study varied the empty bed residence times of 25, 50 and 75 seconds and hydrogen sulfide concentration was fixed at 300 ppm while other conditions remained as before. It was found that both biofilters could still achieve 100% removal efficiency with residence time as low as 25 seconds if the filter depth was increased to 1.5 meter. The downflow biofilter had removal efficiency similar to upflow biofilter, but slightly less when the filter depth was 1 meter and the empty bed residence time was only 25 seconds.
273
274
WONGPUN LIMPASENI
AND
NATTAPOL RATTANAMUK
1 INTRODUCTION Hydrogen sulfide gas is harmful to health. It is highly odorous and highly toxic. The gas can be found in industrial area and wastewater treatment plant. Hydrogen sulfide gas from industry come from industrial processes such as oil refinery, food processing, and pulp industry for example. The concentrations of hydrogen sulfide from the mentioned industrial processes are in the range of 5-70 ppm and can be as high as 300 ppm (Barona et al., 2004). U.S.A. and Thailand limit allowable concentration of hydrogen sulfide gas at 20 ppm and maximum concentration averaging 10 minutes at 50 ppm (OSHA, 2005). Control of hydrogen sulfide gas can be achieved by activated carbon adsorption, ozone oxidation and incineration (Ying et al., 1996). Presently, there is an interest in using biofilter to treat hydrogen sulfide gas. The method has an advantage over physical and chemical treatment because there is little maintenance and low operating cost while yielding high treatment efficiency. Past research works show that the preferred filter media are compost, peat and soil (Hartlikainen et al., 2001; Bohn and Bohn, 1988). It was also observed that lowering retention time led to lower efficiency and lower treatment capacity (Elias and Barona, 2002). Operation of the system over a long period caused a drop in pH of the media and lower treatment efficiency (Oyarzun et al., 2003). The optimum moisture content of the media is around 60%, humidity of air should be higher than 98% and pH of the media should be in the range of 6-8 (Devinny et al., 1999). This research has an objective to study the efficiency of biofilter in removing hydrogen sulfide gas and determine suitable filter media and optimum operating conditions comparing upflow and downflow biofilter.
2 MATERIALS AND METHODS The experiment used a laboratory scale biofilter made from acrylic column diameter 0.054 meter, 2.00 meters high, filled with media 1.50 meter with total volume of 3.43 liters (Figure 1). Other accessory equipment included hydrogen sulfide gas generator and air humidifier. The experiment compared four filter media, which were compost, bamboo fluff, lava rock and activated carbon. Additional media were bio-sludge from municipal wastewater treatment plant served as microorganism seeds, chicken manure served as nutrients, and coconut shell to increase void. The mixing ratio for main filter media : coconut shell : manure : sludge were 60 : 20 : 10 : 10 by volume. The physical properties of the media are shown in Table 1. The experiment consisted of 2 stages. The first experiment employed an upflow filter to compare different filter media and study optimum operating conditions. The
REMOVAL OF HYDROGEN SULFIDE USING UPFLOW AND DOWNFLOW BIOFILTERS
275
second experiment compared an upflow filter to a downflow filter using the most suitable filter media and maximum load (Table 2). The experiments were conducted at room temperature (26-33 °C), control air humidity between 70-80%, and varied hydrogen sulfide concentration between 50-300 ppm and retention time between 2575 seconds. Parameters to be investigated include Empty Bed Residence Time - EBRT, Mass loading, Removal efficiency, and Elimination capacity. List of measuring instruments are shown in Table 3.
Figure 1. Diagram of biofilter apparatus experimental setup.
276
WONGPUN LIMPASENI
AND
NATTAPOL RATTANAMUK
Table 1. Properties of filter media. Parameters Particle diameter d50 (mm) Density (g/cm3) Void (%) Moisture content (%) pH
Compost 0.35 0.60 50.04 29.97 8.12
Bamboo fluff 2.00 1.15 62.57 6.39 7.82
Lava rock 4.35 0.88 59.24 0.71 6.82
Activated Carbon 1.70 0.64 49.98 10.20 9.21
Table 2. Experimental plan. Run
Day
Start – up 1
1 - 20 21 – 35 36 – 50 51 – 65 66 – 80 81 - 95
2
Retention time (sec) 90 75 - 45 75 - 45 75 - 45 75 - 45 75 - 25
Gas flowrate (liter/min) 2.29 2.74 - 4.58 2.74 - 4.58 2.74 - 4.58 2.74 - 4.58 2.74 - 8.23
H2S concentration (ppm) 50 50 100 200 300 300
Table 3. List of measuring instruments. Parameters H2S concentration Pressure drop Relative humidity Gas flowrate pH of media Moisture content of media Temperature of media
Instruments / Model VOCs Analyzer/MiniRAE2000 Manometer/Dwyer Hygrometer/Barigo Rotameter/Dwyer pH meter Moisture meter Thermometer
REMOVAL OF HYDROGEN SULFIDE USING UPFLOW AND DOWNFLOW BIOFILTERS
277
3 RESULTS AND DISCUSSION Comparison of the four types of filter media under variable loading ranging from 3 34 g/m3-hour and retention time between 45-75 seconds showed that for all four filters hydrogen sulfide removal efficiency decreased as the load increased. However, the compost media yielded best efficiency and could still retain 100% efficiency at the media height of 1.00 meter while being subjected to hydrogen sulfide concentration 300 ppm and 45 seconds retention time. On the other hand, other filter media need height of media at least 1.25 meter to achieve the same efficiency (Figure 3). Maximum elimination capacity and Critical loading of each filter can be derived from the relationship between elimination capacity and loading as shown in Figure 4. The compost filter had highest maximum elimination capacity at 122 g/m 3-hour followed by bamboo fluff, activated carbon and lava rock at 72 g/m3-hour (Table 4). The compost filter also had highest critical load at 64 g/m3-hour followed by bamboo fluff, activated carbon and lava rock at 22 g/m3-hour. The second experiment was to determine the influence of gas flow direction using compost filter, which was the best filter media and varied retention time between 25-75 seconds and increase the loading to 20 - 60 g/m3-hour. It was shown that for both filters hydrogen sulfide removal efficiency decreased as the load increased. The downflow filter had removal efficiency similar to upflow filter, but slightly less when the filter depth was only 1.0 meter and the EBRT was only 25 second (Figure 5). However, both filters could still achieve 100% removal efficiency with residence time as low as 25 seconds if the filter depth was increased to 1.50 meter (Figure 6). It can be concluded that the biofilter can be operated using both upflow and downflow configuration and still achieve 100% removal efficiency using a proper loading and adequate filter depth.
278
WONGPUN LIMPASENI
AND
NATTAPOL RATTANAMUK
A. Compost
B. Bamboo Fluff
C. Lava Rock
D. Activated Carbon
Figure 2. Relationship between removal efficiency and loadings for each filter.
REMOVAL OF HYDROGEN SULFIDE USING UPFLOW AND DOWNFLOW BIOFILTERS
A. Compost
B. Bamboo fluff
C. Lava Rock
D. Activated Carbon
Figure 3. Relationship between removal efficiency and height of filters at different EBRT for H2S concentration of 300 ppm.
279
280
WONGPUN LIMPASENI
AND
NATTAPOL RATTANAMUK
A. Compost
B. Bamboo fluff
C. Lava rock
D. Activated carbon
Figure 4. Relationship between maximum elimination capacity and critical loading of the four biofilters.
Table 4. Maximum elimination capacity and critical loading of the four biofilters. Filter Media Compost Bamboo fluff Lava rock Activated carbon
Maximum Elimination Capacity(g/m3-hour) 122 111 72 108
Critical Loading (g/m3-hour) 64 58 22 44
REMOVAL OF HYDROGEN SULFIDE USING UPFLOW AND DOWNFLOW BIOFILTERS
A. Upflow Filter
281
B. Downflow Filter
Figure 5. Relationship between removal efficiency and loadings of upflow and downflow filters.
A. Upflow Filter
B. Downflow Filter
Figure 6. Relationship between removal efficiency and height of filters at different EBRT for upflow and downflow filters for H2S concentration of 300 ppm.
4 CONCLUSIONS 1. Compost was found to be the most suitable filter media among the four media investigated. It can achieve removal efficiency as high as 100% with moderate filter height.
282
WONGPUN LIMPASENI
AND
NATTAPOL RATTANAMUK
2. To completely remove H2S at 300 ppm using residence time of 45 seconds, the compost biofilter needed media height of 1.00 meter only while other filter media including bamboo fluff, lava rock and activated carbon needed media height atleast 1.25 meter. 3. The four filter media which are compost, bamboo fluff, lava rock and activated carbon had maximum elimination capacity of 122, 111, 72 and 10 g/m3-hour, respectively and critical loading of 64, 58, 22 and 44 g/m3-hour, respectively. 4. The downflow filters had similar removal efficiency to the upflow filter but slightly less when the filter depth was 1 meter and EBRT was 25 seconds.
5 ACKNOWLEDGEMENTS Acknowledgements are made to the Graduate School and the Department of Environmental Engineering of Chulalongkorn University.
REFERENCES Barona, A., Elias, A., Arias, I., Ray, S. and Cano., R. (2004) Biofilter response to gradual and sudden variations in operating conditions. Biochemical Engineering Journal. 22: 25-31. Bohn, H. and Bohn, R. (1988) Soil bed weed out air pollutant. Chemical Engineering. 95: 73-76. Devinny, S., Deshusses, A. and Webster, S. (1999) Biofiltration for air pollution control. New York: Lewis Publishers. Elias, A. and Barona, A. (2002) Evalution of a packing material for the biodegradation of H2S and product analysis. Proc. Biochem. 37: 813-820. Hartikainen, T., Ruuskanen, J. and Martikainen., P.J. (2001) Carbon disulfide and hydrogen sulfide removal with a peat biofilter. J. Air & Waste Manage. Assoc. 51: 387-392. Occupational Safety & Health Administration. The appropriate method for assessing hydrogen sulfide peak exposure levels [Online]. 1995. Available from: http://www.osha.gov [2005, June 8] Oyarzun, P., Arancibia, F., Canales, C. and Aroca, G. (2003) Biofiltration of high concentration of hydtogen sulphide using Thiobacillus thioparus. Proc. Biochem. 39: 165-170. Ying, C.C., Chung, H. and Ching, P.T. (1996) Operation optimization of Thiobacillus thioparus CH11 biofilter for hydrogen sulfide removal. J. Biotechnol. 52: 31-38.
283
Proposing a new batch method for assessment of biological activity in H2S degrading biotrickling filters L. OTEGI AND L. LARREA Environmental Engineering Department, CEIT and Tecnun (University of Navarra), Manuel de Lardizábal, 15, 20018 – San Sebastián, Spain
ABSTRACT The proposed batch method consists on measuring the sulfate production rate (SPR) at maximum rate of a set of polyurethane cubes extracted from an ongoing pilot-scale biotrickling filter (BTF) for hydrogen sulfide (H2S) removal. Saturation of the system was achieved by applying high gaseous pollutant concentrations. Under these operational conditions the measured activity is proportional to the concentration of H2S degrading biomass (XSH) present in the system. This method has been used to follow the performance of a pilot-scale BTF under selected conditions. The activity at the inlet zone of the packed bed was found to be between two and three times that measured at the outlet zone. An increase in elimination capacity of 14% corresponded to a very similar average activity increase in the reactor. In addition to this, the method provided a means by which the reactor recovery after a starvation period could be studied.
1 INTRODUCTION The amount of active biomass present in a system greatly influences its buffering capacity in terms of pollutant elimination under varying operational conditions or disruptions. Biotrickling filters operating in wastewater treatment plants are generally exposed to a considerable variety of fluctuating conditions. Among these, natural variations in the waste air composition leading to varying inlet loads (Cox et al., 2002) and interruptions in the plant operation, e.g: breakdown of equipment or electrical failure (Gabriel et al., 2004), are the most common transient-state phenomena BTFs must face. The buffering capacity of these bioreactors has traditionally been assessed
283
284
L. OTEGI AND L. LARREA
by introducing artificial pollutant spikes at the gas inlet and by subjecting the reactor to pollutant starvation (Wani et al., 1998; Chung et al., 2007). Different methods of determining biomass abundance or concentration and its activity in a bioreactor are found in the literature. Changes in the active cell number of microbial species are commonly assessed (Gabriel and Deshusses, 2003; Sercu et al., 2005) using plate counts to measure colony forming units (CFUs). Other authors, however, measure the protein content by the bicinchoninic acid (BCA) method to determine biomass concentration (Kan and Deshusses, 2005; González-Sánchez et al., 2006). For the determination of biofilm activity substrate-induced oxygen uptake rate (OUR) measurements are employed most. In this case, the biofilm is washed out from the packing and transferred to the liquid solution where the decrease in oxygen concentration is measured (Cox and Deshusses, 2002). The objectives of this work were twofold: first of all, to propose a new batch method that enables the biological activity of hydrogen sulfide degrading BTFs to be followed through the measurement of sulfate production rates. Secondly, to demonstrate the suitability of this method of observing the changes in activity of a H2S degrading pilot-scale BTF under the following selected applications: a) along the packed bed height for constant applied load; b) for increasing pollutant loads; c) for elimination capacity recovery after a 15 day starvation period; d) for increasing packed-bed temperature. The information provided by this type of test, that is by the off-line assessment of the H2S degrading biomass concentration, also promises to be very valuable for the calibration of mathematical models for H2S removal in BTFs.
2 MATERIALS AND METHODS 2.1 BASIC PRINCIPLES OF THE PROPOSED METHODOLOGY The proposed method consists on measuring the maximum sulfate production rate of a set of polyurethane foam cubes placed in a batch reactor. First of all it is noted that the main difference between the proposed methodology for activity measurements and those found in the literature is that the biofilm is not washed and remains attached to the packing material. This situation is preferable since the presence of the whole biomass in the batch test is ensured. In addition to this, environmental conditions in the batch reactor are closer to those in the pilot reactor. Secondly, the off-line assessment is carried out under maximum degradation rate conditions; that is, in a completely H2S saturated system where the activity is maximum. In a continuous reactor, however, the actual activity is normally below its maximum. The stoichiometry and kinetics for substrate utilization and biomass growth are shown in Table 1 where CSH (g S-H2S·m-3), CO2 (g O·m-3) and CSO4 (g S-SO42-·m-3)
PROPOSING A NEW BATCH METHOD FOR ASSESSMENT OF BIOLOGICAL ACTIVITY
285
are the hydrogen sulfide, oxygen and sulfate concentrations in the water phase; K is the maximum substrate utilization rate (g S·g COD-1·day-1); and KSH is the halfsaturation constant (g S-H2S·m-3). Table 1. Stoichiometry and kinetics of H2S oxidation. Process/Component
CSH
XSH
CO2
CSO4
Substrate utilization and biomass growth
-1
YSH
2-YSH
1
Kinetics
If kinetics are expressed in terms of biomass growth instead of substrate utilization, the K parameter is substituted by the term YSH·K = μSH, where YSH is the biomass yield coefficient (g COD·g S-1) and μSH is the maximum specific biomass growth rate (day-1). According to the above expression, for values of CSH higher than three to five times KSH, the term CSH/KSH + CSH can be approximated to unity, meaning that kinetics will be saturated. Under these conditions the maximum SPR will be proportional to μSH·XSH. A completely saturated system implies that both the liquid and the biofilm are saturated. The gas to liquid mass flux can be described as qex (Cgas/ HSH·- CSH) where qex is the mass exchange coefficient (m3·day-1) that depends on, among others things, the gas-liquid contact area; Cgas is the H2S concentration in the gas phase; and HSH is Henry’s constant. 2.2 DESCRIPTION OF THE EXPERIMENTAL SET-UP The batch reactor employed for the activity tests consisted of a clear PVC column with an internal diameter of 0.15 m and a 0.29 m bed height (see Figure 1). For each SPR experiment 60 cubes (4-cm polyurethane foam) randomly picked from an ongoing H2S degrading pilot-scale reactor described in detail elsewhere were used (Otegi et al., 2006). This number was considered to be representative enough of the studied location at the pilot plant. The batch reactor was fed through a mixing chamber with a mixture containing air provided by air pumps and pure H2S from a cylinder. A theoretical inlet pollutant concentration of 1640 ppm was continuously applied for two hours. It was operated at an average constant air flow rate of 518 l·h-1 providing an empty bed residence time (EBRT) of 36 seconds. Two litres of recirculation liquid from the same BTF containing an initial sulfate concentration and nutrients were continuously recycled at a constant rate of 42 l·h-1.Unreacted gaseous H2S was absorbed by a sodium hydroxide solution.
286
L. OTEGI AND L. LARREA
Figure 1. Schematic of the batch reactor. 1: PUF cubes; 2: recirculation pump; 3: mixing chamber; 4: NaOH solution
2.3 BATCH METHODOLOGY The maximum SPR of the reactor was determined at different heights of the pilot plant packed-bed; namely at the inlet (INL), intermediate (INT) and outlet (OUT). In order to experimentally verify biological maximum degradation rate conditions in the batch reactor, specific SPR tests were carried out where theoretical inlet H2S concentrations up to 3360 ppm were applied. The possible effect of the gas-liquid contact area on H2S mass transfer was also studied by increasing the recycle liquid flow from 42 to 72 l·h-1. For these activity tests INL and OUT cubes were used. The tests showed identical results for biomass corresponding to the same height independently of the applied concentration (Figure 2) or recycle flow (Figure 3) indicating that for the studied operational conditions the proposed methodology succeeded at working under the desired maximum rate conditions. Abiotic control tests in which PUF cubes containing no biomass were used were also carried out so that any possible non-biological sulfate formation could be disregarded (results not shown). For these, an inlet concentration of 1640 ppm was
PROPOSING A NEW BATCH METHOD FOR ASSESSMENT OF BIOLOGICAL ACTIVITY
287
tested under different temperature conditions. The kinetics for sulfate formation in the absence of biomass were found to be negligible.
Figure 2. Influence of gaseous inlet H2S concentration on SPR methodology (INL and OUT cubes).
Figure 3. Influence of trickling rate on SPR methodology (INL cubes).
The biological activity results presented in this paper were obtained from SPR tests carried out to demonstrate selected specific applications of this methodology. Therefore, care should be taken when comparing rates corresponding to different applications to the overall elimination capacity of the ongoing pilot scale reactor because here the steady-state situation cannot always be ensured.
288
L. OTEGI AND L. LARREA
Water samples from the recirculation liquid were periodically withdrawn and immediately analyzed for pH and conductivity using a regular pH meter and a conductimeter (Crison). For sulfate content measurements a turbidimetric method was employed (APHA, 1995). Ammonia, phosphate and temperature measurements were also carried out at the beginning and end of each test to check nutrient availability and any temperature variations in the recycle liquid.
3 RESULTS AND DISCUSSION 3.1 EFFECT OF REACTOR HEIGHT It is well known that BTF type reactors exhibit biomass distribution patterns along the packed bed height (Kennes and Veiga, 2001; Jin et al., 2005). The usefulness of the proposed methodology was tested for biological activity measurements along the pilot-scale BTF packed-bed height. For that aim, SPR tests were carried out with INL, INT and OUT cubes at a relatively constant packed bed temperature (21±1). The activity value (or the biomass concentration) at the inlet zone of the packed bed was found to be roughly three times of that at the outlet zone as shown in Table 2. Accordingly, and as expected, polyurethane foam cubes extracted from an intermediate reactor height exhibited mid activity values. In order to explain these results, it should be taken into account that the biomass XSH is on the one hand proportional to YSH, Q and ΔS, where Q is the air flow rate (m3·d-1) and ΔS is the eliminated concentration for a selected packed bed volume (g S·m-3). However, at the same time, XSH is inversely proportional to the detachment rate of the biofilm. As explained by Otegi et al. (2006) the ongoing pilot BTF was operated to achieve relatively high H2S outlet concentrations giving rise to a relatively saturated system. For that reason, a lower biomass difference than the one measured experimentally would have been expected between the inlet and outlet zones. The higher difference observed herein is attributed to differences in the detachment rate, this parameter being higher at the outlet zone of the bed where water is trickled over more directly. Table 2. Activity values along the packed bed height of the BTF. Origin of the packing INL INT OUT
SPR (mg S-SO42-·l-1·min-1) 1.6142 1.1481 0.5841
R2 0.97 0.90 0.88
289
PROPOSING A NEW BATCH METHOD FOR ASSESSMENT OF BIOLOGICAL ACTIVITY
3.2 EFFECT OF POLLUTANT LOAD When subjected to a 30% load increase under relatively constant packed bed temperature conditions (21 ± 1ºC), the BTF showed an overall increase in elimination capacity of only 14% (see Table 3). The SPR measurements carried out with INL and OUT cubes revealed an increase in biological activity of 25% and 9%, respectively. From these result it can be stated that the average SPR increase observed along the reactor height (~ 17%) agrees reasonably well with the increase in the whole elimination capacity. This is believed to be a logical result because the maximum activity is proportional to the removed load. Table 3. Effect of load on activity values. Origin of the packing
Applied load (g H2S·m-3·h-1)
EC (g H2S·m-3·h-1)
INL OUT INL OUT
34.8 34.8 45.1 45.9
32.5 32.5 36.7 37.5
SPR (mg S-SO42-· l-1·min-1) 1.1920 0.5985 1.4902 0.6528
R2
0.99 0.99 0.99 0.99
3.3 EFFECT OF A STARVATION PERIOD The effect of a 15-day starvation period on the BTF performance was also tested. During that period no air or pollutant was fed to the reactor and only a low water trickling rate was maintained to prevent the biofilm from drying out. Operation of the reactor was restarted with an applied load of 7.3 g H2S·m-3·h-1 which was gradually increased up to 40.5 g H 2S·m-3·h-1 by changing both the air flow and the inlet concentration (results not shown). As shown in Table 4 the activity of INL cubes was measured one, two and almost six weeks after the starvation period began for a packed bed temperature range of 19 ± 1.5ºC. From the long starvation experiment an important activity loss was expected to happen at the reactor and this fact was confirmed by the low activity value measured in the SPR test carried out at the end of the first week of operation. A roughly two-fold increase in the elimination capacity of the continuous reactor (from 5.4 to 10.4 g H2S · m-3·h-1) was similarly reflected in the batch tests as a double activity value. For the further increase in the overall elimination capacity of the reactor (from 10.4 to 40.5 g H2S ·m-3·h-1) the four-fold activity rise was also found to be almost proportional to the gained elimination capacity.
290
L. OTEGI AND L. LARREA
Table 4. Reactor recovery after a starvation period. Origin of the packing INL INL INL
Time from restart (days) 7 14 40
SPR (mg S-SO42-·l-1·min-1) 0.2359 0.4926 1.9479
R2 0.93 0.98 0.99
3.4 EFFECT OF TEMPERATURE Temperature is known to be an important parameter that affects physical absorption, biological oxidation, and especially microbial growth. Results corresponding to SPR measurements carried out to study the effect of this parameter on biological activity are summarized in Table 5. Cubes corresponding to the INL zone were used and the effect was studied for an 8ºC temperature difference. It was found that at 30ºC, the biological activity was three-fold of that measured at the lower temperature. As stated in the basic principles of this methodology, the measured maximum SPR is not only proportional to XSH but to the product μSH·XSH. Therefore, the observed three-fold rise in activity is attributed to both an increase in the value of μSH and XSH. Table 5. Effect of temperature on biological activity. Origin of the packing INL INL
Packed bed temperature (ºC) 30 22
SPR (mg S-SO42-·l-1·min-1) 5.2120 1.5796
R2 0.98 0.83
4 CONCLUSIONS A new batch method has been proposed for maximum sulfate production rate determination of a set of polyurethane foam cubes taken from a continuous H2S degrading BTF. From the results presented in this paper it can be concluded that the proposed method has proven to be a useful tool in following the H2S degrading biomass concentration of a BTFs. This has been demonstrated through the satisfactory application of the methodology to selected conditions. Results have confirmed, for
PROPOSING A NEW BATCH METHOD FOR ASSESSMENT OF BIOLOGICAL ACTIVITY
291
example, that an activity (or biomass) gradient exists along the studied BTF packed bed. The activity at the inlet zone was found to be two to three times that of the activity measured at the outlet zone. In addition, the recovery of the activity in a reactor subjected to a long starvation period is proportional to the gained elimination capacity as the load is gradually increased.
REFERENCES APHA-AWWA-WEF (1995) Standard Methods for the Examination of Water and Wastewater, 1995. 19th ed., American Public Health Association/American Water Works Association/ Water Environment Federation, Washington DC, USA. Chung, Y-C., Ho, K-L. and Tseng, C-P. (2007) Two-stage biofilter for effective NH3 removal from waste gases containing high concentrations of H2S. J. Air & Waste Manage Assoc. 57: 337-347. Cox, H.H.J. and Deshusses, M.A. (2002) Effect of starvation on the performance and re-acclimation of biotrickling filters for air pollution control. Environ. Sci. Technol. 36(14): 3069-3073. Cox, H.H.J., Deshusses, M.A., Converse, B.M., Schroeder, E.D. and Iranpour, R. (2002) Odor and volatile organic compound treatment by biotrickling filters: pilot-scale studies at Hyperion Treatment Plant. Water Environ. Res. 74(6): 557-563. Gabriel, D. and Deshusses, M.A. (2003) Performance of a full-scale biotrickling filter treating H2S at a gas contact time of 1.6 to 2.2 seconds. Environ. Prog. 22(2): 111-118. Gabriel, D., Strauss, J.M., Sheridan, B.A., Brown, J., Torres, E. and Deshusses, M.A. (2004) Short contact time biotrickling filters for odor treatment: performance of a full-scale reactor at Orange County sanitation district In: Proceedings of the USC-TRG Conference on Biofiltration for Air Pollution Control, October 19-22, Redondo Beach, California, USA. González-Sánchez, A., Revah, S. and Deshusses, M.A. (2006) Deployment of extremophilic alkaliphilic bacteria in biotrickling filters for H2S treatment. In: Proceedings of the USCTRG Conference on Biofiltration for Air Pollution Control, October 18-20, Long Beach, California, USA. Jin, Y., Veiga, M.C. and Kennes, C. (2005) Effects of pH, CO2 and flow pattern on the autotrophic degradation of hydrogen sulfide in a biotrickling filter. Biotechnol. Bioeng. 92 (4): 462-471. Kan, E. and Deshusses, M.A. (2005) Continuous operation of foamed emulsion bioreactors treating toluene vapors. Biotechnol. Bioeng. 92(3): 364-371. Kennes, C. and Veiga, M.C. (2001) Bioreactors for waste gas treatment. Kluwer Academic Publishers, Dordrecht, The Netherlands. Otegi, L., Albizuri, J. and Larrea, L. (2006) Calibration and initial validation of the MCB biofilm model for H2S removal in a biotrickling filter pilot plant. In: Proceedings of the USC-TRG Conference on Biofiltration for Air Pollution Control, October 18-20, Long Beach, California, USA.
292
L. OTEGI AND L. LARREA
Sercu, B., Núñez, D., Van Langenhove, H., Aroca, G. and Verstraete, W. (2005) Operational and microbiological aspects of a bioaugmented two-stage biotrickling filter removing hydrogen sulfide and dimethyl sulfide. Biotechnol. Bioeng. 90(2): 259-269. Wani, A.H., Branion, R.M.R. and Lau, A.K. (1998) Effects of periods of starvation and fluctuating hydrogen sulfide concentration on biofilter dynamics and performance. J. Hazard. Mat. 60: 287-303.
293
Effect of various factors to ammonia biodegradation by two stage biofiltration system SILVIJA STRIKAUSKA1, DZIDRA ZARI 2 ANDREJS
A 2,
OLGA MUTERE2, ULDIS VIESTURS1,2, 3
AND
1
University of Agriculture, Latvia, Jelgava Institute of Microbiology and Biotechnology, University of Latvia, Latvia, Riga 3 Latvian State Institute of Wood Chemistry, Latvia, Riga 2
ABSTRACT An autotrophic ammonia-biodegrading PNNS association was isolated from the biological activated sludge of the fish factory wastewater treatment plant and used in the two-stage biofiltration system with the ammonia load 0.78 g/m3h ensured the total removal efficiency up to 0.69 g/m3h as the result of the denitrification process. Additional investigations were made to study physiological and biochemical properties of individual strains of the PNNS association in order to control their growth under various cultivation conditions with the aim to find out the most optimal conditions for biomass preparation and immobilisation. Individual strains of the association can be revealed and counted because of their different colony morphology using selected medium. Cultivation of individual strains of the PNNS association under aerobic conditions revealed a stimulation effect of (NH4)2SO4 in the concentration range of 0.21 – 4.45 g N/l to their growth. Addition of saccharose, glucose, fructose and/or cabbage leaf extract (CLE) in various combinations to agarized medium resulted in the growth stimulation of individual strains of the PNNS association, i.e. Pseudomonas sp., Nitrosomonas sp., Nitrobacter sp. and Sarcina sp. The whole association was cultivated in the liquid mineral medium with amendments mentioned above. Stimulation of the growth in the presence of CLE and some reducing sugars was observed. The results obtained in these experiments will be used for further optimisation of the two-stage biofiltration system using the PNNS association.
293
294
S. STRIKAUSKA, D. ZARI A, O. MUTERE, U. VIESTURS AND A. BÇRZIÒÐ
Abbreviations CLE AM SA TGA EMB PNNS association CFU
Cabbage Leaf Extract ammonium containing mineral medium Saccharose Agar Tryptone Glucose Yeast Extract Agar Eosin Mehtylene Blue Pseudomonas sp., Nitrosomonas sp., Nitrobacter sp., Sarcina sp. colony forming units.
1 INTRODUCTION Biofiltration is a technology for reducing of odorous emissions, which involves the biochemical capabilities of native or modified biological systems and has some advantages as compared to physical – chemical, burning or mechanical methods. Ammonia is an important component of odorous gases produced in cases of ventilating intensive cattle breeding facilities, manure handling and waste composting (Hong et al., 2005; Kim et al., 2007; Mola et al., 2004; Pagans et al., 2007; Schmidt, 2002). Changes in microbial community structure during biofiltration play a crucial role in air treatment process as a whole (Sakano et al., 1998; Sakano et al., 2002; Steele et al., 2005). Rapid and available methods for monitoring of the total count and separate groups of the association allow to control and manage the nitrification and denitrification processes. These approaches would be useful also for the study on optimisation of cultivation conditions for preliminary preparation and immobilisation of microbial biomass for further successful biofiltration process. In our previous experiments the 294biofiltration technique for the purification of polluted air was developed (Strikauska et al., 1999; Viesturs et al., 2002; Viesturs et al., 2003). However, it is still unclear, what are the most optimal conditions for biomass pre-cultivation and immobilization. The aim of this work was to study the effect of various amendments, in particular cabbage leaf extract and reducing sugars, to the growth of the PNNS association and its individual strains.
2 MATERIALS AND METHODS 2.1 MICROORGANISMS AND CULTIVATION An autotrophic ammonia-biodegrading association was isolated from the biological activated sludge of the fish factory wastewater treatment plant. The
EFFECT OF VARIOUS FACTORS TO AMMONIA BIODEGRADATION
295
composition of the liquid medium AM was (in g.l-1): (NH4)2SO4, 1.0; K2HPO4, 1; NaCl, 2.0; MgSO4.7H2O, 0.5; FeSO4.7H2O, 0.001; and CaCO3, 10 (Strikauska et al., 1999). pH of the medium was 7.7, Eh -40.7 mV. Cabbage leaf extract was prepared from white cabbage leafs. 500 g leafs were washed with top water, boiled at 100 °C for 30 minutes, cooled, afterwards a liquid fraction of prepared extract was filtered with 0.45 μm hydrophilic PTFE membrane filter with 1.0 μm APFB glasfiber prefilter (Millipore) and autoclaved at 0.5 atm. Cultivation of microorganisms in the liquid AM medium was performed in 300ml flasks containing 200 ml liquid medium at +26 °C with periodic agitation in the dark. Samples for analysis were taken after 24h, 72h and 144h of incubation. For determination of the number of colony forming units on agarized medium, AM medium with 1.6 % agar, Saccharose agar, Tryptone Glucose agar, Eosin Methylene Blue agar (Sifin, Germany) were used. Plates with inoculated samples were incubated at +37°C for 48 h and longer for observing the growth of colonies on the different medium. Biochemical characterisation of strains was performed using API®32E (BioMeriéux, France). 2.2 ANALYTICAL METHODS The concentration of carbohydrates was determined by HPLC, Agilent 1100 HPLC system was used for the chromatography work and an Agilent Chemstation software for the data analysis. Carbon and sulphur were measured using the C, S analyzator (ELTRA). Total ammonium was determined according to ISO 5983-2:2005. Concentration of NH3, NO3- and NO2- were determined colorimetrically with Nessler reagent, salicylic acid, and NitriVer® 3 Reagent, correspondingly. pH and RedOx potential were measured by electrode (Hanna pH213). All chemicals used in these experiments were analytical grade.
3 RESULTS AND DISCUSSION 3.1 BIOFILTRATION IN MODIFIED SOLID STATE FERMENTATION SYSTEM Our effort was focused on the development of a biofiltration method ensuring a high ammonia concentration and a limited oxygen environment. Biofiltration process was realized in modified solid-state fermentation system (SSF). The investigations were made at different ammonia concentrations in inlet gas and packing loads. The biodegradation of volatile compounds was investigated in one and two stage systems with inert packing material and chemoautotrophic microorganisms, i.e. the PNNS association. A one-stage biofiltration system with the ammonia load 0.41 g/m3h ensured the biological elimination capacity 0.33 g/m3h due to the nitrification processes. A two-stage system with the ammonia load 0.78 g/m3h ensured increased total removal
296
S. STRIKAUSKA, D. ZARI A, O. MUTERE, U. VIESTURS AND A. BÇRZIÒÐ
efficiency up to 0.69 g/m3h as the result of the denitrification process (Fig. 1). To increase a biodegradation activity of chemoautotrophic microorganisms and biological elimination capacity, the further study was focused on the characterization of microorganisms used for nitrification processes, in order to optimize conditions for culture preparation and immobilization for ammonia biofilter.
Figure 1. Two-stage biofiltration system: 1 - compressor, 2 - vessel for contaminant under degradation; 3 - valve; 4 - measurement of volumetric flow rate; 5 - gas part of the bioreactor; 6 - biofilter; 7 - sampling points; 8 – submerged.
3.2 CHARACTERISATION OF THE PNNS ASSOCIATION BY PLATING METHOD The colony growth on AM, SA, TGA and EMB medium was compared. Among tested agarized medium, SA and TGA were the most sensitive medium to the specific biochemical properties of individual strains resulted in different colony color, size and shape. Although Sarcina sp. was not detected on SA agar, nevertheless this strain was forming the yellow colonies on TGA agar. EMB medium usually is used for cultivation of Gram negative bacteria, however, in our experiments this medium was not appropriate in order to reveal all Gram negative strains of the PNNS association, i.e. Pseudomonas sp., Nitrosomonas sp. and Nitrobacter sp. Growth of microorganisms on the mineral AM medium was slow and therefore, for rapid monitoring, this medium cannot be used. In other cases, e.g. for confirmation of autotrophic growth, this medium is necessary.
EFFECT OF VARIOUS FACTORS TO AMMONIA BIODEGRADATION
297
3.3 CHARACTERISATION OF INDIVIDUAL STRAINS OF THE PNNS ASSOCIATION BY PLATING METHOD
Cultivation of individual strains of the PNNS association under aerobic conditions revealed a stimulation effect of (NH4)2SO4 in the concentration range of 0.21 – 4.45 g N/l to the growth of Sarcina sp., Nitrosomonas sp., and Nitrobacter sp. The effect of CLE and reducing sugars to the growth of isolates was tested on AM medium amended with these compounds in various concentrations and combinations. It is known that the nitrifying bacteria can obtain the carbon for growth from CO2 and the energy and reductant for growth from the oxidation of NH3. Therefore these bacteria are considered an obligate chemolithoautotrophs. Recently, however, it was reported that Nitrosomonas europaea can utilize limited amounts of certain organic compounds, including amino acids, pyruvate, and acetate, although no organic compound has been reported to support the growth of N. europaea. Moreover, it was shown that N. europaea can be grown in CO2-free medium by using fructose and pyruvate as carbon sources and may now be considered a facultative chemolithoorganotroph (Hommes et al., 2003). Besides, the growth of Nitrosomonas sp. and Nitrobacter sp. on glucose was reported (Pan et al., 1972). Thus, in our study, the scheme of the experiment provided various combinations of organic compounds, which are supposed to be as carbon source for the growth of individual strains of the PNNS association. After 72h incubation the colonies of isolates were compared by size. The growth of Sarcina sp. was stimulated by addition of the mixture of CLE, (NH4)2SO4, and saccharose in the concentration-dependent manner. All these compounds added as a single amendment resulted in the growth stimulation in lower extent. In a parallel way, the same compositions of medium with inoculated strains were cultivated under capnophylic conditions to study an effect of CO2 to autotrophic growth of tested isolates. An enhanced concentration of CO2 did not reveal any changes in growth rate as compared to the samples cultivated under aerobic conditions. It is known that ammonia oxidizing bacteria obtain usable energy and reductant solely from ammonia and fix carbon autotrophically (Prosser, 1989). Most probably, the effect of CO2 to colony growth can be visible after the more long incubation. 3.4 TESTING
OF BIOCHEMICAL PROPERTIES OF INDIVIDUAL STRAINS OF THE
PNNS
ASSOCIATION
Strains demonstrated an enhanced growth at enhanced concentrations of ammonia were tested for their biochemical properties, using API test systems ID 32E and API 20 NE. Usually these test systems are used for identification of Gram-negative bacteria. In the PNNS association Sarcina sp. is the Gram-positive bacteria. However, the study of tested strains with substrates, provided by API ID 32 E and API 20 NE, was used for biochemical characterisation of microorganisms, not for identification. Results showed that Sarcina sp. does not reduce nitrates, in turn Nitrosomonas sp.
298
S. STRIKAUSKA, D. ZARI A, O. MUTERE, U. VIESTURS AND A. BÇRZIÒÐ
and Nitrobacter sp. reduce nitrates to nitrites. Formation of N2 was not detected. Urease, α-galactosidase and α-maltosidase was detected only for Nitrosomonas sp. among tested strains, in turn, 5 ketogluconate was detected for Sarcina sp. and Nitrobacter sp. The abilities to hydrolyze urea as a source of ammonia and carbon dioxide and to use the products of ureolysis for modification of the pH in the vicinity of the cell appear to be important ecologically selected traits provided by the urease enzyme (Koper et al., 2004). These biochemical properties will be taken into consideration for further study on conditions optimisation for cultivation and immobilisation of the PNNS association. 3.5 EFFECT
OF THE LIQUID MEDIUM CONTENT TO THE DEVELOPMENT OF THE
PNNS
ASSOCIATION
Inoculation of the PNNS association into the liquid AM medium amended by CLE, (NH4)2SO4 and reducing sugars resulted in the noticeable changes in biomass growth and physically-chemical data among tested medium combinations. The scheme of the experiment is shown in Table 1. Table 1. Addition of various amendments to AM liquid medium, a scheme of the experiment. Variant of medium 1 2 3 4 5 6 7 8 9
AM medium, ml 200 195 190 190 195 195 185 185 185
(NH4)2SO4, 200g/l stock, ml
5
CLE, ml
Saccharose 20% stock, ml
Glucose 20% stock, ml
Fructose, 20% stock, ml
5 10 5 5 5
5 5 5
5 5 5
5 5 5
The total count of microorganisms determined on TGA medium, varied in dependence on the medium composition. Thus, the higher number of colony forming units was detected in the samples containing 5% CLE (Fig. 2, Table 1, and variant No. 3). Regarding the diversity of microbial community developed in the tested medium compositions, Sarcina sp. was detected only in the samples 4, 7 and 9 (Table 1). Colony growth of the tested samples on EMB medium was detected only for the
EFFECT OF VARIOUS FACTORS TO AMMONIA BIODEGRADATION
299
samples 5 and 6. No colony formation was detected on SA medium. The results shown above indicate to an important role of the medium for the total count determination. Besides, an adequate choice of a cultural medium could provide the valuable information related to the composition of the association, i.e. proportions between single strains of the association. Application of this approach in our further investigations could provide the more detailed information regarding an effect of medium composition to the microbial growth. In particular, it is of great importance to design conditions for cultivation, which provide an efficient equilibrium between single strains of the association for its efficient work in biofiltration system.
Figure 2. Effect of liquid AM medium content to the total count of microorganisms cultivated during 144 h at +26 °C.
Figure 3. Effect of liquid AM medium content to NO2- formation during cultivation of the PNNS association.
300
S. STRIKAUSKA, D. ZARI A, O. MUTERE, U. VIESTURS AND A. BÇRZIÒÐ
Figure 4. Changes of pH during incubation of the PNNS association (+26 °C, 144 h).
Figure 5. Changes of RedOx potential during incubation of the PNNS association (+26 °C, 144h).
It was shown that formation of NO2- by growing biomass was occurred in the presence of ammonia as well as CLE (Fig. 3). The nitrites can be formed as the first stage of nitrification process (for the sample 1), or both, as nitrification and nitrate reduction. It is necessary to note, that the AM medium amended with CLE contains additional amount of carbon and nitrogen, which can be utilized by growing biomass. The analysis of CLE (undiluted) showed that this extract contained the total nitrogen in concentration of 5 g/l, carbon - 13.7 g/l, sulphur – 0.38 g/l. The concentration of saccharose, glucose and fructose was found to be in undiluted CLE 1.29 g/l; 13.62 g/
EFFECT OF VARIOUS FACTORS TO AMMONIA BIODEGRADATION
301
l and 10.42 g/l, correspondingly. These data should be taken into account in further experiments. According to the data shown in the Figures 3 and 4, pH and RedOx potential were changed upon cultivation. Thus, the medium pH level in the tested samples varied at the beginning of cultivation in the range of 7.67 – 8.12, after 144 h cultivation - 7.42 – 8.04 (Fig. 4 and 5). The simultaneous growth of microorganisms association leads to the noticeable changes of physico-chemical properties of environment. NO3- was not detected during 144h incubation. No noticeable changes in ammonia concentration in the samples were detected during 144h incubation. These results are quite predictable because it is known that an oxidation of nitrites to nitrates by association occurs after a longer period of incubation, i.e. 10-12 days. The second stage of nitrification as well as denitrification is supposed to be investigated in our further experiments.
4 CONCLUSIONS Assessment of changes in microbial community structure during operation of an ammonia biofilter is of a great importance in the context of process optimisation. The PNNS association used in the two-stage ammonia biofiltration system was studied in the both, fermentation system and batch cultures. Experiments with fermentation system demonstrated a high capacity of the PNNS association in nitrification processes. Batch cultivation showed an important role of various amendments to the growth of microorganisms. In conclusion, further work is needed to understand the relationship between biofilter performance and microbial community dynamics at all stages of the process, i.e. biomass pre-cultivation, immobilisation, adaptation, ammonia oxidation.
5 ACKNOWLEDGEMENTS The present research was funded by the Latvian Council of Science Project No 05.1484 and No 04.1076. We are grateful for HPLC analysis to Rita Scherbaka.
302
S. STRIKAUSKA, D. ZARI A, O. MUTERE, U. VIESTURS AND A. BÇRZIÒÐ
REFERENCES Bergey’s Manual® of Systematic Bacteriology. Ed. J.G. Holt, The Williams and Wilkins Company/ Baltimore, 1980. Hommes, N.G., Sayavedra-Soto, L.A. and Arp, D.J. (2003) Chemolithoorganotrophic growth of Nitrosomonas europaea on fructose. J. Bacteriol. 185(23): 6809-6814. Hong, J.H. and Park, K.J. (2005) Compost biofiltration of ammonia gas from bin composting Biores. Technol: 741-745 Kim, J.H., Rene, E.R. and Park, H.S. (2007) Performance of an immobilized cell biofilter for ammonia removal from contaminated air stream. Chemosphere 68(2): 274-280. Koper, T.E., El-Sheikh, A.F., Norton, J.M. and Klotz, M.G. (2004) Urease-encoding genes in ammonia-oxidizing bacteria. Appl. Environ. Microbiol. 70(4): 2342-2348. Molla, A.H., Fakhru’l-Razi A., Hanafi, M.M. and Zahangir Alam, M. (2004) Optimization of process factors for solid-state bioconversion of domestic wastewater sludge, International Biodeterioration and Biodegradation 53: 49-55. Pagans, E., Font, X. and Sánchez, A. (2007) Adsorption, absorption, and biological degradation of ammonia in different biofilter organic media. Biotechnol. Bioeng. 97(3): 515-525. Pan, P. and Umbert, W.W. (1972) Growth of obligate autotrophic bacteria on glucose in a continuous flow-through apparatus. J. Bacteriol. 109(3): 1149-1155. Prosser, J. I. (1989) Autotrophic nitrification in bacteria. Adv. Microb. Physiol. 30: 125-181. Sakano, Y. and Kerkhof, L. (1998) Assessment of changes in microbial community structure during operation of an ammonia biofilter with molecular tools. Appl. Environ. Microbiol. 64(12): 4877-4882. Sakano, Y., Pickering, K.D., Strom, P.F. and Kerkhof, L.J. (2002) Spatial distribution of total, ammonia-oxidizing, and denitrifying bacteria in biological wastewater treatment reactors for bioregenerative life support. Appl. Environ. Microbiol. 68(5): 2285-2293. Schmidt, D. (2002) Odor, hydrogen sulphide, and ammonia emissions from the composting of caged layer manure - Conference of the American Society of Agricultural Engineering, USA, 10 October 2000, pp. 1-9. Steele, J.A., Ozis, F., Fuhrman, J.A. and Devinny, J.S. (2005) Structure of microbial communities in ethanol biofilters. Chem. Engin. J. 113: 135-143. Strikauska, S., Zarina, S., Berzins, A. and Viesturs, U. (1999) Biodegradation of ammonia by two stage biofiltration system. Environmental Engineering and Policy. 1(3): 175-179. Viesturs, U., Zarina, Dz., Strikauska, S., Berzins, A. and Zilevica, A. (2002) Solid state systems for bioremediation and biodegradation - VI International Symposium on Environmental Biotechnology and IV International Symposium on Cleaner Bioprocesses and Sustainable Development, Mexico, Veracruz, 1.2 pdf (CD). , A. (2003) Emission of odour Viesturs, U., Zari a, Dz., Strikauska, S., Dubova, L. and gases during the composting processes of different organic wastes - European Congress on Biotechnology, Basel, Switzerland, 24-29 August, p. 129.
303
MEMBRANE BIOREACTORS
This page intentionally left blank
305
Removal of dimethyl sulfide in a thermophilic membrane bioreactor MUNKHTSETSEG LUVSANJAMBA, AMIT KUMAR
AND
HERMAN VAN LANGENHOVE
En VOC Research Group, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, B-9000 Gent, Belgium
ABSTRACT In this study, the removal of dimethylsulfide was investigated in a flat composite membrane bioreactor (MBR) at elevated temperature (52°C). A composite membrane with a 2 μm polydimethylsiloxane (PDMS) coating layer and porous polyvinylidene diflouride (PVDF) support layer was used. The effect of variable operating conditions such as empty bed retention time (EBRT), nutrient supply, mass loading rate, temperature on elimination capacity was investigated. A maximum elimination capacity (ECmax) of 54 g m-3 h-1 was obtained at mass loading rate of 64 g m-3 h-1, removal efficiency (RE) 84%. The reactor was sensitive to temporary temperature decrease and recovery was slow compared with the biotrickling filter operated at thermophilic condition.
1 INTRODUCTION Treatment of volatile organic sulfur compounds (VOSCs) such as methanethiol, dimethyl disulfide, dimethyl sulfide (DMS) and carbon disulfide takes special attention in waste gas treatment technologies because they cause odour nuisance due to their very low odor threshold value. The potential sources of VOSCs are waste water treatment plants, rendering and composting processes, Kraft and paper pulp industries (Smet et al., 1998). For VOSC removal, due to the chemical complexity of low concentrated waste gas streams and the high flow rates to be handled, biotechnological techniques and scrubbers are recommended. However, biological waste gas treatment technologies have attracted industrial attention because they do not produce secondary waste streams and have low investment cost. A number of studies have proven the effective VOSCs
305
306
MUNKHTSETSEG LUVSANJAMBA, AMIT KUMAR
AND
HERMAN VAN LANGENHOVE
removal in biofilters, biotrickling filters, and membrane bioreactor (De Bo et al., 2003; Ruokojarvi et al., 2001; Sercu et al., 2005; Smet and Van Langenhove, 1998) at mesophilic conditions. But in practice, some processes such as pulp and paper manufacturing, rendering and composting emit VOSCs at elevated temperature (4075°C) and cooling down the waste gas is necessary prior to the biotreatment (Chan, 2006). Cooling down the humid gas is expensive and also produces secondary waste stream. Therefore the treatment of these compounds at elevated temperature using thermophilic and thermotolerant microorganisms should be considered. For the waste treatment at thermophilic condition, the challenge will arise related with the mass transfer limitation. With increasing temperature, Henry’s coefficient will increase which will result in lower driving force for interphase mass transfer (Cox et al., 2001). In biofilters and biotrickling filters air flows through a packed bed of a carrier material on which microorganisms grow as biofilm. The biofilm is covered with a water layer, forming a barrier between the microorganisms and hydrophobic compounds in the air phase. Therefore for hydrophobic pollutants, mass transfer limitation is highly possible in biofilters and biotrickling filters at elevated temperature. According to Matteau and Ramsay (1999), the lower degradation rate at 60°C was due to the physical limitation rather than biological in toluene degrading biofilter. Dhamwichukorn et al. (2001) also suggested that high temperature affected ineffective mass transfer for α-pinene. But in a membrane bioreactor liquid phase and air phases are separated by a membrane thus allowing the transfer of hydrophobic compounds to the biofilm. Pollutants diffuse through the membrane and subsequently degraded in the biofilm. Therefore membrane bioreactors could be advantageous for treatment of waste gases at thermophilic conditions. Also in a membrane bioreactor it is easy to control operational parameters such as pH and nutrients. Aim of this study was to investigate the possibility of removal of DMS at thermophilic condition (52°C) in a MBR using PDMS/PVDF composite membrane.
2 MATERIAL AND METHODS EXPERIMENTAL SETUP The schematic of the thermophilic MBR is shown in Figure 1. Flat membranes were clamped between two identical Perspex reactor halves. The effective membrane area was 40 cm2, and the volume of each compartment was 8 mL. The commercially available composite membrane provided by GKSS Research Centre Geesthacht (Germany), consisted of a 2 μm PDMS layer coated on PVDF was used. Through one compartment mineral medium was recirculated by a Heidolph peristaltic pump (PD5006, Heidolph Instruments GmbH & Co., Schwabach, Germany). Through the other compartment the air was passed in countercurrent along the porous support
REMOVAL OF DIMETHYL SULFIDE IN A THERMOPHILIC MEMBRANE BIOREACTOR
307
layer of the composite membrane. The membrane contactor was placed in a thermostat at 52°C to maintain the constant temperature. The temperature of the inlet gas and liquid entering the reactor was 52°C. Both gas and liquid phases were operated at atmospheric pressure. Dimethyl sulfide was continuously dosed to the inlet air stream using a dynamic vapor-generating system (Smet et al., 1993). The gas sampling ports were inserted before and after the reactor. The inlet air flow was provided at 14, 19.2, 40, 50 and 60 ml min-1, resulting in an EBRT of 36, 25, 12, 10 and 8 s. The air flow was controlled with a calibrated mass flow controller (Model 5850S, Brooks Instrument Division, Emerson Electric Co., Veenendaal, The Netherlands). The reactor was inoculated by recirculating pre-enriched culture (TSS = 4 g L-1) along the PDMS layer of the membrane. When the reactor liquid was clear and biofilm development on the membrane was observed, the microbial suspension was drained and replaced by fresh nutrient solution. The nutrient solution was placed in a water bath (53°C). Periodically, the nutrient solution was drained completely and replaced by a fresh nutrient solution. A mineral medium containing 3.0 g L-1 K2HPO4, 3.0 g L-1 KH2PO4, 3.0 g L-1 NH4Cl, 0.5 g L-1 MgSO4.7H2O, and 0.01 g L-1 FeSO4.7H2O was recirculated. The pH of the liquid media was controlled daily and readjusted to 7 with 1 M NaOH when it reached a value lower than 6.
Figure 1. Schematic of flat membrane bioreactor.
308
MUNKHTSETSEG LUVSANJAMBA, AMIT KUMAR
AND
HERMAN VAN LANGENHOVE
ANALYTICAL METHODS The concentration of DMS was determined by injection of 1 mL gas samples into a GC Agilent 4890D (Hewlett Packard Inc., Agilent Technologies Inc., USA) equipped with a flame ionization detector (FID), using a 15 m HP-5 column (internal diameter 0.53 mm; film thickness 1.5 μm) with helium as carrier gas. The pH of the recirculating liquid media was measured with an electronic pH sensor (Jenway Ltd., Essex, England 3310).
3 RESULTS AND DISCUSSION The membrane bioreactor was operated for four months at 52°C. The reactor was started-up at inlet DMS concentration of 70 ± 10 g m-3 h-1 and EBRT of 10 s. As can be seen in Figure 2, the removal efficiency was low (< 20%) for the first 12 days. Therefore on day 13, the EBRT was increased to 25 s. As soon as the EBRT was increased, immediate improvement was observed and RE increased from 16 to 45%. When the loading rate was decreased stepwise down to 20 g m-3 h-1, the RE stayed at 45% (EC ∼10 g m-3 h-1). When the EBRT was further increased to 36 s on day 19, no change was observed in RE. Instead the concentration of DMS in the head space in the bottle containing the liquid medium increased (results not shown). It showed that at high EBRT of 36 s, DMS was only transported to the liquid medium but not degraded. After increasing the EBRT back to 25 s, decreased RE of 30% was observed. This change could be caused by accumulation of DMS in liquid medium at high EBRT. On day 24, the microbial suspension was added to the MBR, and RE increased from 30 to 50% in 2 days and further to >90% in 9 days. Removal efficiencies exceeding 90% were obtained on day 33 at mass loading rate of 18 g m-3 h-1 and EBRT of 25 s. Shorter start-up period of 9 days was observed in a MBR inoculated with Hyphomicrobium VS for DMS at ambient temperature (De Bo et al., 2003). Similar start-up period of 25 days was observed in a thermophilic biotrickling filter treating DMS (Luvsanjamba et al., 2006). When the reactor reached the steady state, EC max was determined in a thermophilic MBR (Figure 3). From day 68, the mass loading rate was increased stepwise up to 70 g m-3 h-1 at EBRT of 25 s. ECmax of 54 g m-3 h-1 was observed in MBR at mass loading rate of 64 g m-3 h-1 (RE 84%). De Bo et al. (2003) reported ECmax of 100 g m-3 h-1 (EBRT = 24 s) for DMS in a MBR inoculated with Hyphomicrobium VS using PDMS/PVDF membrane. The lower EC obtained in this study compared with the latter was most probably due to the lower DMS degrading capacity of the biofilm established in thermophilic MBR. But compared with thermophilic biotrickling filter removing DMS (ECmax = 45 g m-3 h-1), somewhat higher EC of 54 g m-3 h-1 was obtained in MBR (Luvsanjamba et al., 2006). Also shorter EBRT of 25 s was required to obtain
REMOVAL OF DIMETHYL SULFIDE IN A THERMOPHILIC MEMBRANE BIOREACTOR
309
Figure 2. Start-up period of the MBR at 52°C. The dashed line indicates the change in EBRT from 10 to 25s.
Figure 3. Elimination capacity for DMS as a function of loading rate in a thermophilic MBR.
310
MUNKHTSETSEG LUVSANJAMBA, AMIT KUMAR
AND
HERMAN VAN LANGENHOVE
this EC compared with the thermophilic biotrickling filter (EBRT = 200 s). This difference could be attributed to enhanced mass transfer rate in the MBR compared with a biotrickling filter. After determining the maximum EC in the thermophilic MBR, effect of variable operating conditions such as mass loading rate, EBRT, temperature, and nutrient supply on reactor performance was investigated. Firstly the mass loading rate was increased from 20 to 60 g m-3 h-1. At the same EBRT of 25 s, RE decreased from > 95% to 70 - 80%. When longer EBRT of 36 s was tested at the loading rate of 60 g m-3 h-1, the RE increased from 80 to 88%. However this RE did not remain constant and decreased to 80% next days. The concentration of DMS in the head space of the liquid bottle increased, proving that DMS was accumulated instead of being degraded. Thereafter the EBRT was decreased from 25 to 12 and 8 s at a constant mass loading rate of 60 ±3 g m-3 h-1. As indicated in Figure 4, the RE decreased from 76 to 56 and 40%, respectively. It shows that the reactor performance is greatly influenced by EBRT. Thirdly the effect of temperature decrease to mesophilic range on EC was investigated. The temperature of the isothermal chamber was kept at 25°C for 3 days. The EC dropped from 26 g m-3 h-1 (RE = 100%) to 15 g m-3 h-1 (RE = 51%). It took 11 days to reach the RE 95% at loading rate of 25 g m-3 h-1 when the temperature was set back to 52°C. In contrast, it took only 2 hours in a DMS degrading thermophilic biotrickling filter inoculated with activated sludge (Luvsanjamba et al., 2006).
Figure 4. Effect of the EBRT on removal efficiency of the MBR.
REMOVAL OF DIMETHYL SULFIDE IN A THERMOPHILIC MEMBRANE BIOREACTOR
311
Different nutrient supplies such as vitamin, trace elements and tap water were tested in a recirculating liquid medium to examine the effect of medium composition on EC of the reactor. However no significant effect was observed on the elimination capacity. From the third month of an operation the elimination capacity of the reactor decreased to 30 g m-3 h-1 (RE = 50% at an EBRT of 25 s). The decreased EC could be influenced by ageing of the biofilm as well as formation of thick biofilm layer on the membrane lowering the transportation of the pollutant.
4 CONCLUSIONS This study demonstrated the possibility of removal of DMS in membrane bioreactor at thermophilic condition. Higher elimination capacity (54 g m-3 h-1) was obtained in MBR in comparison with biotrickling filter operated at 52°C (45 g m-3 h-1). The optimum EBRT was found to be 25 s. Since the declined performance of was observed after an operation of three months, the long-term efficiency of thermophilic MBR should be further investigated.
REFERENCES Chan, A.A. (2006) Attempted biofiltration of reduced sulphur compounds from a pulp and paper mill in northern Sweden. Environ. Prog. 25(2): 152-160. Cox, H.H.J., Sexton, T., Shareefdeen, Z.M. and Deshusses, M.A. (2001) Thermophilic biotrickling filtration of ethanol vapors. Environ. Sci. Technol. 35(12): 2612-2619. De Bo, J., Heyman, J., Vincke, J., Verstraete, W. and Van Langenhove, H. (2003) Dimethyl sulfide removal from synthetic waste gas using a flat poly(dimethylsiloxane)-coated composite membrane bioreactor. Environ. Sci. Technol. 37(18): 4228-4234. Dhamwichukorn, S., Kleinheinz, G.T. and Bagley, S.T. (2001) Thermophilic biofiltration of methanol and alpha-pinene. J. Ind. Microbiol. Biotechnol. 26(3): 127-133. Luvsanjamba, M., Van Peteghem, J., Sercu, B. and Van Langenhove, H. (2006). Thermophilic biofiltration of dimethyl sulfide loaded waste gas. Paper presented at the USC – TRG Conference on Biofiltration, October 20-22, 2006, Long Beach, CA, USA. Matteau, Y. and Ramsay, B. (1999) Thermophilic toluene biofiltration. J. Air. Waste. Manage. 49(3): 350-354. Ruokojarvi, A., Ruuskanen, J., Martikainen, P.J. and Olkkonen, M. (2001) Oxidation of gas mixtures containing dimethyl sulfide, hydrogen sulfide, and methanethiol using a two-stage biotrickling filter. J. Air. Waste. Manage. 51(1): 11-16.
312
MUNKHTSETSEG LUVSANJAMBA, AMIT KUMAR
AND
HERMAN VAN LANGENHOVE
Sercu, B., Nunez, D., Aroca, G., Boon, N., Verstraete, W. and Van Langenhove, H. (2005) Inoculation and start-up of a biotricking filter removing dimethyl sulfide. Chem. Eng. J. 113(2-3): 127-134. Smet, E., Keymeulen, R. and Van Langenhove, H. (1993) Dynamic vapour generating system: practicability and environmental application. in: Proceedings Environmental Platform, Leuven, Belgium, p. 121-141. Smet, E., Lens, P. and Van Langenhove, H. (1998) Treatment of waste gases contaminated with odorous sulfur compounds. Crit. Rev. Env. Sci. Tec. 28(1): 89-117. Smet, E. and Van Langenhove, H. (1998) Abatement of volatile organic sulfur compounds in odorous emissions from the bio-industry. Biodegradation. 9(3-4): 273-284.
313
Biological waste gas purification using membranes: Opportunities and challenges N.J.R. KRAAKMAN1, N. VAN RAS2, D. LLEWELLYN3, D. STARMANS4 AND P. REBEYRE5 1
Bioway international b.v., P.O. Box 361, 6710 BJ Ede, The Netherlands (
[email protected]) Bioclear b.v., P.O. Box 2262, 9704 CG, Groningen, The Netherlands 3 University of Guelph, 50 Stone Road East, N1G 2W1, Guelph, Canada 4 Wageningen University & Research Center, P.O. Box 17, 6700 AA Wageningen, The Netherlands 5 European Space Agency/ESTEC, P.O. Box 299, 2200 AG Noordwijk, The Netherlands 2
1 INTRODUCTION Biotechnology to purify air and waste gasses has been applied more frequently in recent years, because they eliminate many of the drawbacks of classical physicalchemical techniques (Kennes and Veiga 2001; Shareefdeen and Singh, 2005). The disadvantages of the traditional air-treatment techniques are high-energy costs (incinerators), the use of chemicals (chemical scrubbers) and the production of waste products (incinerators, scrubbers, activated carbon filters). Biological waste gas purification is a ‘green’ technology that requires only minimal energy inputs and produces little to no waste. Biological waste gas purification is safe as it is operated at ambient temperatures without the requirement of storage and handling of chemicals and is often applied because of its low operational costs. However, biological waste gas purification is still relatively new in many application fields. For example, applications of biological waste gas treatment for high contaminant concentrations are still scare at this moment. The main reason is that biological treatment systems face operational limitations for the treatment of high contaminant concentrations as it may lead to biomass clogging inside the biological treatment system. Biomass clogging will result in poor airflow distribution, high pressure drop over the system and unstable operation, that eventually will lead to a reduction of performance. This severely limits the applicability of conventional biofiltration systems for the treatment of airstreams with high contaminant
313
314
N. J. R. KRAAKMAN, N. VAN RAS, D. LLEWELLYN, D. STARMANS AND P. REBEYRE
concentrations. Traditional technologies like incineration are commonly used to treat waste gases with high concentrations of contaminants. This is attractive when the concentrations are high enough to thermodynamically support burning the contaminants without additional fuel. If this is not the case, incineration is relatively expensive because of the additional fuel requirement. The cost of operating an incinerator is then directly related to the energy prices, which have been increasing significantly over the last couple of years. Conventional biofiltration systems can have limitations in terms of process control, and often higher degradation capacities or smaller footprints are required. Bioreactors using membranes are of interest for new applications as they have some important advantages over conventional biofiltration systems. The first advantage is that any biomass accumulation does not interfere with the gas phase and that biomass accumulation can be better controlled. Secondly, it eliminates the risk of unintentionally drying out of the biofilm and ensures a sustainable control of moisture, which is often a problem in the operation of conventional biofilters. Another advantage is the improved homogenous airflow distribution as well as better controlled nutrient addition to the biofilm. Finally, poor water-soluble compounds with high membrane permeability can be treated effectively with membrane bioreactors as has been demonstrated for hexane (Reiser, 1994). The membrane separates the gas phase from the liquid phase holding the biology, which improves the possibilities to control and optimize the biological process. This paper describes reactor design considerations of membrane bioreactor for waste gas treatment. Current limitations and challenges for further development of applications are discussed including some possible interesting application fields.
2 PROCESS MECHANISM OF A MEMBRANE BIOREACTOR In a membrane bioreactor for gas treatment, the membrane is the interface between the contaminated gas phase and a liquid phase containing nutrients. The pollutants diffuse from the waste gas through the membrane to the biofilm that is attached on the membrane at the liquid membrane interface. The micro-organisms in the biofilm will obtain oxygen from the gas phase, while the nutrients are obtained from the liquid phase. The liquid with the nutrients is usually recirculated, buffered to sustain a suitable pH and refreshed occasionally to add nutrients or remove degradation products.
BIOLOGICAL WASTE GAS PURIFICATION USING MEMBRANES
315
Figure 1. Process flows in a membrane bioreactor.
A membrane bioreactor will have an overall mass transfer coefficient Kov (m s-1) that is defined by the mass transfer coefficient on the gas phase side of the membrane kg, the mass transfer coefficient for the membrane km, and the mass transfer in the liquid phase kl as written as 1/Kov = 1/kg + 1/km + 1/kl
(1)
The mass of pollutants transferred through the membrane, the flux J (g m-2 s-1), is an important design parameter for membrane bioreactors and is defined as the overall mass transfer multiplied by the concentration gradient dC (g m-3). J = Kov * dC
(2)
The mass transfer coefficient for pollutants in the membrane km is defined as the ratio of the permeability of the pollutant P (m2 s-1) in the membrane and the membrane thickness d (m). Permeability has been described by Solubility of the air/ membrane partition coefficient S (g m-3membrane / g m-3air) multiplied by the diffusion coefficient Dm through the membrane material (Mulder, 1996). Km = P / d = (S * Dm) / d
(3)
There are two basic types of membrane that are used in biological waste gas purification: micro-porous membranes and dense-phase membranes. High membrane permeability for the pollutants is an important factor in choosing a membrane type or membrane material for a specific application.
316
N. J. R. KRAAKMAN, N. VAN RAS, D. LLEWELLYN, D. STARMANS AND P. REBEYRE
3 DESIGN CONSIDERATIONS 3.1 INTRODUCTION Membrane bioreactors for waste gas treatment are relatively new. Suitable membrane material and reactor configuration are essential in the further development and successful application of membrane bioreactors for waste gas purification. The relatively high cost of the membrane materials, and the lack of experience with the different types of membranes and possible reactor configurations on a wide range of pollutants have been an important limitation to the advancement of membrane bioreactors. Membrane biotechnology has evolved over the last couple of years in other areas like industrial and municipal wastewater treatment and creates new possibilities for biological waste gas treatment. 3.2 TYPES OF MEMBRANES There are basically two membrane types: micro-porous membranes and densephase membranes. Micro-porous membranes have a porous structure with a porosity of up to 30-85% (Hartmans et al., 1992). The pollutants can cross the membrane by diffusing through the gas-filled pores, yet the pores are small enough to prevent microorganisms to pass the membrane. The membrane material is often chosen for having hydrophobic properties, so that at relatively low trans-membrane pressures the risk of water penetration is reduced. Dense-phase membranes have no macroscopic pores meaning that the pollutant has to diffuse through the membrane material. This imparts some potential for contaminant selectivity when choosing a type of dense-phase membrane. It is critical that a membrane material with a high gas diffusion coefficient should be used for the specific contaminants to minimise the mass transfer resistance of the dense-phase membrane applied. In theory, micro-porous membranes have significantly lower mass transfer resistance than dense-phase membranes. Micro-porous membranes have higher permeability and a poor to no selectivity in permeation compared to dense-phase membranes. However, at high air pressures, systems deploying micro-porous membranes run the risk of trans-membrane gas flow, which may compromise the integrity of the membrane. In addition, micro-porous membranes are not a complete definite barrier for micro-organisms, which could be important in certain applications. Furthermore, micro-porous membranes are subject to fouling due to blocking of the micro-pores, leading to a decline in performance over time (van Reij, 1996; de Bo et al., 2003). Dense-phase membranes seem, in principal, more suitable than microporous membranes for long term sustainable operation. Since the contaminant may actually dissolve into the membrane material, the polymer phase of dense-phase membranes might also act as a buffering medium for fluctuating inlet pollutant loads.
BIOLOGICAL WASTE GAS PURIFICATION USING MEMBRANES
317
To combine the best characteristics of both types of membranes, composite membranes made of micro-porous membrane supports coated with a dense-phase polymer layer have been studied (de Bo et al., 2003) and seem very promising. The very thin layer of dense-phase membrane is located on the liquid side. The supporting porous layer is located on the gas-phase side. The composite membranes will be inherently more complex and more expensive to manufacture than the dense-phase or the micro-porous membranes. The choice of membrane material depends on the pollutants to be treated. Different membrane materials have been studied for gas treatment as has been illustrated by Fitch (Fitch, 2005). Examples of membrane materials are poly(butadiene), latex, poly(vinylalcohol), poly(sulfone), poly(styrenesulfone), poly(amide), poly(ethylene), poly(tetrafluoethene), poly(propylene) and poly(dimethylsiloxane). Different membranes are commercially available as they have been successfully used in areas such as the medical and the food processing fields. The polymer selection for the dense-phase membrane material needs to have a high solubility for the pollutant. Otherwise, mass transfer resistance across the membrane will hinder the rate of biodegradation. Poly(dimethylsiloxane) (PDMS) dense-phase membranes are relatively permeable and seems to be relatively unselective towards pollutants (Merkel et al., 2000; De Bo et al., 2003). The risk of dense phase membrane failures due to dissolution or swelling of polymer when solubility is extremely high should be avoided. The effect has been demonstrated in membrane bioreactors for benzene removal in using a latex membrane (Fitch et al., 2003). Parameters that counteract these side-effects are the degree of cross linking and the molecular weight of the membrane polymer used. 3.3 MEMBRANE THICKNESS Micro-porous and dense-phase membranes are made in a variety of different thicknesses, but most membranes have a thickness of around 150-800 um. The membrane thickness is preferably as small as possible, but needs to have a certain thickness for mechanical stability. When composite membranes are used, the microporous membrane is used only for support of the dense-phase layer. The dense-phase layer can be kept very thin (< 10 um) and is usually located on only one side of the microporous membrane. 3.4 MEMBRANE AND REACTOR CONFIGURATION The reactor configuration is dictated by the membrane configuration. Membrane configurations are possible in two basic shapes: tubular or flat. The choice of membrane configuration can be based on optimizing biomass removal, but is most likely based on optimizing mass-transfer. Mass-transfer is a transport phenomenon that is similar to heat transfer in heat exchangers. In general the tubular configuration is most likely
318
N. J. R. KRAAKMAN, N. VAN RAS, D. LLEWELLYN, D. STARMANS AND P. REBEYRE
the most optimal configuration for a membrane bioreactor because of a higher surface area to volume ratio. However, the cost of the various membrane materials, their shape and their availability currently determines choices of membrane and reactor configurations. 3.5 AIRFLOW DYNAMICS Air flow dynamics of a membrane bioreactor are important for optimal treatment as well as energy requirements during operations. The energy requirement is directly related to pressure drop over the bioreactor system. Low pressure drop values should be an important objective when designing a membrane bioreactor for waste gas treatment. Figure 2 shows an example of the pressure drop as a function of the inner diameter of tubular membranes. Larger diameters are associated with lower high pressure drop. However larger diameters also reduce the total surface area of the membrane, which limits the total mass-transfer and the size of the active biofilm.
Figure 2. The pressure drop versus internal diameters of a single hollow fiber PDMS membrane at gas flow rates of 4, 6, 8 and 10 liters per minute.
3.6 WATER FLOW Water with nutrients is usually recirculated, buffered to sustain a neutral pH and refreshed occasionally to add nutrients or remove degradation products. The direction of the water flow can be current or counter-current to the gas flow direction.
BIOLOGICAL WASTE GAS PURIFICATION USING MEMBRANES
319
The water flow recirculation rate is an important parameter in order to maintain minimal differences in pH or nutrients concentration in the membrane bioreactor. The water flow rate is especially of importance for the control of the amount of biomass in the bioreactor that might accumulate at high inlet pollutant loads. Water flow rate will generation shear forces that can discharge inactive biomass. 3.7 BIOMASS CONTROL The membrane separates the gas phase from the biofilm layer on the liquid side, which improves the possibilities to control the biological process. The most active micro-organisms of the biofilm layer are located directly adjacent to the membrane surface, which makes it possible to easily discharge inactive biomass from the reactor. Water flow can generate shear forces on the biofilm layer that erodes only the top layer. An increase of water flow rates will enhance this process called biomass sloughing. The most active zone in the biofilm is likely limited to 0.2 mm or less. Studer showed that a biofilm of less than 1 mm could easily be maintained at relatively low shear forces (Studer, 2005). An alternative method of removing biomass is the intermittent use of aerators. Large gas bubbles will generate shear forces that removes excess biomass. This has already become common practice over the last years in full-scale membrane bioreactors for industrial and municipal wastewater treatment.
4 APPLICATION FIELDS An application field of increasing interest for newly developed biological gas treatment technologies is the treatment of waste gases from chemical industries that emit waste gasses with relatively high pollutant concentrations. At what pollutants concentration do conventional biotechniques like biofilters or biotrickling filters starts to have difficulties treating waste gas, especially in relation to control the biomass growth? Biomass accumulation in conventional biofilters and biotrickling filters relates to the total pollutant load per reactor volume (Ozis, 2005). For most pollutants, an average pollutant loading higher than 50 g/m3 reactor volume per hour will most likely sooner or later lead to biomass accumulation in a conventional biofilter. Up to what concentration in the waste gas is the conventional technique incineration not self-supporting and does it require input of fuel for the incineration process? Calculations show that waste gas with a concentration range between approximately 0.4 – 4 g/m3 is interesting for most contaminants for future applications of newly developed biological waste gas treatment systems (Bioway, 2007). Other specific fields of application are the printing and paint industries, where relatively poorly water-soluble solvents are used. That normally leads to high treatment
320
N. J. R. KRAAKMAN, N. VAN RAS, D. LLEWELLYN, D. STARMANS AND P. REBEYRE
costs as these compounds are difficult to remove by conventional waste gas treatment techniques. Relatively small facilities often face high capital investments for incinerators that often can not operate under economically feasible circumstances. Biological treatment is only possible when temperature is in a suitable range for the micro-organisms that are used. As the gas phase is separated from the biomass at the liquid side of the membrane, membrane bioreactors can also be applied for the treatment of relatively hot off-gasses. The recirculating liquid can be used for heat exchange to keep the biofilm at the required temperature range for optimal biological degradation. Not only airstreams with relatively high temperatures or high concentrations are suitable for further development of membrane bioreactors. Also airstreams with very low concentrations, such as indoor air, form an interesting application field (Llewellyn and Dixon, 2006) as for example an increase in air humidity (often unwanted in the case of indoor environments) can be avoided using membranes technology. As the air stream is completely separated from the biomass by the membrane, the risk of airborne particles possibly containing unwanted microbes from the biological treatment system is also eliminated. The application of membrane biofilters for air quality control in space crafts (van Ras et al., 2006) as well as a variety of occupied terrestrial environments is currently being investigated. Long term space missions require a sustainable air purification technology that uses a minimum energy demand and does not generate any waste products, which membrane biotechnology can provide.
5 CONCLUSIONS Bioreactors using membranes are of interest for new applications as they have some important advantages over conventional waste gas treatment systems, including conventional biological waste gas treatment systems. Challenges are still present as an extra resistance for mass transfer from the gas phase to the biofilm might be introduced using a membrane. Thinner and cheaper membranes are required. Research and development have a current focus on the design of the systems that require input of basic parameters like permeability of pollutants for the membrane material and optimal reactor configurations.
BIOLOGICAL WASTE GAS PURIFICATION USING MEMBRANES
321
REFERENCES de Bo, I., Van Langenhove, H., Pruuost, P., De Neve, J., Pieters, J., Vankelecom, I.F.J. and Dick, E. (2003) Investigation of permeability and selectivity of gases and volatile organic compounds for polydimethylsiloxane membranes. J. Membrane Sci. 215: 303-319. Bioway. (2007) Internal study. Fitch, M.W. (2005) Membrane bioreactor biotechnology. In (Shareefdeen, Z. and Singh, A., Eds.), Biotechnology for Odour and Air Pollution Control, Springer-Verlag, Heidelberg, Germany, pp. 195-212. Fitch, M.W., Neeman, J. and England, E. (2003) Mass transfer and benzene removal from air using latex rubber tubing and a hollow-fiber membrane module. Appl. Biochem. Biotechnol. 104: 199-214. Hartmans, S., Leenen, E.J.T.M. and Viskuilen, G.T.H. (1992) Membrane bioreactors with porous hydrophobic membranes for waste-gas treatment. In (Dragt A.J., van Ham, J., Eds.), Biotechniques for air pollution abatement and odour control policies, Elsevier, Amsterdam, pp. 103-106. Kennes, C. and Veiga., M.C. (2001) Bioreactors for Waste Gas Treatment. Kluwer Acadamic Publishers, Dordrecht, The Netherlands. Llewellyn, D. and Dixon, M. (2006) A botanical-membrane hybrid for the biofiltration of indoor air. In: Proceedings of the USC-TRG Conference on Biofiltration for Air Pollution Control, October 18 to 20, Long Beach, California, pp. 75-84. Merkel, T.C., Bondar, V.I., Nagai, K., Freeman, B.D. and Pinnau, I. (2000) Gas sorption, diffusion and permeation properties of poly(dimethylsiloxane) films. J. Polym. Sci. Polym. Phys. 28: 415. Mulder, M.H.V. (1996) Basic Principles of Membrane Technology, 2nd ed., Kluwer Academic Publishers, Dordrecht. Ozis, F. (2005) A percolation biofilm-growth model for biomass clogging in biofilters. Dissertation, University of Southern California. Reij, M.W. and Hartmans, S. (1996) Propene removal from synthetic waste gas using a hollowfibre membrane bioreactor. Appl. Microbiol. Biotechnol. 45: 730-736. Reiser, M., Fischer, K. and Engesser, K.H. (1994) Kombination aus Biowasher- under Biomembranverfahren zur Reinigung von Abluft under hydrophilen under hydrofoben Inhaltstoffen, VDI Berichte 1104, p. 1003. Shareefdeen, Z. and Singh, A. (2005) Biotechnology for Odour and Air Pollution Control, SpringerVerlag, Heidelberg, Germany. Studer, M.H. (2005) Novel membrane based biological waste gas treatment systems. Dissertation. Swiss federal Institute of Technology Zurich. Van Ras, N., Ogink, N., D’amico, A., Godia, F., Dixon, M., Llewellyn, D., Demey, D., Darlington, A., Kraakman, N.J.R., Eckhard, F. and Zoma, G. (2006) Biological air filter for air quality control. In: Microgravity applications programme: Successful teaming of science and industry. ESA publication SP-1290, Andrew Wilson (ed.), Noordwijk, the Netherlands, ISBN 92-9092-971-5. pp. 270-280.
This page intentionally left blank
323
Treatment of the confined air of a spacecraft cabin AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD AND VINCENT PLATEL Laboratoire de Thermique, Energétique et Procédés, Université de Pau et des Pays de l’Adour, Quartier Bastillac, 65000 Tarbes, France Frédéric Bellossi, Astrium Space Transportation, TO65, BP20011, 33165 Saint-Médard-en-Jalles, France
ABSTRACT A test bench at scale 1/15 has been implemented in order to evaluate the performance of photocatalysis and membrane bioreaction to treat the polluted air of a spacecraft cabin containing 30 mice. Ten pollutants could be generated with low concentration and seven of them could be analysed simultaneously. Formaldehyde was tested alone because of the needed volume for its analysis. First results have shown that photocatalysis process could lead to good efficiency except for methane and formaldehyde. This one is not sufficiently eliminated. A new configuration of photocatalysor allowing higher illumination in order to try to increase formaldehyde elimination will be tested. Membrane bioreactor have to be tested too. Interesting results are expected testing the two processes in series.
1 INTRODUCTION Indoor Air Quality (IAQ) has been a matter of great interest for several years now. Indeed a very large number of pollutants, among which a lot of organic compounds and especially volatile organic compounds (VOCs), have been detected in residential indoor (Mosqueron and Nedellec, 2004; Edwards et al., 2001). Moreover, as the cabin of a car can be considered to be a part of the living environment (because of the spent time in cars) some authors have also focused on identification of airborne organic compounds in a cabin of a car (Yoshida and Matsunaga, 2006). In the same manner, cabins of spacecrafts are also concerned with indoor air pollution. Improvement of IAQ lies in three kinds of solution:
323
324
AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD
AND
VINCENT PLATEL
– removal of , or, at least, reduction in pollution sources, – ventilation and thus air renewal, – polluted air treatment. As far as spacecrafts are concerned, ventilation and air renewal being obviously impossible, indoor air treatment is unavoidable. Since porous materials offer high capacity for VOCs, adsorption is a widely used process. But a non negligible airborne quantity of porous materials can be needed according to the duration of the space flight. This is a major disadvantage since little room is available in spacecraft. Regeneration of porous materials that could be considered would also need significant energy amount that could be prohibitive to space application. Hence the present study focuses on two other treatment processes: photocatalysis and membrane bioreactor. The objective of the study is to demonstrate the applicability and performance of these processes to the removal of pollutants in a spacecraft containing 30 mice. In the present study there are no outdoor pollutant sources neither sources arisen from human activities. Only pollutants coming from the mice metabolism, from the waste decomposition and from the material emission have then been considered to establish the list of the ten chosen pollutants (Yu and Crump, 1998; Phillips et al., 1999; Ruzsanyi et al., 2005; Sato et al., 2001; Jungbluth et al., 2001; Van Kempen et al., 2001; Mackie et al., 1998; Miller et al., 2001 and 2002; Powers et al., 2005; Schade et al., 1995; Zahn et al., 2001). That’s why the list is quite different from lists usually encountered in IAQ studies. For each pollutant, two concentrations have been taken into account : the maximum allowed concentration (MAC) for long term exposure and the short term exposure limit (STEL) which is the maximum acceptable concentration level during 15 minutes. Moreover, daily production rate have been estimated. Table 1 provides the list of the chosen pollutants and the MAC, the STEL and the production rate of each pollutant. Actually isoprene has been replaced by limonene (C10H15) which is easier to handle for performance assessment of the photocatalysor and of the bioreactor. An adequate airstream in the mice cages is necessary to provide removal of waste towards the filters of the cages. A fraction of this stream is led to the air conditionning system loop which contains the treatment process (as well as HEPA (high efficiency particulate air) filter, charcoal bed, condensing heat exchanger...). Thus the global removal efficiency depends on the proper efficiency of the treatment process and on this fraction of the air stream as well. The higher this fraction is, the lower the proper efficiency of the process can be to maintain the same global removal efficiency. Two values of the air flow have been chosen : 4 m3/h and 10 m3/h.
325
TREATMENT OF THE CONFINED AIR OF A SPACECRAFT CABIN
Table 1. Main gaseous pollutants to be removed and relevant concentrations (P=1atm, T=298K). Compounds
Short Term Exposure Limit (STEL) ppm
Maximum Allowed Concentration (MAC) ppm
Daily production rate (30 mice) ppm
Ammonia (NH3) Methane (CH4) Acetic Acid (CH3COOH) Propionic Acid (C2H5COOH) Acetone (CH3COCH3) 4-methylphenol (m-cresol) (CH3C6H4OH) Hydrogen Sulfide (H2S) Isoprene (C5H8) Formaldehyde (HCHO) Trimethylamine ((CH3)3N)
35 Not relevant 15
25 0,5 vol % * 10
1260 900 10,5
15
10
10,5
750
250
10,2
5
2,3
1,7
10
5
1,3
Not available
2
0,8
0,1
0,016
0,3
15
2
0,2
* The Lower Explosion Limit of methane is 5 vol %
The global removal efficiency of the two processes will be considered according to two parameters: Capacity to come back from the STEL value to the MAC value within less than 15 minutes Capacity to maintain the pollutant level below the MAC value on the long term taking daily production into account.
326
AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD
AND
VINCENT PLATEL
2 MATERIALS AND METHODS 2.1 PHOTOCATALYSOR The photocatalysor is cylindrical. The UV lamp stands on the axis with a photocatalyst layer all around. The inside walls of the shell are UV reflective in order to optimise the illumination level. The catalyst layer is a silica felt containing TiO2. The polluted air is forced through it (Figure 1). In order to test different illumination levels the following parameters can be changed : – – – –
UV lamp length : 300 mm, 400 mm, or 900 mm UV lamp electrical power : 10, 15, 25 or 30 W distance photocatalyst/UV lamp : 1 or 2 cm photocatalyst surface.
In the present study, different configurations of the photocatalysor have been tested depending on the chosen values of these parameters.
Figure 1. Shematic representation of the photocatalysor.
2.2 BIOREACTOR Since we deal with a space application, it is important to ensure the confinement of the biomass and the complete separation of the gaseous phase and the liquid phase containing the biomass. A classical biofilter would not allow this, that’s why a membrane bioreactor has been used. It is made up of a shell containing a bundle of 2600 hollow fibres with diameter of 600 μm. The polluted gaseous stream is forced through the fibres while the liquid phase and the biofilm stand on the outer faces of the fibres, in the shell of the bioreactor. The fibres are microporous hydrophobic membranes. Indeed dense membrane would have been probably too selective regarding the number of pollutants which must diffuse through it. Figure 2 shows the principle
TREATMENT OF THE CONFINED AIR OF A SPACECRAFT CABIN
327
of the microporous membrane bioreactor. The micro-organisms come from liquid sludge of waste water treatment station. During six months the liquid sludge tank has been stirred, aerated and nutrients have been added. The liquid sludge has then been filtered with 100 μm sieve and introduced in the bioreactor shell.
Figure 2. Principle of the microporous membrane bioreactor.
2.3 ANALYTICAL METHODS NH3 concentration was determined using a gas sensor (Dräger Sensor NH3 8101711). H2S concentration was measured by means of a specific apparatus dedicated to the sulphur compounds analysis, the ChromatoSud airMEDOR which is a gas chromatograph with an electro-chemical detector. Varian 3800 gas chromatograph equipped with a flame ionization detector allows the measurement of CH4 and acetone. Formaldehyde in the polluted air is sampled by an active-DNPH-silica cartridge (Waters XPosure Sampler WATO47205), extracted with acetonitrile and analysed with a Varian Pro Star HPLC-UV detector. All the other pollutants are analysed by means of a gas chromatograph with spectrometry mass detector (Thermofinnigan Trace GC-MS). A cryogenic preconcentrator is used in order to ensure a good detection. Indeed it removes water and carbon dioxide which would be responsible for interferences in the analyse if not, and concentrates the effluent at the same time. The preconcentration is made using a semi-volatile method. The detection mode of MS is Single Ion Monitoring (SIM) which decreases the detection limit of all compounds.
328
AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD
AND
VINCENT PLATEL
2.4 EXPERIMENTAL SETUP The test bench is a closed circuit which total volume is representative of real volume of mice habitats at scale 1/15. So the air flow is 4/15 m3/h or 10/15 m3/h. Experiments are performed at atmospheric pressure, at temperature about 20°C and a humidity about 40% (obtained by injection of distilled water by means of a syringe in a heated point). At the start of each experiment, the circuit is first open and a pure air is flushed from a zero air generator (dry air with maximal concentration in carbon compounds about 50 ppbv ) in the circuit to remove any impurity. Then the gaseous mixture is generated by different means. In order to introduce formaldehyde, trimethylamine and hydrogen sulphur, a calibration gas generator constituted with two permeation ovens is fed by zero air. Methane and ammonia are introduced in the air stream using gas tight syringes. Then a gaseous mixture of acetic and propionic acids, acetone, limonene and m-cresol is generated by introducing these components one by one in a heated air stream using a syringe diver (Harvard Apparatus PHD2000). These liquids are vaporised in the air stream which is dragged by a pump to a twenty liters balloon flask (in order to represent the mice habitat). Two mass flow meters allow the regulation of the air flow within the two treatment processes (photocatalysor and membrane bioreactor) which are implemented on this loop. Numerous valves allow numerous configurations : – by-pass of the two processes (this configuration is used when the gas mixture is generated), – by-pass of only one process, – both processes in series (photo-bio or bio-photo), – both processes in parallel. After the generation phase, the treatment phase begins with the choice of one of the previous configurations. It can be either a STEL to MAC test (Capacity to come back from the STEL value to the MAC value within less than 15 minutes) or a MAC to MAC test (Capacity to maintain the pollutant level below the MAC value on the long term taking daily production into account). Hence in the first case every pollutant has been first generated till the STEL concentration and in the second case till the MAC. At the end of the treatment phase, a sample is made using a tedlar gas sampling bag. Actually formaldehyde must be tested alone since the whole volume of the circuit must be forced through the active-DNPH-silica cartridge to ensure its detection. It can be noted that the two processes have been tested within a closed circuit to ensure the test bench the more representative as possible of real conditions. This choice led to the following constraints :
TREATMENT OF THE CONFINED AIR OF A SPACECRAFT CABIN
329
– to ensure airtightness, we had to check for leaks regularly and carefully, – only one analysis could be performed for each test because of the needed sample volume. Moreover, formaldehyde have been tested alone as it has been explained previously.
Figure 3. Experimental setup.
3 RESULTS AND DISCUSSION 3.1 CALIBRATION Numerous problems had to be solved during calibration step. m-cresol is a quite viscous liquid. So it is introduced in the circuit within acetone. But it seams that during the analysing step, it is difficult to eliminate it from the preconcentrator. Hence the calibration is very difficult, and at present we have eliminated it from our tests in order to ensure that it will not interfere with other analyses. We will reintroduce it at the end of the tests when an optimal process configuration will have been chosen. The MS detector doesn’t ensure a separate detection of acetone and trimethylamine. So tests are performed with a gas mixture containing either acetone or trimethylamine. Thus the number of calibration tests and the number of treatment tests is twice.
330
AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD
AND
VINCENT PLATEL
No calibration is required for H2S because airMEDOR apparatus is equipped with an internal calibration. No calibration is either required for ammonia. Calibrations of methane, acetone, limonene, propionic acid and formaldehyde are quite good with linear regression coefficient being 0,9 or more. Calibrations of acetic acid and trimethylamine are not so good (linear regression coefficient between 0,7 and 0,8). Moreover, the quantification limit of acetic acid is higher that the MAC. At last, as is has been already said, apparatus can’t be properly calibrated for m-cresol. 3.2 TESTS Only the photocatalysor has been tested by now. At present, three configurations have been used for STEL to MAC tests. Table 2 provides each pollutant concentration after treatment. STEL (which is initial concentration of each test) and MAC are recalled. It can be seen that pollutants are eliminated with good efficiency except methane. It is expected that methane will be removed by bioreactor process. Concentrations of ammonia, propionic acid, acetone, hydrogen sulfide, limonene and trimethylamine are lower than MAC that is an encouraging result. Since the quantification limit of acetic acid is higher than its MAC, no conclusion can be drawn from the previous results for this component. But since it is a smaller molecule than propionic acid which is well eliminated, we can expect that acetic acid is also eliminated. Formaldehyde is not sufficiently removed : its concentration is still higher than the MAC. A new configuration allowing higher illumination has to be tested.
4 CONCLUSIONS The following conclusions can be drawn from the results presented in this study. A test bench at scale 1/15 has been implemented in order to evaluate the performance of photocatalysis and membrane bioreaction to treat the polluted air of the spacecraft cabin containing 30 mice. Ten pollutants could be generated with low concentration and seven of them could be analysed simultaneously. Only one pollutant could not be analysed (m-cresol). Formaldehyde was tested alone because of the needed volume for its analysis. First results have shown that photocatalysis process could lead to good efficiency except for methane and formaldehyde. This one is not sufficiently eliminated.
331
TREATMENT OF THE CONFINED AIR OF A SPACECRAFT CABIN
Table 2. Pollutants concentrations after 15 minutes of treatment (P=1atm, T=298K). illumination level (mW) Air flow (m3/h) Ammonia (NH3) Methane (CH4) Acetic Acid (CH3COOH) Propionic Acid (C2H5COOH) Acetone (CH3COCH3) Hydrogen Sulfide (H2S) Limonene (C10H15) Formaldehyde (HCHO) Trimethylamine ((CH3)3N) QL : Quantification Limit
390
390
394,2
4/15
10/15
4/15
< 2,5
< 2,5
< 2,5
35
25
8624 7062 7507 <12,5 (QL)
8097 <12,5 (QL)
8136 <12,5 (QL)
10 000 15
5000 10
<6 (QL)
<6 (QL)
<6 (QL)
15
10
213 193 2,50,6
210 0,07
87 0,01
750 10
250 5
0,380,38
0,390,44
0,82
4
2
0,063
0,031
0,1
0,016
0,99
1,74
0,78 0,74
15
2
Short Term Exposure Limit (STEL) ppm
Maximum Allowed Concentration (MAC) ppm
Concentration after 15 minutes of treatment (ppm)
Next steps of this study will be : – test of a new configuration of photocatalysor allowing higher illumination in order to try to increase formaldehyde elimination, – MAC to MAC tests on the photocatalysor selected configuration after STEL to MAC tests, – STEL to MAC tests with the membrane bioreactor (methane removal is expected) – MAC to MAC tests with the membrane bioreactor, – tests with the two processes in series.
332
AUDREY RAMIS, CÉCILE HORT, SABINE SOCHARD
AND
VINCENT PLATEL
Interesting results are expected testing the two processes in series. Use of photocalatysis as a pre-treatment of bioreaction have already been investigated for wastewater (Mohanty et al., 2005) and more recently for airborne toluene and oxylene (Moussavi et al., 2006). Indeed, photocatalysis partially oxidized pollutants into more biodegradable intermediates. Moussavi et al. have shown that the coupled UV-bioflitration system provided up to 60 % additional contaminant removal compared to the sum of that offered by UV and reference biofilter. Thus such a coupled treatment applied in the present study could lead to reduce the electrical power needed for photocatalysis and to reduce the volume of the bioreactor too.
5 ACKNOWLEDGEMENTS The present research has been supported by ASTRIUM-Space Transportation within the frame of a project of the European Space Agency (ESA).
REFERENCES Edwards, R.D., Jurvelin, J., Saarela, K. and Jantunen, M. (2001) VOC concentrations measured in personal samples and residential indoor, outdoor and workplace microenvironments in EXPOLIS-Helsinki, Finland, Atmospheric Environment 35: 4531-4543. Jungbluth, T., Hartung, E. and Brose, G. (2001) Greenhouse gas emissions from animal houses and manure stores. Nutrient Cycling in Agroecosystems 60: 133-145. Van Kempen, T.A.T.G. (2001) Dietary adipic acid reduces ammonia emission from swine excreta Journal of Animal Science 79(9): 2412–2417. Mackie, R.I., Stroot, P.G., and Varel, V.H. (1998) Biochemical identification and biological origin of key odor components in livestock waste. Journal of Animal Science 76: 1331-1342. Miller, D.N. and Varel, V.H. (2001) In vitro study of the biochemical origin and production limits of odorous compounds in cattle feedlots. Journal of Animal Science 79: 2949-2956. Miller, D.N. and Varel, V.H. (2002) An in vitro study of manure composition on the biochemical origins, composition, and accumulation of odorous compounds in cattle feedlots. Journal of Animal Science 80: 2214-2222. Mohanty, S., Rao, N.N., Khare, P. and Kaul, S.N. (2005) A coupled photocatalytic-biological process for degradation of 1-amino-8-naphthol-3, 6-disulfonic acid (H-acid). Water Research. 39(20): 5064-5070. Mosqueron, L. and Nedellec, V. (2004) Revue des enquêtes sur la qualité de l’air intérieur dans les logements en Europe et aux Etats-Unis, Paris, Observatoire de la qualité de l’air intérieur : 1-55.
TREATMENT OF THE CONFINED AIR OF A SPACECRAFT CABIN
333
Moussavi, G. and Mohseni, M. (2006) Using UV pretreatment to enhance biofiltration of mixtures of aromatic VOCs, Journal of Hazardous Materials 144: 59-66 Phillips, M., Herrera, J., Krishnan, S., Zain, M., Greenberg, J. and Cataneo, R.N. (1999) Variation in volatile organic compounds in the breath of normal humans. Journal of Chromatography B 729: 75-88. Powers, W.J., Bastyr, S.B., Harmon, J. and Kerr, B.J. (2005) Proceedings of the AWMA. Minneapolis, MN, June, Paper # 1107. Ruzsanyi, V., Baumbach, J.I., Sielemann, S., Litterst, P., Westhoff, M. and Freitag, L. (2005) Detection of human metabolites using multi-capillary columns coupled to ion mobility spectrometers. Journal of Chromatography A 1084: 145-151. Schade, G.W. and Crutzen, P.J. (1995) Emission of aliphatic amines from animal husbandry and their reactions: potential sources of N2O and HCN. Journal of Atmospheric Chemistry 22 (3): 319-346. Sato, H., Hirose, T., Kimura, T., Moriyama, Y. and Nakashima, Y. (2001) Analysis of malodorous volatile substances of human waste : feces and urine. Journal of Health Science 47: 483-490. Yoshida, T. and Matsunaga, I. (2006) A case study on identification of airborne organic compounds and time courses of their concentrations in the cabin of a new car for private use. Environment International 32: 58-79. Yu, C. and Crump D. (1998) A review of the emission of VOCs from polymeric material used in buildings. Building and Environment 33: 357-374. Zahn, J.A., Hatfield, J.L., Laird, D.A., Hart, T.T. and Do Y.S. and DiSpirito, A.A. (2001) Functional classification of swine manure management systems based on effluent and gas emission characteristics. J. Environ. Qual. 30 (2):635-647.
This page intentionally left blank
335
Gas-phase toluene biodegradation by Burkholderia vietnamiensis G4 in a biofilm membrane reactor AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA, HERMAN VAN LANGENHOVE* Research Group EnVOC, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, B-9000, Ghent, Belgium
ABSTRACT A laboratory-scale biofilm membrane bioreactor inoculated with Burkholderia Vietnamiensis G4 was examined to treat toluene vapors from a synthetic waste gas stream. The gas feed side and nutrient solution were separated by a composite membrane consisting of a porous polyacrylonitrile (PAN) support layer coated with a very thin (0.3 μm) dense polydimethylsiloxane (PDMS) top layer. After inoculation, a biofilm developed on the dense layer. The biofilm membrane bioreactor was operated continuously at different residence times (28-5 sec) and loading rates (1.2-17.7 kg m-3 d-1), with an inlet toluene concentrations ranging from 0.21-4.1 g m-3. The overall performance of the membrane bioreactor was evaluated over a period of 151 days. Removal efficiencies ranging from 78-99% and elimination capacities ranging from 4.2-14.4 kg m-3 d-1 were observed depending on the mode of operations. A maximum elimination capacity of 14.4 kg m-3 d-1 was observed at a loading rate of 17.4 kg m-3 d-1. Overall, the results illustrate that biofilm membrane reactors can potentially be more effective than conventional biofilters and biotrickling filters for the treatment of air pollutants such as toluene.
1 INTRODUCTION Biological methods for treating contaminated air are usually divided into four categories: biofilter, biotrickling filters, bioscrubbers, and membrane bioreactors. Biological treatment is advantageous compared to physical/chemical treatments when the VOCs are biodegradable and the concentration is low. These advantages include
335
336
AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA AND
HERMAN VAN LANGENHOVE
low capital and operating cost, low energy requirement, and the absence of waste products that require further treatment or disposal (Wu et al., 1999; Zilli et al., 2000). Biofiltration has been widely studied for the control of biodegradable and odorous VOCs in air. However, studies and field application of these systems have been limited to inlet VOC loading rates of less than 50 g m-3 h-1 (Wu et al., 1999). At high VOC loading rates, microbial growth results in the clogging of media pore spaces with microbial biomass. This causes channelling in the packed bed, which consequently results in deterioration of the unit performance. Finally, the system fails due to high head losses across the bed. In addition, these systems are of limited use where degradation results in the accumulation of acidic compounds (Zilli et al., 2000; Ergas et al., 1995). Moreover, control of humidity and moisture contents of the packing materials is a difficult task in biofiltration processes (Sun et al., 2002). In a membrane bioreactor for waste gases (MBRWG), liquid phase and waste gas remain separated by a membrane and are subsequently degraded by the microorganisms in the biofilm attached to the membrane surface. A conceptual diagram of a membrane bioreactor is shown in Figure 1. Kumar et al. (2007) conducted a review of developments concerning MBRWG. Several bench-scale studies have demonstrated the value of dense phase membrane bioreactors (Attawayet et al., 2001; Ficth et al., 2003; Freitas dos Santos et al., 2003), while others have focused on the removal of contaminants from air using a porous membrane module (Ergas et al., 1999; Keskiner and Ergas 2000). In a composite membrane bioreactor, the porous layer is used as support, while the thin dense layer prevents microbial growth through the membrane (Van Langenhove et al., 2004). Prior studies on toluene biotreatment have highlighted challenges in obtaining effective toluene treatment. The volumetric degradation rates of toluene were often too low for the process to be practical. Usually, this was due to low activity of the culture or the system became biokinetic and/or mass transfer limited over a period of time (Kumar et al., 2007). So far MBRWG for toluene removal have been seeded by pure culture (Pseudomonas putida) or by bacterial consortia enriched from activated sludge as biofilm or suspended cells (Kumar et al., 2007). Biological treatment of VOCs in air depends on the ability of certain microorganisms to metabolise these VOCs and use them as their sole source of carbon and energy producing carbon dioxide, water vapor, and biomass (Mutafov et al., 2004). Thus, a microbially engineered bioreactor system that could effectively treat toluene over extended period of time would be desirable. The Burkholderia cepecia complex members possess considerable biotechnological potential as agents of bioremediation (O’Sullivan and Mahenthiralingam, 2005). Burkholderia cepecia G4 proficiently degraded toluene in a foamed emulsion bioreactor (Kan and Deshusses, 2005). It is expected that Burkholderia Vietnamiensis G4, a member of genus Burkholderia can proficiently degrade toluene in a MBRWG.
GAS-PHASE TOLUENE BIODEGRADATION BY BURKHOLDERIA VIETNAMIENSIS
337
The aim of present study was to evaluate the long-term performance of a MBRWG treating toluene vapors by Burkholderia Vietnamiensis G4 under various operating conditions. A comparison between present and prior study on MBRWG for toluene removal was also made.
2 MATERIALS AND METHODS 2.1 LAB-SCALE MEMBRANE BIOREACTOR SET-UP MBRWG was set up as shown in Figure 1. Commercially available PDMS/ PAN composite membrane (GKSS, Germany, 40 cm2 effective membrane area) was used, consisting of PDMS as hydrophobic dense top layer with a thickness of 0.3 μm and PAN as hydrophobic support layer material with a thickness of 185 μm. The membrane was incorporated into the Perspex reactor module. Through one compartment, mineral medium was recirculated at the dense membrane side at a flow rate of 75 cm3 min-1 by a peristaltic pump (2) (Masterflex, Cole Parmer). For all the experiments described herein, the MBR was rinsed with ethanol, and the mineral medium and heat resistant reactor parts were autoclaved prior to the experiments. This ensured that Burkholderia Vietnamiensis G4 remained the dominant organism in the system.
Figure 1. Experimental set-up of the membrane bioreactor.
338
AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA AND
HERMAN VAN LANGENHOVE
The mineral medium (MM) used for MBR consisted of 1 g L-1 KH2PO4, 1 g L-1 K2HPO4, 1 g L-1 KNO3, 1 g L-1 NaCl, 0.2 g L-1 MgSO4, 26 mg L-1 CaCl2.2H2O, 5.2 mg L-1 EDTA Na4 (H2O)2, 1.5 mg L-1 FeCl2, 4H2O, 0.1 mg L-1 MnCl2. 2H2O, 0.012 mg L-1 CoCl2.6H2O, 0.07 mg L-1 ZnCl2, 0.06 mg L-1 H3BO3, 0.025 mg L-1 NiCl2 6H2O, 0.025 mg L-1 NaMo4.2H2O, 0.015 mg L-1 CuCl2.2H2O. Between the pump and the module (4), a pulse dampener (3) (Cole Parmer) was placed. The MM was magnetically stirred at 400 rpm (1) (IKA RCT basic, IKA labortechnik). 2.2 ANALYTICAL METHODS Gas phase toluene concentration was measured using a Varian 3700 gas chromatograph (Varian Associates, Inc.) coupled with FID detector. Gas samples were taken in triplicate with a 1 mL Vici gas syringe. The residual standard deviation on the measurements were less than 10%. Water phase toluene concentrations were determined by taking 1 mL water samples with a plastic syringe (BD plastipak). The samples were brought into a 4.5 ml vial with a Teflon®-lined Mininert® screw cap and placed in a thermostatic bath at 30.0°C. After 2 hours, 1 mL of the gas phase was sampled and injected into the gas chromatograph. Cell dry weight was determined gravimetrically (APHA, 1980). The pH was measured with a Jenway 3310 apparatus, equipped with a Hanna Instruments electrode.
3 RESULTS AND DISCUSSION 3.1 MEMBRANE BIOREACTOR PERFORMANCE The reactor was seeded with the Burkholderia Vietnamiensis G4, which had been grown in a mineral medium with toluene as a sole carbon and energy source. During the operation period of 151 days, toluene loading rate, gas residence time, and removal efficiency of toluene are shown in Figure 2, air flow rates and toluene feeding controlled by mass flow regulator determined the gas residence time and toluene loading rate in the membrane bioreactor. The performance of the membrane bioreactor was evaluated by the following performance parameters: toluene loading rate, removal efficiency, elimination capacity. The definitions of these parameters are set out below: (2)
(3)
GAS-PHASE TOLUENE BIODEGRADATION BY BURKHOLDERIA VIETNAMIENSIS
339
(4)
3.2 MEMBRANE BIOREACTOR START-UP (PERIOD I: 1 – 43 DAYS) In membrane bioreactor, composite membrane (PDMS/PAN) was incorporated in the Perspex reactor module, TOL loaded air and mineral medium remain separated by the composite membrane. The inoculum was recirculated along the dense side of the membrane, while TOL loaded air diffuses through the porous side of the membrane and subsequently degraded by the microorganisms in the biofilm attached to the dense membrane. After two days, more than 60 % TOL removal was obtained. The microbial suspension was replaced by fresh MM, and thus all non-adhering cells were removed. During the first 43 days, the gas residence time (τ) was set at 11 s. Toluene removal efficiency increased and reached 74 % with an average loading rate of 7.2 kg m-3d-1. During the start-up period water condensation at the feed side was observed but after a period of 15 days it was no longer observed. This may be due to the development of biofilm growth (visible) on the dense side. During period of 22-34 days, a 30% decrease in removal efficiency was observed, but after replacement of mineral medium (day 23, 35) it could recover to 74 % removal efficiency. However, after increasing the gas residence time to 28 s consequently decreasing TOL average loading rate to 1.2 kg m-3 d-1 could recover to 99 % TOL removal efficiency. This could be explained by biomass growth and enzyme production is necessary for TOL removal. 3.3 INFLUENCE OF LOADING RATE AND GAS RESIDENCE TIME ON THE REACTOR PERFORMANCE After period I (start-up), different periods (II to VIII) were established with decreasing residence time from 28 s (period II), 24 s (period III), 20 s (period IV), 15 s (period V), 10 s (period VI), and 5 s (period VII). During each of these periods, the MBRWG was subjected to a range of load conditions to determine the removal characteristics through the unit. TOL inlet concentrations (Cin) were changed between 0.21 to 4.10 TOL g m-3. The gas residence time was switched between 28 s and 5 s. Consequently, the mass loading rate (LR) was increased from to 0.67 to 17.7 kg m-3 d-1. At day 44 gas residence time was increased from 11 to 28 s. During period II (44-51 d) at LR of 0.84 to 1.88 kg m-3 d-1 at τ = 28 s removal efficiency was 99%. During period III (52-84 d) at LR of 1.89 to 14.4 kg m-3 d-1 at a τ = 24 s removal efficiency reached 99%. During period IV (85109 d) at LR of 4.1 to 13.87 kg m-3 d-1 at τ = 20 s removal efficiency decreased to 86%. During period V (110-126 d) at LR of 4 to 16.68 kg m-3 d-1 at τ = 15 s removal efficiency dropped to 86%. During period VI (127-140 d) at LR of 6.9 to 15.52 kg m-3 d-1 at τ = 10 s removal efficiency of 78% was observed. During period VII (141-151 d) at LR of 3.66 to 16.41 kg m-3 d-1 at τ = 5 s removal efficiency was 78%. As shown
340
AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA AND
HERMAN VAN LANGENHOVE
Figure 2. Performance of membrane bioreactor under different operating conditions.
in Figure 4, the TOL removal efficiency decreased as gas residence was decreased. For a gas residence time longer than 5 s, the removal efficiency was always >78 %. After changing the concentrations and/or the gas residence time, removal efficiency and elimination capacity became stable after 20-24 h. Each setting was kept constant for 4-5 days to be sure that reactor performance was stable over time. Overview of the results plotted in Figure 2 demonstrates that the removal efficiency depends on both the gas residence time and the inlet concentration. The removal efficiency was maintained at 78 % for an inlet load of 16.7 kg m-3 d-1 at a gas residence time of 5 s, but declined at higher loads. It appears that growth of micro-organisms based on dry matter determination (data not shown) is inhibited at higher toluene loading rates. The result obtained during the present study is compared and discussed with prior studies in Table 1. 3.4 ELIMINATION CAPACITY Elimination capacity (EC) is one important parameter to evaluate the MBR performance. The performance of membrane bioreactor under different operational parameters can be summarized by plotting the EC against the LR. It can be seen from Figure 3 that > 90 % removal efficiency was obtained at organic loading rate up to 14.4 kg m-3 d-1 (τ = 20 s). At LR of 16.4 kg m-3 d-1 (τ = 5 s), removal efficiency decreased to 20%. There was a trend of increasing elimination capacity with increasing inlet loading and then reaching a constant level, which was named as maximum elimination capacity (Figure 3).
GAS-PHASE TOLUENE BIODEGRADATION BY BURKHOLDERIA VIETNAMIENSIS
341
Figure 3. Average elimination capacity (EC) for TOL as a function of loading rate, operate at a residence time of 24, 20, 10 and 5s. The straight line represents 100% removal efficiency, while dotted lines are best fits of data.
4 COMPARISON OF THE PERFORMANCE OF VARIOUS MEMBRANE BIOREACTORS FOR TOLUENE REMOVAL In Table 1 entries include reactor design, operation and performance parameters, observed range of toluene, reactor dimensions, types of membrane, and inoculum type. Compared to a flat and capillar membrane configuration, hollow fibres have large specific gas-membrane contact area. Because of the large range in these specific membrane areas used in membrane bioreactor experiments, data on mass loading rate, LR, and elimination capacity, EC, should be compared per unit of available (specific) membrane area. Volumetric ECs suggest that a flat membrane configuration is inferior to hollow fibres. However, on the basis of the available membrane area, data are in the same order of magnitude. As can be seen in the Table 1, per unit of membrane area, ECm, max amounts 28.8 g toluene m-2 d-1, is the highest than obtained with other membrane bioreactors in the same range of loading rates. Only England and Fitch (2002) reported higher elimination capacity, but at loading rates that were more than 100 times larger than the loadings applied in this study. Differences in removal percentage between the current study and prior studies may be attributed to differences in compound mass transfer in membranes, air flow rates, membrane surface areas, and/or biofilm composition (Kumar et al., 2007).
342
AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA AND
HERMAN VAN LANGENHOVE
Table 1. Comparison of the performance of various gas-phase membrane bioreactors for the treatment of toluene.
Inoculum (co-substrate); Days 90 <1 120 168 n.r. 150 339 37 165
b = biofilm, s = suspend. cells Pseudomonas putida Tol1A Pseudomonas GJ40 Activated sludge Activated sludge n.r. Pseudomonas putida A1 Pseudomonas putida TVA8 Activated sludge Burkholderia Vietnamiemsis G4
Reactor set-up Configuration, b s b b b b b b b
type, material HF, P, PE F, P, PP HF, P, PP C, P, PSf* C, NP, PDMS HF, PE CM,PDMS/PVDF T, NP, PDMS CM, PDMS/PAN
τ
a 2
-3
m m 10256 500 20000 2622 n.r. n.r. 500 558 500
s 0.8 - 4.2 1.6 - 9.6 0.9 - 1.8 16 / 32 n.r. 0.5 – 1.3 8 - 24 1.0 5-28
Reactor performance ECm, max LRm η -2 -1
gm d 1.6 2.8 3.0 3.9 16 n.r. 19 144 28.8
1.6 8.1 8.6 4.7 84 n.r. 23 720 35.4
% 97 35 35 84 20 86 84 20 82
Ref. 1 2 3 4 5 6 7 8 This work
Configurations: HF: hollow fibre (ID < 0.5 mm), C: capillary (0.5 mm < ID < 10 mm), T: tubular (ID > 10 mm), SW:spiral-wound, F = flat membrane Membrane type: P : porous, NP : nonporous, CM : composite membrane, Membrane polymer: PP: polypropylene, PSf : polysulfone, PE : polyethylene, PDMS : polydimethylsiloxane; * indicates pores are water-filled, PVDF : polyvinylidenefluoride, Zrf : zirfon, n.r. : not reported or not sufficient data to calculate Notations : a: specific membrane area (m2 membrane per m3 air volume); LR: volumetric loading rate; LRm : loading rate per unit of available membrane area; η: removal efficiency; ECmax :maximum volumetric elimination capacity. [1] Ergas et al., 1997; [2] Parvatiyar et al., 1996a; [3] Ergas et al., 1999; [4] Parvatiyar et al., 1996b; [5] Reiser et al., 1994; [6] Dong et al., 2005; [7] Jacobs et al., 2004 ; [8] England and Fitch, (2002).
5 CONCLUSIONS The results presented herein clearly demonstrate that toluene can be effectively treated in a MBRWG. Depending on the conditions, high elimination rate or high removal percentage of toluene was obtained. Following conclusions can be drawn based on this study: 1) This study demonstrates the stability and good reactor performance of a composite membrane (PDMS/PAN) bioreactor for treatment of toluene contaminated air. The bioreactor was inoculated with Burkholderia
GAS-PHASE TOLUENE BIODEGRADATION BY BURKHOLDERIA VIETNAMIENSIS
343
vietnamiensis G4. The bioreactor performance was affected by the gas residence time and inlet concentration. Lowering the gas residence time at a constant loading rate resulted in lower reactor performance. A TOL maximum elimination capacity of 14.4 kg m-3 d-1 was observed, which is the highest degradation reported in the literature for similar loading rates to those used in the experiments. 2) In the beginning water condensation at the feed side was observed but after a period of 15 days it was no longer observed. It may be due to the development of biofilm growth on the dense side. 3) During period II, increasing the residence time from 11 to 28 s gives 99% removal efficiency at TOL LR of 1.2 kg m-3 d-1. 4) Compared to other MBRWG for toluene removal present study shows that use of Burkholderia vietnamiensis G4 is a good option for the treatment of toluene.
6 ACKNOWLEDGEMENTS The author acknowledges the Ghent University, Gent, Belgium for providing Special Research Grant (BOF) for doctoral research.
REFERENCES APHA. (1980) Standard methods for the examination of water and waste water, 15th ed.; American Public Health Association: Washington, DC. Attaway, H., Gooding, C. and Schmidt, M. (2001) Biodegradation of BTEX vapors in a silicone membrane bioreactor system. J. Ind. Microbiol. Biotechnol. 26: 316-325. Dong, K. and Jim, Heonki. (2005) Degradation of toluene vapor in a hydrophobic polyethylene hollow fiber membrane bioreactor with Pseudomonas putida. Proc. Biochem. 40: 2015-2020. England, E. and Fitch, M. (2002) Heat transfer and toluene removal in bench-scale membrane bioreactors. Proceedings of the Air and Waste Management Association Conference, Maryland, United States, 23-27. Ergas, S.J. and McGrath, M.S. (1997) Membane bioreactor for control of volatile organic compound emissions. J. Environ. Eng.-ASCE 123 (6): 593-598. Ergas, S.J., Schroeder, E. D., Chang, D.P.Y. and Morton, R. (1995) Control of VOC emissions from a POTW using a compost biofilter. Water Env. Fed. 67: 816-821.
344
AMIT KUMAR, JO DEWULF, MUNKHTSETSEG LUVSANJAMBA AND
HERMAN VAN LANGENHOVE
Ergas, S.J., Shumway, L, Fitch, M.W. and Neemann, J.J. (1999) Membrane process for biological treatment of contaminated gas streams. Biotechnol. Bioeng. 63 (4): 431-441. Fitch, M., Neemann, J. and England, E. (2003) Mass transfer and benzene removal from air using latex rubber tubing and a hollow fiber membrane module. Appl. Biochem. Biotechnol. 104: 199-214. Freitas dos Santos, L., Hommerich, U. and Livingston, A. (1995) Dichloroethane removal from gas stream by an extractive membrane reactor. Biotechnol. Prog. 11: 194-201. Jacobs, P., De Bo, I., Demeestere, K., Verstraete, W. and Van Langenhove, H. (2004) Toluene removal from waste air using a flat composite membrane bioreactors. Biotechnol. Bioeng. 85: 68-77. Kan E. and Deshusses M.A. (2005) Continuous operation of foamed emulsion bioreactor treating toluene vapors. Biotechnol. Bioeng. 92: 364-371. Keskiner, Y. and Ergas, S. (2000) Hollow fiber membrane bioreactor for aqueous and gas phase ammonia removal by nitrification. Hazard Ind. Wastes. 32: 867-876. Kumar, A., Dewulf, J. and Van Langenhove, H. (2007) Membrane based biological waste gas treatment. Chem. Eng. J. doi:10.1016/j.cej.2007.06.006 (in press). Mutafov, S., Angelova, B., Schmauder, H.P., Avramova, T. and Boyadijieva, L. (2004) Stoichiometry of microbial continuous-flow purification of toluene contaminated air. Appl. Microbiol. Biotechnol. 65: 222-234. O’Sullivan, L.A. and Mahenthiralingam, E. (2005) Biotechnological potential within the genus Burkholderia. Lett. Appl. Microbiol. 41: 8-11. Parvatiyar M.G., Govind R. and Bishop D.F. (1996a) Biodegradation of toluene in a membrane biofilter. J. Membr. Sci. 115: 121-127. Parvatiyar, M.G., Govind, R. and Bishop, D.F. (1996b) Treatment of trichloroethylene in a membrane biofilter. Biotechnol. Bioeng. 50: 57-64. Reiser, M., Fischer, K. and Engesser, K.H. (1994) Kombination aus Biowascher-und Biomembranverfahren zur reinigung von Abluuft und hydrophilen und hydrofoben Inhaltsstoffen, VDI Berichte, 1104 103. Sun, Y., Quan, Y., Chen, J., Yang, F., Xue, D. and Liu, Y. (2002) Toluene vapor degradation and microbial community in biofilter at various moisture content. Process. Biochem. 38: 109-113. Van Langenhove, H., De Bo, I., Jacobs, P., Demeestere, K. and Dewulf, J. (2004) A membrane bioreactor for the removal of dimethyl sulphide and toluene from waste air. Water Sci. Technol. 50: 215-224. Wu, G., Conti, B., Leroux, A., Brzeinski, R., Viel, G. and Hetiz, M. (1999) A high performance biofilter for VOC emission control. J. Air Waste Manage. Assoc. 49: 185-192. Zilli, M., Del Borghi, A. and Converti, A. (2000) Toluene vapour removal in a laboratory-scale biofilter. Appl. Microbiol. Biotechnol. 54: 248-254.
345
VOC REMOVAL IN CONVENTIONAL AND BIOTRICKLING FILTERS
This page intentionally left blank
347
Treatment of gas phase styrene in a biofilter under steady-state conditions ELDON R. RENE, MARÍA C. VEIGA AND CHRISTIAN KENNES Faculty of Sciences, Chemical Engineering Laboratory, University of La Coruña, 15071 – La Coruña, Spain
ABSTRACT Preliminary studies on the performance of a laboratory scale perlite biofilter inoculated with a mixed culture taken from petrochemical refinery sludge was evaluated for gas phase styrene removal under various operating conditions. Initially, the biofilter was acclimatized for 53 days at constant loading rates (40-60 g/m3.h), wherein the performance gradually improved with fluctuations in the removal profiles. Experiments were carried out by subjecting the biofilter to different flow rates (150, 300 l/h) and concentrations (0.5 – 5 g/m3), that corresponds to inlet loading rates between 60 – 200 g/m3.h. The results from this study show100% styrene removal with a maximum elimination capacity of 190 g/m3.h.
1 INTRODUCTION Styrene is an important chemical feedstock, which is used commonly as a raw material for the synthesis of plastics, synthetic resins, butadiene-styrene latex, styrene copolymers and unsaturated polyester resins (Jorio et al., 2000). Due to improper practices and treatment, a substantial amount of vapours containing styrene are being emitted to the ambient atmosphere from process industries. It is reported to have significant effect on human health and natural environment. Exposure to even low concentrations of styrene could cause contact-based skin inflammation, irritation of the eyes, nose and respiratory tract, and may induce narcotism (Fielder and Lowing, 1981). This has led to increased attention from the regulatory authorities, that has helped to achieve continuous modifications/development in the existing control technologies.
347
348
ELDON R. RENE, MARÍA C. VEIGA AND CHRISTIAN KENNES
Biodegradation is a promising alternative for the mineralization of volatile organic compounds (VOCs). The most widely used biological processes for waste gas treatment are bioscrubbing, trickling biofilters and biofilter. The simplicity in the operation of biofilters has resulted in its emergence as a more practical treatment option (Kennes and Veiga, 2001). Biofilters have also proven to be effective in treating large volumes of VOCs at relatively high concentrations (Kennes et al., 1996; Mohammad et al., 2007). A complex phenomenological step consisting of adsorption, absorption, diffusion and biodegradation takes place in a biofilter where the pollutant is converted to non-toxic end products (Devinny et al., 1999). Furthermore, the removal and oxidation rates of these hazardous contaminants depend principally on the biodegradability, reactivity and largely on the solubility of the pollutant in the liquid layer of the biofilm. Biofiltration studies have been tested with different packing materials and with a wide variety of pollutants having different degradation rates. Typical examples are biofilters packed with perlite as inert carrier material that have been used to treat styrene or alkylbenzene vapours (Kennes et al., 1996; Cox et al., 1997; Paca et al., 2001; Veiga and Kennes, 2001). For styrene removal in biofilters, individual or mixed species of bacteria have generally been used according to literature. Pseudomonas sp. represents the most common group of isolates capable of styrene degradation and has been shown to produce styrene mono-oxygenase, which plays a major role in styrene degradation (O’Leary et al., 2002). Jang et al (2004) used Pseudomonas sp in a biofilter packed with peat and ceramic beads to treat styrene vapours and was able to show a maximum elimination capacity (EC) of 170 g/m3.h. Similarly, a mixed culture biofilter packed with perlite showed maximum EC of 145 g styrene/m3.h (Weigner et al., 2001). Ryu et al. (2004) used a polyurethane foam biofilter inoculated with activated sludge and achieved EC ranging between 580~635 g/m3.h at a space velocity (SV) of 50~200 per hour. These experimental studies have proved biofiltration as an efficient waste gas treatment process and a reliable technology for the control of off gases containing styrene. This paper present the performance of a perlite based mixed culture biofilter treating styrene vapours at different concentrations and flow rates.
2 MATERIALS AND METHODS 2.1 MICROBIAL SEED A mixed microbial culture obtained from a petrochemical refinery sludge was used to inoculate the biofilter. This was done by filling the biofilter with the sludge and draining it after 12 hours. The procedure was repeated several times until visible biomass was noticed on the surface of perlite.
TREATMENT OF GAS PHASE STYRENE IN A BIOFILTER UNDER STEADY-STATE CONDITIONS
349
2.2 BIOFILTER The biofilter was made of glass having a diameter (ID) of 10 cm and 70 cm in height. The packing in the biofilter consisted of sieved perlite beads (4-6 mm). A perforated plate at the bottom provided the support for the packing while another plate at the top acted as a distributor for gas flow and mineral salt media addition. Gas sampling ports sealed with rubber septa were provided at equal intervals along the biofilter height. 2.3 EXPERIMENTAL A schematic of the experimental setup is given in Figure 1. Humidified styrene vapors at constant flow and concentration, controlled through valves were passed through the bed in a down flow mode. The bed moisture was maintained constant by periodic addition of fresh mineral salt medium (Kennes et al., 1996) from the top. Experiments were carried out by varying the flow rates of the styrene vapors and humidified air independently to get different initial concentrations and residence times in the biofilter. Gas samples were collected from different ports and analyzed for residual styrene and CO2 concentration.
Figure 1. Schematic of the perlite biofilter.
350
ELDON R. RENE, MARÍA C. VEIGA AND CHRISTIAN KENNES
2.4 ANALYTICAL METHODS Styrene concentration in gas samples were measured by gas chromatography on an HP 5890 gas chromatograph, using a 50 m TRACER column and a FID detector. The flow rates were 30 ml/min for H2 and 300 ml/min for air. Helium was used as the carrier gas at a flow rate of 2 ml/min. The temperatures at the GC injection, oven and detection ports were 150, 150 and 150°C respectively. CO2 was analyzed with a HP 5890 gas chromatograph equipped with a TCD detector. The injection and oven temperatures were 90 and 25 oC respectively, with the TCD set at 100 oC. Biomass concentration, as g of dry biomass/ g of perlite was measured according to the procedure outlined by Mohammad et al. (2007).
3 RESULTS AND DISCUSSIONS The performance of the biofilter was evaluated in terms of two parameters, the removal efficiency (RE, %) and the elimination capacity of the filter bed (EC, g/m3.h). The biofilter was initially acclimatized to styrene vapours by passing low concentrations and low gas flow rates (150 LPH) for 53 days to obtain sufficient biomass concentration in the filter bed. The biofilter was run under these conditions to achieve stable and high removal efficiencies. During the operation, the relative humidity of the air stream was maintained at around 95%. Scanning electron microscopy (SEM) was used to visualize the perlite particles colonized with microbial populations. These images are shown in Figure 2, which clearly show the presence of different colonies of microbes that are potential styrene oxidizers. The degree of acclimatization primarily depends on the adaptive capability of the microorganisms inoculated on the perlite, substrate concentrations, nutrient concentration and its availability and other necessary environmental conditions. After acclimatization, the combined effect of styrene inlet concentration and gas flow rate was investigated in two phases that correspond to residence times of 2 min and 1 min (Figure 3). On day 54, when the concentration was increased to 4 g/m3, the removal efficiency dropped to about 82%. However, the inlet concentration was later increased in small time steps to values as high as 5 g/m3, where the biofilter showed 100% removal efficiencies. In the next phase, the flow rate was increased to 300 LPH corresponding to inlet loads varying between 80 to 190 g/m3.h. The removals were high and consistent (100%), indicating the stability of the biomass and high performance of the biofilter. The elimination capacity, which reflects the capacity of the biofilter to remove the pollutants, is plotted as a function of the inlet styrene load in Figure 4. Though there were fluctuations in the EC values during start-up, under steady state conditions, a linear relation between the two variables was observed with a maximum EC of 190 g/m3.h. The results from this study are higher than most of the studies
TREATMENT OF GAS PHASE STYRENE IN A BIOFILTER UNDER STEADY-STATE CONDITIONS
351
reported in literature using biofilters for handling styrene vapours, which could be due to the dominant presence of fungi as shown in Figure 2. Indeed, it has been reported that fungal dominant biofilters would allow reaching a better performance than usual for hydrophobic VOCs (Kennes and Veiga, 2004).
Figure 2. SEM images of different biomass on the surface of perlite.
Figure 3. Start-up of the biofilter and effect of flow rate and concentration on the performance of perlite biofilter.
352
ELDON R. RENE, MARÍA C. VEIGA AND CHRISTIAN KENNES
Figure 4. Effect of inlet loading rate on the elimination capacity of biofilter.
For better understanding the styrene elimination mechanism within the reactor, the concentration profile at different heights was measured at a constant loading rate. The results indicate that styrene removal is more efficient in the higher part (Port 2) than in the lower section of the filter bed. Nearly 40% of styrene was removed in the first section followed by 30% respectively in the other two sections (Figure 5). This may be due to a higher concentration of microbial populations and higher moisture content in the upper section of the filter bed. In any biofiltration process, the volatile organic compounds are aerobically degraded to water, carbon dioxide, and biomass. Hence it is important to monitor the profile of CO2 in gas phase, at the inlet and outlet of the biofilter. These profiles are also shown in Figure 5, when the inlet loading rate was 84 g/m3.h. Complete mineralization was confirmed from these CO2 profiles measured along the biofilter height. It is also well know that biological waste gas treatment is an exothermic process. Hence the temperature profile across the biofilter is expected to vary depending on the inlet load applied to the filter, as also observed by others (Mohammed et al., 2007). At a constant inlet load of 84 g/m3.h, the temperature difference across the biofilter was 2.5 oC (Figure 6). The biomass concentration was also monitored periodically. However there were only minor variations in their concentrations across the biofilter height. This variation may be attributed in part to the variations of microbial dynamics in different sections and the corresponding specific activities and metabolic pathways utilized by the dominant strains involved in the styrene degradation.
TREATMENT OF GAS PHASE STYRENE IN A BIOFILTER UNDER STEADY-STATE CONDITIONS
Figure 5. Removal of styrene and CO2 evolution profile along the biofilter height (Inlet concentration – 1.2 g/m3, flow rate – 300 l/h).
Figure 6. Temperature and biomass dry weight profile along the biofilter height (Day 99: Inlet loading rate – 84 g/m3.h, Removal efficiency – 100%).
353
354
ELDON R. RENE, MARÍA C. VEIGA AND CHRISTIAN KENNES
4 CONCLUSIONS The results from this preliminary study show that the removal of styrene vapours from off gas emissions can be performed with high efficiencies by means of a biofilter. A petrochemical sludge provides easily adaptable microbial strains capable of degrading VOCs. Removal efficiencies as high as 100% were achieved under the present operational conditions at inlet loads varying between 60 – 190 g/m3.h, reaching a high maximum EC of 190 g/m3.h, which could be due to the dominant presence of fungi.
5 ACKNOWLEDGEMENTS The present research was funded by the Spanish Ministry of Education and Science (Project No: CTM2004-00427/TECNO) and through European FEDER funds. The post doctoral research of E.R.R was funded by the same Ministry.
REFERENCES Cox, H.H.J., Moerman, R.E., van Baalen, S., van Heiningen, W.N.M., Doddema, H.J. and Harder, W. (1997) Performance of styrene degrading biofilter containing the yeast Exophiala jeanselmei. Biotechnol. Bioeng. 53: 259-266. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. Lewis Publisher, Boca Raton. Fielder, R.J. and Lowing, R. (1981) Toxicity Review 1: Styrene. Health and Safety Executive, London. Jang, J.H., Hirai, M. and Shoda, M. (2004) Styrene degradation by Pseudomonas sp. SR-5 in biofilters with organic and inorganic packing materials. Appl. Microbiol. Biotechnol. 65(3): 349-355. Jorio, H., Bibeau, L. and Heitz, M. (2000) Biofiltration of air contaminated by styrene: effect of nitrogen supply, gas flow rate and inlet concentration. Environ. Sci. Technol. 34: 1764-1771. Kennes, C., Cox, H.H.J., Doddema, H.J. and Harder, W. (1996) Design and performance of biofilters for the removal of alkylbenzene vapours. J. Chem. Technol. Biotechnol. 66: 300-304. Kennes, C. and Veiga, M.C. (2001) Conventional Biofilters, In: Bioreactors for waste gas treatment, (Eds: Kennes, C., Veiga, M.C), Kluwer Academic Publishers, Dordrecht, The Netherlands, 47-98. Kennes, C. and Veiga, M.C. (2004) Fungal biocatalysts in the biofiltration of VOC polluted air. J. Biotechnol. 113: 305-319.
TREATMENT OF GAS PHASE STYRENE IN A BIOFILTER UNDER STEADY-STATE CONDITIONS
355
Mohammad, B.T., Veiga, M.C. and Kennes, C. (2007) Mesophilic and thermophilic biotreatment of BTEX polluted air in reactors. Biotechnol Bioeng. 97: 1423-1438. O’Leary, N.D., O’Connor, K.E. and Dobson, A.D.W. (2002) Biochemistry, genetics and physiology of microbial styrene degradation. FEMS Microbiol. Rev. 26: 403-417. Paca, J., Koutsky, B., Maryska, M. and Halecky, M. (2001) Styrene degradation along the bed height of perlite biofilter. J. Chem. Technol. Biotechnol. 76: 873-878. Ryu, H.W., Kim, J., Cho, K-S., Jung, D.J. and Lee, T.H. (2004) Biological treatment of air contaminated with styrene. In: Proceedings of the Better Air Quality Conference, CAIAsia, Agra, India. Veiga, M.C. and Kennes, C. (2001) Parameters affecting performance and modelling of biofilters treating alkylbenzene polluted air. Appl. Microbiol. Biotechnol. 55: 254-258. Weigner, P., Paca, J., Loskot, P., Koutsky, B. and Sobotka, M. (2001) The start-up period of styrene degrading biofilters. Folia Microbiol. 46: 211-216.
This page intentionally left blank
357
Degradation of solvent mixture vapors in a biotrickling filter reactor: Impact of hydrophilic components loading and loading release dynamic JAN PACA1, ONDREJ MISIACZEK1, MARTIN HALECKY1 AND KIM JONES2 1
Institute of Chemical Technology, Department of Fermentation Chemistry and Bioengineering, Technicka 5, 160 28 Prague, Czech Republic 2 Environmental Engineering, Texas AM University-Kingsville, MSC 213, Kingsville, Texas 78363, USA
ABSTRACT Interactions amongst the degradation rates of toluene, xylenes, methyl ethyl ketone (MEK), methyl isobutyl ketone (MIBK), n-butyl acetate (n-BA), and acetone (Ac) were investigated in a biotrickling filter reactor. The reactor was packed with polypropylene High-Flow rings in a counter-current airwater mode of operation. Performance evaluation of the reactor with increased hydrophilic compound loading while maintaining a steady loading rate of hydrophobic components, were evaluated. The dynamic responses of the individual solvent components following a drop of the ketone loading rate are also described in a second phase of experiments. The degradation rate of aromatics became partially inhibited at OLKET of 15 g.m-3.h-1; below this level all of the ketones were totally degraded. Once the organic loading exceeded a value of 40 g.m-3.h-1 the removal efficiency of all the components (except n-BA) began to drop sharply. At OLKET of 85 g.m-3.h-1 the RE of aromatics dropped to below 10 %, that of acetone to 10 %, MEK and MIBK to 20 %, but n-BA removal remained above 97 %. A step-decrease of the OLKET from 85 to 5 g.m-3.h-1 resulted in a rapid increase of REAROM to 30 % (in 20 min). After the decrease, the level of REKET quickly reached 90 %, specifically: for MEK this occurred in about 4 min, for MIBK in about 25 min but for acetone, this was not achieved until after a period of 3.5 h. The significantly longer time period of REAc to achieve the original value was a consequence of: (1) its slower degradation rate resulting from a degradation competition with the other components, (2) the inhibitory effect resulting from acetone unlimited water solubility, and (3) a high quantity of acetone being accumulated in a circulating aqueous medium.
357
358
JAN PACA, ONDREJ MISIACZEK, MARTIN HALECKY AND KIM JONES
1 INTRODUCTION Paint solvents are mixtures of VOCs containing both hydrophobic and hydrophilic components. Their removal from waste air has been studied using biotrickling filters (Webster et al., 1999; Song et al., 2002; Hekmat et al., 2003; Bastos et al., 2003; Chang and Lu, 2003; Kim et al., 2004) and biofilters (Atoche and Moe, 2003, Moe and Qi, 2004; Qi et al., 2005; Moe and Qi, 2005; Qi and Moe, 2006). The objectives of this study were as follows: (1) To evaluate the simultaneous degradation of the individual hydrophobic and hydrophilic components in a paint solvent mixture. (2) To test the effect of increasing loading by only ketones on the above mentioned degradation. (3) To characterize the reactor’s dynamic response after a step-rate ketones loading decrease.
2 MATERIALS AND METHODS A schematic diagram of the bench-scale biotrickling filter is shown in Figure 1. The height of the reactor was 1.7 m and the internal diameter was 0.15 m. The sump was separated from the column by a perforated plate. The packing material consisted of Pall rings made of hydrophilized polypropylene. The parameters of Pall rings were as follows: 15 x15 x 1 mm, void volume of 0.8624, specific surface area of 350 m2.m-3 and a bulk density of 120 kg.m-3. The packed bed height was 1 m. The mixed microbial culture used to inoculate the reactor contained the following bacterial strains: Sphingobacterium multivorum (G- rods), Comamonas testosteroni (G- rods), Pseudomonas putida (G- rods) and Bacillus cereus (G+ rods). All the bacterial strains were primary toluene and xylene degraders (i.e. each strain was able to use the individual pollutants as the sole carbon and energy source for growth). Key biodegradation experimental conditions included a temperature of 22°C, a mineral medium (MM) for growth which contained 0.4 g/L (NH4)2SO4, 0.3 g/L KNO3, 0.1 g/L NaCl, 0.125 g/L K2HPO4, 0.085 g/L KH2PO4, 0.34 g/L MgCl2.6H2O, 0.02 g/L and 1 mg/L trace elements (Weigner et al., 2001). The pH of the circulating water phase was maintained at 7.0, along with a hydraulic loading rate of 0.224 m.h-1. Mixtures of the following pollutants were applied to the biological treatment systems to simulate air contamination: toluene (TOL), xylenes (XYL), methyl ethyl ketone (MEK), methyl isobutyl ketone (MIBK), n-butyl acetate (n-BA) and acetone (Ac). The pollutants in the gas phase were determined using an Agilent 6890 N gas chromatograph. Details of the conditions have been published (Paca et al., 2006). The reactor had been in operation degrading solvent mixtures for seven months when this experiment started. The reactor loading during this experiment was as follows: The ratio of TOL/XYL was kept constant at 50%w/50%w levels. The ratio of the ketones
DEGRADATION OF SOLVENT MIXTURE VAPORS IN A BIOTRICKLING FILTER REACTOR
359
Figure 1. Schematic of the trickle bed reactor set-up 1- blower, 2 - needle valve for air-flow rate control, 3 - humidifier, 4 - vessel with pollutant, 5 syringe pump, 6 – flow meter, 7 - manometer, 8 - packing, 9 - sampling ports, 10 - membrane pump, 11 - pH electrode, 12 - thermometer, 13 - NaOH solution, 14 - peristaltic pump.
inlet concentrations was kept constant: Acetone/MEK/MIBK/nBA = 1/1.13/0.93/1.43. Their inlet concentrations were changed within a range of 94 – 1500 mg.m-3. The empty bed retention time (EBRT) for the experiment was kept at 43 s.
3 RESULTS AND DISCUSSION 3.1 IMPACT OF LOADING RATE BY KETONES The experiment started on day 220 and lasted 85 days. During this time period, the constant loading rate of aromatics was 4 gc.m-3.h-1 (Fig. 2b). The reactor was loaded by gradually increasing inlet concentrations of ketones (Fig. 2a). At day 226 (A arrow), when a drop of the REAROM was observed, the syringe pump supplying a mixture of TOL and XYL into the inlet air failed. It resulted in the drop of REAROM only. At day 256 (B arrow), biomass backwashing was carried out. The biomass removal was followed by a drop of RE of both the hydrophilic and hydrophobic groups of
360
JAN PACA, ONDREJ MISIACZEK, MARTIN HALECKY AND KIM JONES
compounds. However, the recovery lasted only three days. The next decrease of both the groups of components on day 260 was a consequence of the basic nutrients starvation that was eliminated by replacing of the entire volume of aqueous medium on day 265 (C arrow).
Figure 2. Loading rates (full dots) and their impact on RE (empty dots) of hydrophilic (a) and hydrophobic (b) components.
DEGRADATION OF SOLVENT MIXTURE VAPORS IN A BIOTRICKLING FILTER REACTOR
361
Figure 3. Impact of loading rates on RE of the individual components in the solvent mixture. b – ( ) TOL; ( ) XYL a – ( ) MEK; (x) MIBK; ( ) nBA; (Δ) Ac
Increasing the OLKET from 6 to 12 gc.m-3.h-1, there was no effect on the REKET (still above 95 %) but the REAROM dropped from 85 % to 65 % (on day 245). Looking at Fig. 3 it is evident that the RE of all the individual ketones remained above 95 %. Due to preferentially degraded ketones, the RE of TOL and XYL decreased to 85 %
362
JAN PACA, ONDREJ MISIACZEK, MARTIN HALECKY AND KIM JONES
and 55 %, respectively. Further increase of OLKET to 24 gc.m-3.h-1 still resulted in a low REKET drop (above 90 % on day 48). The REAROM dropped to 35 %. Supplying the cells with a sufficient nutrient (day 50 to 63) at the OLKET of 30 gc.m-3.h-1, the gradual degrease of both aromatics as well as the ketones occurred (Fig. 3). Only the RE of n-butyl acetate still was kept above 90 %. 3.2 DYNAMIC RESPONSE TO THE STEP-DROP OF THE LOADING RATE Figure 4 shows the dynamic response following the release of ketones loading to the original value of OLKET = 7 gc.m-3.h-1 (before starting this experiment) that was carried out on day 305. As only the OLKET was dropped, and since the OLAROM was not changed the REAROM increased to 32 % (Fig. 4b). Before reaching the new steady state condition both the RETOL and REXYL showed damping oscillations lasting 4 h. The new values were RETOL = 37 % and REXYL = 28 %. The transient of the REKET response lasted 10 h. Since the REBA was not affected by the high OLKET there was no response either. The fastest increase of RE showed MEK that reached 90 % during 5 min while with MIBK it took 25 min. A completely different response was observed with acetone. Due to its unlimited water solubility, the acetone acummulated in five litres of the circulating water medium during the previous loading time period. Therefore, after the loading drop at time 0 (Fig. 4a) the REKET showed 3 h zero value. From the acetone analyses it was proven that during a time period of 2 h 40 min its concentrations in the outflow air were higher than those in the inflow air. Nevertheless, the REAc of 98 % was achieved after the four hours transient phase. Comparing the final RE values on day 305 with those on day 220 (Figs. 2 and 4) it was found that the cells were able to reach back the original RE value just for ketone degradations. After a period of the high ketone loading rates and despite of a continuous low TOL and XYL loading rates, the REAROM only achieved 32 % instead of the originally 90 %. This fact can be explained by a reduction of catabolic activity by the microbial cells in degrading the toluene and xylene hence, more slowly degraded compounds. A recovery of the degradation activity to aromatics could be achieved during 2 – 4 days, as it has previously been proven (Paca et al., 2006).
4 CONCLUSIONS • The cells prefer to dissimilate ketones to aromatic hydrocarbons. • The lack of nutrients caused a significant loss of the ability to dissimilate xylene, toluene, and acetone. A little bit milder was this effect observed with MIBK and MEK degradations. The degradation rate of nBA was not affected at all by the nutrient limitation.
DEGRADATION OF SOLVENT MIXTURE VAPORS IN A BIOTRICKLING FILTER REACTOR
363
Figure 4. Dynamic response to loading release by ketones on REKET (a) and REAROM (b). OL – full dots; RE – empty dots. Test carried out on day 305.
• High OLKET of 85 gc.m-3.h-1 suppressed the RE of all the compounds below 20 % with the exception of n-butyl acetate. • Under conditions of the breakthrough of the solvents (overloading conditions) in the biotrickling filter, acetone can be dissolved into the circulating aqueous medium to a much higher concentrations than the other
364
JAN PACA, ONDREJ MISIACZEK, MARTIN HALECKY AND KIM JONES
solvents with a limited water solubility. After a drop below the overloading conditions, the degradation of acetone, evolved from the aqueous phase, significantly prolongs the transient phase before reaching the new steady state conditions.
5 ACKNOWLEDGEMENTS This study was financially supported by the Czech Science Foundation, Joint Project 104/05/0194 and the Ministry of Education of the Czech Republic, Research Project MSM 6046137305. This study also was supported by U.S. DOE contract DE-FC3604GO14310, and partially by the National Science Foundation under Cooperative Agreement No. HRD-0206259. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.
REFERENCES Atoche, J. and Moe, W.M. (2003) Treatment of mixtures of methyl ethyl ketone (MEK) and toluene using continuous and sequencing batch operated biofilters, Proceedings of A&WMA´s 96th Ann. Conference & Exhibition, San Diego, Paper 69925. Bastos, F.S.C., Castro, P.M.L. and Jorge, R.F. (2003) Biological treatment of a contaminated gaseous emission from a paint and varnish plant – from laboratory studies to pilot-scale operation. J. Chem. Technol. Biotechnol. 78: 1201-1207. Chang, K. and Lu, C. (2003) Biofiltration of toluene and acetone mixtures by a trickle-bed air biofilter, World J. Microbiol. Biotechnol. 19: 791-798. Hekmat, D., Feuchtinger, A., Stephan, M. and Vortmeyer, D. (2003) Microbial composition and structure of a multispecies biofilm from a trickle-bed reactor used for the removal of volatile aromatic hydrocarbons from a waste gas. J. Chem. Technol. Biotechnol. 79:13-21. Kim, D., Cai, Z. and Sorial, G.A. (2004) Evaluation of trickle-bed air biofilter performance for removal of paint booth VOCs under stressed operating conditions, Proceedings of A&WMA´s 97th Ann. Conference & Exhibition, Indianopolis, Paper 36. Moe, W.M. and Qi, B. (2004) Performance of a fungal biofilter treating gas-phase solvent mixtures during intermittent loading. Wat. Res. 38: 2258-2267. Moe, W.M. and Qi, B. (2005) Biofilter treatment of volatile organic compound emissions from reformulated paint: Complex mixtures, intermittent operation, and start-up, J. Air & Waste Manage. Assoc. 55: 950-960.
DEGRADATION OF SOLVENT MIXTURE VAPORS IN A BIOTRICKLING FILTER REACTOR
365
Paca, J., Klapkova, E., Halecky, M., Jones, K. and Webster, T.S. (2006) Interactions of hydrophobic and hydrophilic solvent component degradation in an air-phase biotrickling filter reactor, Environ. Prog. 25: 365-372. Qi, B. and Moe, W.M. (2006) Performance of low pH biofilters treating a paint solvent mixture: Continuous and intermittent loading, J. Hazard. Mat. B135: 303-310. Qi, B., Moe, W.M. and Kinney, K.A. (2005) Treatment of paint spray booth off-gases in a fungal biofilter. J. Environ. Engin. 131: 180-189. Song J.H., Kinney, K.A. and Cooke, J. (2002) Effect of nitrogen avaibility on paint VOC mixtures removal. Proceedings 2002 USC-TRG Conference on Biofiltration, Reynolds Jr. F.E., Ed., The Reynolds Group, Tustin, CA, pp. 191. Webster, T.S., Togna, A.P., Guarini, W.J. and McKnight, L. (1999) Application of a biological trickling filter reactor to treat volatile organic compound emission from a spray paint booth operation, Metal Finishing. 97: 20-26. Weigner, P., Paca, J., Loskot, P., Koutsky, B. and Sobotka, M. (2001) The start-up period of styrene degrading biofilters. Folia Microbiol. 46:211-216.
This page intentionally left blank
367
Performance of peat biofilters treating ethyl acetate and toluene mixtures under non-steady-state conditions F. J. ÁLVAREZ-HORNOS, C. GABALDÓN*, V. MARTÍNEZ-SORIA, P. MARZAL AND J.M. PENYA-ROJA Department of Chemical Engineering, University of Valencia, Dr. Moliner, 50, 46100 Burjassot, Spain
ABSTRACT This paper presents the response of peat biofilters to loading changes corresponding to industrial practices such as overnight and weekend shutdowns, intermittent emission or inlet concentration peaks. Three laboratory-scale reactors fed with air contaminated with ethyl acetate, toluene or a 1:1 mixture of ethyl acetate and toluene were operated under 65 g m-3 h-1 inlet load and 60 s EBRT during 16 h/day, 5 days/week. Dynamic behavior after feed resumption after night and weekend closures showed a 1-2 h period of transient response to recover stable CO2 production values. No increase in VOC emission was observed, except for biofilters treating toluene for which a transient peak in VOC emission during 4-8 h after weekend closures was detected. More stressful conditions such as intermittent emissions (2 h-on/ 2 h-off, 16 h/day, 5 days/week), or inlet concentration peaks (40-min, 50% increase) were successfully handled in the biofilter treating only ethyl acetate; but deterioration in the operation was observed in presence of toluene. The system performance after 15-days starvation period was fully recovered in less than 8 h of re-acclimation period. Living and dead cells monitoring results are also presented.
1 INTRODUCTION Most of VOC gaseous emissions from chemical process are often generated, in practice, in transient conditions related to flux variations, and daily and weekly rotations in the production. Data indicate that short-term transient loadings within certain range of
367
368
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL AND J. M. PENYA
concentration must not cause performance problems on biofiltration (Deshusses et al., 1996), however, at some severe conditions, high contaminant emissions can be expected (Moe and Li, 2004). Research on characterization of the transient response of biofilters to remove VOCs is still scarce and adequate process monitoring and control of biofilter performance will require improved knowledge of transient loading response characteristics (Wright et al., 2005). By other side, new techniques to explore the microbiological aspects of the biofiltration have just begun to be applied, mainly due to the difficulty of making detailed observations (Steele et al., 2005). In this sense, direct cell count by staining with fluorochromes that distinguish between living and dead bacteria reveals greater bacteria concentrations than the plate count technique, since culture methods exclude any organisms that are not able to grow on the culture media (Tresse et al., 2003). The purpose of the present research was to investigate the biofiltration of ethyl acetate, toluene and the mixture of both pollutants, taking into consideration the following objectives: (1) to evaluate the response of the biofilters to intermittent feeding conditions (16-h on/8-h off, 5 days/week), long-term starvation (2-weeks without VOC feeding), and pulse-step changes in concentration; (2) to evaluate the possible inhibition effect of ethyl acetate on toluene removal; and (3) to determine the dynamics of living and dead cells in the biofilters for better understanding the link between operational and ecological aspects of the biofiltration.
2 MATERIALS AND METHODS 2.1 BIOFILTER SYSTEM The biodegradation was carried out in three laboratory-scale biofilters (total length of 97 cm and an internal diameter of 13.6 cm) treating air polluted with ethyl acetate, toluene or with a 1:1 mixture of both pollutants. The biofilters were equipped with five sampling ports to measure gas concentrations, located at 0 (inlet port), 25, 50, 75, and 95 (outlet port) cm of column length. Additional ports located at 20, 40, 60 and 80 cm were used for temperature measurement and to recover filter bed samples. Peat (ProEco Ambiente, Spain) was used as filter material, and inoculation with an adapted culture from activated sludge was performed; details were described elsewhere (Álvarez-Hornos et al., 2006). Compressed, filtered and dried air was passed through a humidifier to assure a relative humidity value of at least 90 %. The empty bed residence time (EBRT) was adjusted to 60 s by using mass flow controllers (Bronkhorst Hi-Tec, Nederlands). VOC feed was introduced into the air stream by using a syringe pump (New Era, infusion/withdraw NE 1000 model, USA) and then, the contaminated air was flowed downwards into the bed. The strategy adopted to control the moisture content in the biofilter bed was by feeding the air previously
PERFORMANCE OF PEAT BIOFILTERS TREATING ETHYL ACETATE
369
humidified, as well as pouring on top of the biofilter 250-500 mL of a pH-buffered and nutrient solution (5.23 g/L K2HPO4, 0.30 g/L KH2PO4, 3.00 g/L NH4Cl, pH = 8.00) each 2-3 days. 2.2 OPERATIONAL CONDITIONS The biofilters were operated in parallel at 65 g m-3 h-1 of continuous IL (inlet load) for more than three months (phase 0) before starting intermittent loading experiments. Several phases were consecutively performed: – Phase I: To simulate shift work of many industrial facilities, intermittent loading during 16 h/day, 5 days/week at identical instantaneous IL than in phase 0 was applied. Uncontaminated air during night and weekend closures was supplied. – Phase II: To simulate intermittent pattern emission, 4 cycles of 2 h-on/2 hoff during 5 days/week were performed at an IL of 130 g m-3 h-1 during VOC feeding. – Phase III: A long-term starvation period of 15 days without VOC loading, but with air flow through the biofilters, was applied. – Phase IV: After time, intermittent loading similar to phase 1 was carried out to evaluate the re-acclimation of the biofilters under discontinuous operation. In addition, shock loading experiments to study the dynamic response of the biofilters to inlet concentration peaks were carried out. Each biofilter was exposed on three consecutive days to 40-min step of a 50% inlet concentration higher than the regular feed. 2.3 ANALYTICAL TECHNIQUES VOC concentration profiles were monitored daily by using a gas chromatograph (GC 8000 model, CE Instruments, Spain) equipped with a 0.86 mL automated gas valve injection system and a flame ionization detector. The chromatographic packed column was 10 % SP-1000 on 80/100 SUPELCOPORT, 10’ (ID 1/8"). The gas carrier was helium (35 cm-3 min-1) and the temperatures of the injection port, oven and detection port were 200, 60 and 250 ºC, respectively. Response of the systems after feed resumption was evaluated, in terms of total VOC concentration, by using a total hydrocarbon analyzer (Nira Mercury 901 model, Spirax-Sarco, Spain). CO 2 concentrations in the five ports were measured by using a NDIR carbon dioxide analyzer (Gaswork NDIR model, Seda, Spain). The three equipments were calibrated by using standard gaseous mixtures supplied by Carburos Metálicos, Spain. Bacteria were enumerated by fluorescence microscopy using LIVE/DEAD® BacLightTM Bacterial Viability Kit (Invitrogen, USA), thus co-staining was performed
370
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL AND J. M. PENYA
with SYTO 9 to define the number of living cells and with propidium iodine to define the dead ones. Samples were taken from the four sampling ports each 1-2 weeks. Microorganisms were dispersed from peat, stained and recovered onto membrane microfilters, as described in Álvarez-Hornos et al. (2007). Filters were mounted on microscope slides in mounting media (Invitrogen) and examined using an epifluorescence microscope (Nikon Eclipse E800, Japan) equipped with a blue excitation filter (B-2A; ex 450-490 nm, dm 505 nm, ba 520 nm) and a green excitation filter (G-2A; ex 510-560 nm, dm 575 nm, ba 590 nm). Living and dead cells were enumerated counting ten random microscopic fields three times, with an average standard deviation of 8.7% (maximum value of 37.3%).
3 RESULTS AND DISCUSSION 3.1 PERFORMANCE OF BIOFILTERS The results of the biofilters monitoring are presented in Figure 1. For all biofilters, similar RE (removal efficiency) was achieved under intermittent loading than in continuous one, demonstrating the process capacity to handle shift work. However, the response of biofiltration process to shutdown periods depended on the pollutant and on the operational conditions. For pure ethyl acetate (Figure 1a), deterioration in global performance was not detected for overnight and weekend closures. For pure toluene, greater penetration into the bed was obtained after weekend closures (Figure 1b, days 6, 13, 20, 27, 34, 41). The presence of ethyl acetate resulted in a decrease of toluene RE (Figure 1c), especially high toluene penetration was observed after weekend closures (days 4, 11, 18). Under 2 h-on/2 h-off cycling pattern (phase II), only ethyl acetate was successfully degraded, although the contaminant penetrated deeper into the filter bed, showing the adverse response of the biological activity as off-time period and instantaneous IL increased. For pure toluene, RE decreased to values around 70% – 75%, with outlet toluene concentrations between 500 and 650 mg m-3, then the test was stopped. From the first day of the operation after 15 days of starvation (phase IV), RE in the three biofilters was restored to the values obtained in phase I at same feeding conditions. Biomass re-acclimation was shorter than 6 h, demonstrating the feasibility of the peat biofilters to work under prolonged cut-off. Cox and Deshusses (2002) have reported 10-24 h of re-acclimation for 2-9 days of starvation in a biotrickling filter treating toluene.
PERFORMANCE OF PEAT BIOFILTERS TREATING ETHYL ACETATE
371
Figure 1. Monitoring biofilters performance. VOC concentration at the inlet (), first quarter of the biofilter ( ) and outlet (x); overall RE (Δ) and RE at the first quarter of the biofilter (—).
3.2 EFFECT OF STARVATION ON BIOLOGICAL ACTIVITY Determination of the carbon dioxide emissions after feed resumption is a fast respirometric measurement for the assessment of the effect of starvation periods on biological activity. Carbon dioxide production and outlet VOC concentration were monitored after night and weekend closures in phases I and II (Figure 2). For the mixture biofiltration, chromatographic analysis demonstrated that outlet emission was mainly composed of toluene (95% ± 5%). Data from carbon dioxide production after overnight closures, plotted in Figures 2a, 2b and 2d showed similar tendency: after a lag of about 15-30 min, carbon dioxide production increased from values corresponding to endogenous metabolism to the stable total CO2 production in less than 1-2 h, with the main increase observed from 30 min to 1 h period. For pure ethyl acetate, VOC emissions were always lower than 10 mgC m-3 (not shown). For feeds containing toluene (Figures 2c and 2e), no VOC emissions were observed in the first 30 min: toluene adsorption onto the biofilter might have occurred; and then a continuous increase during 1 h was observed until stationary values were reached.
372
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL AND J. M. PENYA
For pure ethyl acetate, weekend closures presented similar CO2 production pattern to overnight ones, although lag-time enlarged to 40-60 min. With toluene, greater time (4 - 6 h) was needed to fully restore the active metabolism, especially for the first weekend feed cut-off (day 6 for pure toluene, day 4 for the mixture). Microbial populations shifted in metabolism and more re-acclimation time to induce the toluene degradation pathway was needed. Besides, high toluene emissions with maximum concentrations corresponding to 30%-50% of toluene inlet concentration were produced in the first 2-4 h after feed resumption, and more than 4-8 h were needed to recover RE. Kim et al. (2005) reported a minimum time of 100-200 min to recover the RE of a biofilter fed with 0.2-2 g m-3 of toluene operated under 2 days/week starvation.
Figure 2. Immediate response of biofilters after night and weekend closure periods. (a) pure ethyl acetate, (b), (c) pure toluene; (d), (e) 1:1 ethyl acetate: toluene mixture.
PERFORMANCE OF PEAT BIOFILTERS TREATING ETHYL ACETATE
373
Figure 3. Immediate response of biofilters after 15-days feed cut-off, phase III. (a) pure ethyl acetate, (b), (c) pure toluene; (d), (e) 1:1 ethyl acetate: toluene mixture.
Results of monitoring the outlet CO2 and total VOC concentration after the long-term starvation period of 15 days (phase IV) are presented in Figure 3. CO2 production and outlet VOC patterns were similar to those observed previously at identical operational conditions (phase I, Figure 2) after night and weekend closures. This observation demonstrates that by using an organic support, with unpolluted air supply and moisture control, the environmental conditions and nutrients supply were enough to assure the bacterial population survival under endogenous metabolism for more than 15-days cut-off period. The evolution of CO2 production immediately after re-startup after 15-days of starvation was almost identical to a weekend closure. However, greater toluene breakthroughs were observed, with maximum outlet concentrations of 92% and 82% of toluene inlet concentration for pure toluene and
374
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL AND J. M. PENYA
the mixture, respectively. The period of 6 h to restore RE was not enlarged after 15 days of starvation in comparison with weekend closures. These data also indicate that the presence of ethyl acetate did not adversely affect the activation of the toluene degradation mechanisms after two weeks of starvation. These results are of great interest to demonstrate the feasibility of the biofiltration process to adequately respond to usual non-use periods related to holiday breaks at industrial sites. 3.3 DYNAMICS OF LIVING AND DEAD CELLS Colonization dynamics monitoring in the four sections of the biofilters resulted in quite stable living bacteria density for the whole experimentation period, including the re-starting feed phase IV. Average living cells per gram of dry peat were 1.5×1010 (6.0×109 st. dev.) for the biofilter treating pure ethyl acetate, 1.3×1010 (4.5×109 st. dev.) for pure toluene, and 1.4×1010 (6.5×109 st. dev.) for the biofilter treating the mixture. Results suggest that a healthy microbial population was maintained during the whole intermittent loading experiment. For pure toluene and the mixture, dead cell percentages maintained in high values around 76% from the beginning until phase III starting. In case of ethyl acetate, dead cell percentage suddenly increased from 34% to 75% when feed was changed from 16 h/d on to 2 h-on/2h-off cycles (phase II), and it remained quite constant until phase III starting. After 15-d of starvation (phase IV), a reduction in total cell density was observed, yielding average dead cell percentages of about 41%, 64% and 41% for the biofilters treating pure ethyl acetate, toluene and the mixture, respectively. This observation suggests that living cells utilized dead ones as main carbon source in the long-term non-fed period. Results corroborated that non-use periods can be considered as a means of biomass control as previously reported by Kim et al. (2005) for low toluene loading biofiltration. 3.4 RESPONSE OF THE BIOFILTERS TO INLET CONCENTRATION PEAK Here, transient response of biofilters to stepwise variation in inlet concentration, common problem that may occur in industrial sites, is examined. For each biofilter, inlet concentration was increased by 50% for 40-min on three consecutive days. Results from monitoring total VOC concentration at the first quarter (for pure ethyl acetate, no emission were detected at the outlet) or at the outlet (for pure toluene and for the mixture) and total CO2 production are plotted in Figure 4. Chromatographic analysis indicated the absence of ethyl acetate at the outlet for the biofilter treating the mixture. For each biofilter, similar evolution was observed for the three days. The biofilter treating ethyl acetate (Figure 4a) was able to fully assimilate the transient peak working in 2 h-on/2 h-off cycles. Nevertheless, an increase in the pollutant breakthrough at the first quarter of the biofilter was rapidly developed and stabilized in less than 15-min
PERFORMANCE OF PEAT BIOFILTERS TREATING ETHYL ACETATE
375
from the peak start. In case of toluene, biofilters were not able to assimilate the transient peak in smoother conditions (16 h/d-on stage). An increase in the pollutant emission was rapidly developed (< 10-min) and average maximum emissions detected at the peak end were 307 mg-C m-3 (9% st. dev.) for pure toluene and 419 mg-C m-3 (14% st. dev.) for the mixture feed. For the three biofilters, return to the initial feed condition at 40-min caused a sudden recovery (< 10-20-min) of the VOC concentration. In all cases, increases in CO2 production were also observed. The 20-min gap between VOC outlet peak end and maximum CO2 emissions indicated that microbial population needed some time to activate their metabolism under higher concentration of carbon source.
Figure 4. Response to inlet concentration peaks. Total VOC concentration at the inlet (—). Solid symbols represent total CO2 production. Open symbols indicate total VOC concentration at first quarter for pure ethyl acetate (a), and at the outlet for toluene (b) and the mixture (c).
376
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL AND J. M. PENYA
4 CONCLUSIONS This study demonstrates the capacity of peat biofilters to handle intermittent loading conditions. Similar removal was achieved operating in a 16 h/day, 5 days/week regular feed mode (at 65 g m-3h-1 of instantaneous IL and 60 s of EBRT) than in continuous loading. The presence of ethyl acetate affected adversely to toluene RE, resulting in greater toluene penetration into the bed in comparison with pure toluene biofiltration. Some differences between toluene and ethyl acetate, more easily biodegradable compound, can be pointed out: – Treating pure ethyl acetate it was possible to achieve low emissions even when more severe conditions (2-h on/2-h off cyclic loading during 16 h/d) were applied. – When inlet concentration peaks (40-min, 50% increase) were supplied, the presence of toluene resulted in high VOC emissions; only the biofilter fed with pure ethyl acetate was able to maintain complete removal efficiency. – For the three biofilters, night closure did not affect the RE, and CO 2 productions were restored in stable values in less than 1 - 2 h. But restart-up after weekend closures resulted in high toluene emission in the first 6 h, while VOC emission was not detected in case of pure ethyl acetate biofiltration. – After 15-days of starvation, CO2 production indicated the recovery of the active metabolism in less than 6 h with full restoration of the RE with toluene. In case of pure ethyl acetate, complete RE was obtained from the beginning, and the CO2 production was restored in less than 2 h. – For the three biofilters, living cell density remained quite stable for the whole experimentation period. Besides, after 15-days of starvation period, dead cell percentages decreased, especially for the biofilter treating pure ethyl acetate.
5 ACKNOWLEDGEMENTS Financial support by Ministerio de Educación y Ciencia, Spain (research project CTM 2004-05714-C02-01/TECNO with FEDER funds) is acknowledged. F. Javier ÁlvarezHornos has a FPU grant from Ministerio de Educación y Ciencia, Spain.
PERFORMANCE OF PEAT BIOFILTERS TREATING ETHYL ACETATE
377
REFERENCES Álvarez-Hornos, F.J., Gabaldón, C., Martinez-Soria, V., Marzal, P., Penya-Roja, J.M. and Izquierdo, M. (2007) Long-term performance of peat biofilters treating ethyl acetate, toluene, and its mixture in air. Biotechnol. Bioeng. 96(4): 651-660. Cox, H.H.J. and Deshusses, M.A. (2002) Effect of starvation on the performance and re-acclimation of biotrickling filters for air pollution control. Environ. Sci. Technol. 36(14): 3069-3073. Deshusses, M.A., Hamer, G. and Dunn, I.J. (1996) Transient-state behavior of a biofilter removing mixtures of vapors of MEK and MIBK from air. Biotechnol. Bioeng. 49(5): 587-598. Kim, D., Cai, Z. and Sorial, G.A. (2005) Behavior of trickle-bed air biofilter for toluene removal: effect of non-use periods. Environ. Progress 24(2): 155-161. Moe, W.M. and Li, C. (2004) Comparison of continuous and sequencing batch operated gas-phase biofilters for treatment of MEK. J. Environ. Eng. 130(3): 300-314. Steele, J.A., Ozis, F., Fuhrman, J.A. and Devinny, J.S. (2005) Structure of microbial communities in ethanol biofilters. Chem. Eng. J. 113(2-3): 135-143. Tresse, O., Lescob, A. and Rho, D. (2003) Dynamics of living and dead bacterial cells within a mixed-species biofilm during toluene degradation in a biotrickling filter. J. Appl. Microbiol. 94(5): 849-855. Wright, W.F., Schroeder, E.D. and Chang, D.P.Y. (2005) Regular transient loading response in a vapour-phase flow-direction-switching biofilter. J. Environ. Eng. 131(2): 1649-1658.
This page intentionally left blank
379
Characterization of a biotrickling filter treating methanol vapours A. ÁVALOS RAMÍREZ, J.P. JONES, AND M. HEITZ Department of Chemical Engineering, Faculty of Engineering, Université de Sherbrooke, 2500, boulevard de l’Université, Sherbrooke (Québec), J1K 2R1, Canada
ABSTRACT The aim of this research is to characterize a biotrickling filter (BTF) treating methanol vapour emissions. The parameters studied were the nitrogen concentration in the nutrient solution and the empty bed residence time (EBRT). The effect of continuously recycling the nutrient solution was also analyzed. At nitrogen concentrations as low as 0.001 gN L-1, the BTF presented removal efficiencies higher than 70 % for an inlet load of 110 g m-3 h-1. A nitrogen concentration of 0.005 gN L-1 was used to study the effect of EBRT and the continuous recirculation of nutrient solution. At a constant methanol inlet concentration of 1500 ppmv, the BTF was operated in a range of EBRT from 20 to 265 s and the removal efficiencies respectively attained were 40 and 90 %. Methanol vapours were absorbed into the lixiviate and were taken into account in analysing the results.
1 INTRODUCTION Methanol is a major pollutant emitted to the atmosphere in Canada. Methanol is toxic to human health and depending on exposure it causes headache, sleep disorders, gastrointestinal problems and optic nerve damage (OPPT, 2006). In the environment, methanol is related to problems like smog generation (Monod et al., 1998). There are chemical, physical and biotechnological treatments which can be used for controlling methanol vapour emissions. Among them, the biotrickling filter is of considerable interest for effluents with low pollutant concentration and high volumetric flow rates (Cooper and Alley, 2002). The BTF advantageous characteristics are: no production of hazardous wastes, low energy consumption and low operation costs (Jorio and
379
380
A. AVALOS RAMÍNEZ, J. P. JONES AND M. HEITZ
Heitz, 1999, Delhoménie and Heitz, 2005). When the pollutant is water-soluble, e. g. ethanol and methanol, the BTF is able to operate in a wide range of inlet concentration (Avalos Ramírez et al., 2007; Avalos Ramírez et al., 2005). There are few studies which discuss methanol emissions control by using BTF (Avalos Ramírez et al., 2005; Prado et al., 2004). The aim of the present study is to determine the effect of nitrogen concentration in nutrient solution, the empty bed residence time (EBRT) and the role of the lixiviate on the performance of a BTF treating methanol vapour emissions.
2 MATERIALS AND METHODS The experiments were performed using three identical biotrickling filters. The experimental setup is shown on Figure 1. The setup consisted of a bubbler containing methanol, a humidification column, a biotrickling filter with an empty bed volume of 0.018 m3, a holding tank and a recycling pump. The nutrient liquid solution was fed into the BTF countercourant to the air flow. The concentrations of methanol were measured continuously with a total hydrocarbon analyzer (Horiba FIA-510, Horiba Instruments Inc., Irvine, CA, USA). The concentrations of carbon dioxide in air were measured continuously with a portable gas analyzer (Ultramat 22P, Siemens AG, Munich, GE).
Figure 1. Schematic representation of biotrickling filter.
CHARACTERIZATION OF A BIOTRICKLING FILTER TREATING METHANOL VAPOURS
381
3 RESULTS AND DISCUSSION 3.1 EFFECT OF NITROGEN CONCENTRATION ON REMOVAL EFFICIENCY Figure 2 shows the BTF removal efficiency as a function of the nitrogen concentration in the nutrient solution at a methanol inlet concentration of 1500 ppmv and an EBRT of 65 s. Urea was added to tap water in order to vary the nitrogen concentration in nutrient solution from 0.0 to 0.1 gN L-1. Since tap water contains traces of nutrients, for example 0.0002 gN-NH3 L-1, the BTF still presented a removal efficiency of 70% without additional nutrients. Removal efficiency increased with nitrogen concentration passing from 70% at 0.0 gN L-1 to 90% at 0.1 gN L-1. As shown in Figure 2, the removal efficiency increased from 70 to 80 % when nitrogen
Figure 2. Variation of removal efficiency with nitrogen concentration in nutrient solution at an EBRT of 65 s and a methanol inlet concentration of 1500 ppmv.
Figure 3. Variation of elimination capacity and carbon dioxide production rate with empty bed residence time, at a methanol inlet concentration of 1500 ppmv and nitrogen concentration in nutrient solution of 0.005 gN L-1. ( ) inlet load, ( ) elimination capacity, (Δ) carbon dioxide production rate.
382
A. AVALOS RAMÍNEZ, J. P. JONES AND M. HEITZ
concentration passed from 0.0 to 0.01 gN L-1. When nitrogen concentration increased 10 fold, from 0.01 to 0.1 gN L-1, the removal efficiency only increased from 80 to 90%. BTF appears to be an appropriate treatment for controlling methanol vapour emissions at small concentrations of nitrogen. 3.2 EFFECT
OF EMPTY BED RESIDENCE TIME ON CARBON DIOXIDE PRODUCTION RATE AND
ELIMINATION CAPACITY
In order to study the effect of EBRT and the role of the lixiviate in the performance of the BTF, a nitrogen concentration of 0.005 gN L-1 was chosen. The BTF was operated at five EBRTs: 20, 30, 65, 130 and 265 s. The methanol inlet concentration was maintained at 1500 ppmv for all experiments. Figure 3 shows the variation of elimination capacity with EBRT. Elimination capacity (EC) decreased with EBRT, from 135 g m-3 h-1 at 20 s to 25 g m-3 h-1 at 265 s. However, removal efficiency increased with EBRT, from 40 to 90%. This can be appreciated in Figure 3, as EBRT increased, the EC tended to the value of methanol inlet load (IL). The BTF presented a competitive performance at EBRT higher or equal to 65 s. Carbon dioxide production rate (PCO2) decreased with EBRT from 13.0 to 9.5 g m-3 h-1.
Figure 4. Variation of methanol concentration in lixiviate with empty bed residence time at a methanol inlet concentration of 1500 ppmv and nitrogen concentration in nutrient solution of 0.005 gN L-1.
3.3 ROLE OF LIXIVIATE ON METHANOL REMOVAL Figure 4 shows that the methanol concentration in lixiviate decreased with EBRT. It passed from 4.5 g L-1 at 20 s to 0.5 g L-1 at 265 s. Since nutrient solution was renewed daily, absorption in the lixiviate was a mechanism for methanol removal. By comparing the curve of IL in Figure 3 with that of methanol concentration in lixiviate
CHARACTERIZATION OF A BIOTRICKLING FILTER TREATING METHANOL VAPOURS
383
in Figure 4, it is evident that there is a direct relationship between the quantity of methanol in the inlet air flow rate to the methanol absorbed by the lixiviate. So that, the renewing of lixiviate is a way to remove methanol vapours from an air stream. The lixiviate in the present study contributed to obtain high removal efficiencies, especially at low nitrogen concentrations.
4 CONCLUSION This study shows that the biotrickling filter is appropriate for controlling methanol vapour emissions at low nitrogen concentrations in nutrient solution. The BTF presents removal efficiencies of at least 70 % for concentrations as low as 0.0002 gN L-1 (tap water). The empty bed residence time greatly influences the BTF performance. BTF produces more carbon dioxide at small EBRTs. The lixiviate contributed to removal of methanol from the air stream.
5 ACKNOWLEDGEMENTS The authors express their sincere acknowledgements to Natural Sciences and Engineering Research Council of Canada for the financial support.
REFERENCES Avalos Ramírez, A., Jones, J. P. and Heitz, M. (2007) Biotrickling filtration of air contaminated with ethanol. J. Chem. Technol. Biotechnol. 82: 149-157. Avalos Ramírez, A., Jones, J.P. and Heitz, M. (2005) Methanol and ethanol treatment by biotrickling filtration: an experimental study. Proceedings of the international congress biotechniques for air pollution control. Universidade da Coruña, La Coruña, Spain, pp. 385-389. Cooper, C. D. and Alley, F. C. (2002) Air pollution Control, A design approach, 3rd edition, Waveland Press, Prospect Heights, p. 429-442. Delhoménie, M.C. and Heitz, M. (2005) Biofiltration of air: A review. Crit. Rev. Biotechnol. 25: 53-72. Jorio, H. and Heitz, M. (1999) Traitement de l’air par biofiltration. Can. J. Civ. Eng. 26: 402-424. Monod, A., Doussin, J.F., Chebbi, A., Carlier, P., 1998. Transformations chimiques des COV dans la troposphère. Impact sur la qualité de l’air. In: Le Cloirec, P. (Ed.). Les composés organiques volatils (COV) dans l’environnement. Technique & Documentation, Paris, France, pp. 119-162.
384
A. AVALOS RAMÍNEZ, J. P. JONES AND M. HEITZ
OPPT (Office of Pollution Prevention and Toxics), 2006. Chemicals in the Environment: Methanol (CAS NO. 67-56-1). Paper No. EPA 749-F-94-013. Environmental Protection Agency, US. Prado, O.J., Veiga, M.C. and Kennes, C. (2004) Biofiltration of waste gases containing a mixture of formaldehyde and methanol. Appl. Microbiol. Biotechnol. 65: 235-242.
385
Performance evaluation of fungal biofilters packed with Pall rings, lava rock, and perlite for α-pinene removal YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES Chemical Engineering Laboratory, Faculty of Sciences, University of La Coruña, Rúa Alejandro de la Sota, 1, 15008 – La Coruña, Spain
ABSTRACT The most suitable packing material for biofiltration of α-pinene was selected among Pall rings, lava rock, and perlite. In the present study, several biofilters fed α-pinene-polluted air were inoculated with a new fungal isolate of Ophiostoma stenoceras. The biofilters were packed either with lava rock or Pall rings alone or with a mixture of perlite and Pall rings. During the approximately 9 months operation, α-pinene’s removal efficiency, pressure drops, pH dependence and removal profile were evaluated. α-Pinene removal efficiencies were above 93.8%, 79.4% and 58.6% at an inlet loading rate of 100 g.m-3.h-1 in the biofilters packed with Pall rings, lava rock, and a mixture of Pall rings and perlite, respectively. The fungus preferred to grow in lava rock and the mixture of Pall rings and perlite instead of Pall rings alone. Moreover, with sufficient nutrients and buffer solution, the biofilter packed with the mixture of Pall rings and perlite reached the highest elimination capacity compared to the other two packing materials. The pressure drop of the biofilter packed with the mixture of Pall rings and perlite did not exceed 11 mm H2O.m-1. The low-pressure drop reached when using the mixture of Pall rings and perlite as packing material allows to conveniently prevent clogging and channeling problems often associated with conventional biofilter operations.
1 INTRODUCTION Alpha-pinene is one of the major hydrophobic organic compounds, which is emitted by the forest product industries, pulp and paper industries, and fragrance manufacturers in the form of monoterpene. These emissions into the atmosphere result in the formation
385
386
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
of particulates and free radicals. The former form a blue haze and reduce visibility and the latter deplete the ozone layer. Since the compound is emitted in low concentrations of, usually, less than 300 ppm, treatment of this waste gas by biofiltration is expected to be cost-effective. Biofiltration is an established technology for air pollution control and the alternative of choice to conventional physico-chemical treatment techniques (Kennes and Veiga, 2001). Biofiltration is a promising technology involving the flow of a polluted air stream through a packed-bed containing microorganisms that are able to degrade pollutants into harmless products. Several studies have been carried out using biofilters based on the action of bacteria. As a result of the low solubility of α-pinene in water (2.5 ppm at 23 °C), that compound is poorly absorbed by the bacterial biofilms. In addition, acidification and drying out of the filter bed often cause biofilter failure. This is why a fungal biofilter was chosen in the present study. For α-pinene abatement, filamentous fungi were isolated from biofilters operated in our laboratory. Fungi develop hyphae which provide a large surface area in contact with the gas phase so that a direct efficient mass transfer from the gas phase to the biological aqueous phase is possible. This allows a faster uptake of hydrophobic compounds than in flat aqueous bacterial biofilms. Furthermore, fungi are generally tolerant to low water activities and a low pH, so that these parameters do not need to be strictly monitored in the biofilters (Kennes and Veiga, 2004). While biofiltration has emerged as an attractive technique in the treatment of waste gases, it is not completely free of problems and still needs to be further optimized. One potential problem is a high pressure drop which causes high energy consumptions. Conventional biofilters are usually packed with natural carriers, such as compost, peat or soil. They decay over time, causes compaction, clogging, short circuiting and increased headloss across the bed. Therefore, the filter bed usually requires blending with some inert materials to prevent this from happening. Polystyrene, gypsum, perlite, wood chips, and branches have been used as inert materials to be blended into the bed (Kennes and Thalasso, 1998; Devinny et al., 1999; Kennes and Veiga, 2001). The oxidation of sulfur, nitrogen, and chlorine-containing compounds produces acidic intermediates or end products, which lowers the bed’s pH and subsequently reduces the efficiency of VOCs removal. Calcium carbonate, marl, and oyster shells have been used to neutralize acid products (Ottengraf and van den Oever, 1983). More recently, inert and synthetic filter beds have been used for biofiltration. Inert carriers present several advantages such as a long lifetime and low pressure drop. Besides, they are physically and chemically inert, when compared to conventional natural media (Kennes and Veiga, 2002). In our previous studies, we have isolated a new strain of Ophiostoma stenoceras which proved to be very effective in removing α-pinene. Therefore, Ophiostoma stenoceras was used as a biofilter inoculum and different packing materials such as
PERFORMANCE EVALUATION OF FUNGAL BIOFILTERS PACKED WITH PALL RING
387
Pall rings, lava rock, and perlite were used as biofilter bed. The biofilters were operated under similar conditions and their capabilities to degrade α-pinene were measured and compared. In addition, possible operational problems such as a high pressure drop, as well as the optimization of operational parameters were also evaluated.
2 MATERIALS AND METHODS 2.1 PACKING MATERIAL In choosing packing materials, our concern was to provide a large surface area for microbial adhesion and efficient mass transfer, along with a minimal pressure drop within the column. Microbial compatibility, low cost, and readily availability were additional considerations. The packing materials tested were either Pall rings, lava rock, and perlite, alone or as mixtures. The polypropylene Pall rings were obtained from VFF GmbH & Co (Germany). The lava rock in this study was bought from Burés S.A., (Spain). The macroporous volcanic stone is commonly crushed and sieved for use in decorative landscaping. The perlite was manufactured from Otavi Ibérica S.L.u. (Spain). Perlite is an expanded mineral with high porosity consisting of SiO2 as the main component. Table 1 summarizes some properties of all three packing materials.
Table 1. Characteristics of the filter bed materials used in the experiment. Packing Pall rings Lava rock Perlite
Density (kg.m-3) 80 866.7 94.5
Void space (%) 91 50 40
Size (mm) 15 4-10 4-6
Specific surface area (m2.g-1) 350 m2.m-3 0.55 8.75
2.2 MEDIA COMPOSITION Batch experiments were undertaken with an aqueous culture medium containing (per liter) (Estévez et al., 2005): 4.5 g KH2PO4, 0.5 g K2HPO4, 2.0 g NH4Cl and 0.1 g MgSO4.H2O. The culture medium was autoclaved at 120 °C for 20 min before adding filter-sterilized solutions of vitamins and trace minerals. The composition of the vitamins solution was (per liter): 0.2 g thiamine·HCl, 0.1 g riboflavin, 1.0 g nicotinic acid, 2.0 g Ca-pantothenate, 0.1 g biotin, 0.1 g thioctic acid, 0.1 g folic acid and 0.25 g pyridoxine HCl. The composition of the trace minerals solution was (per liter): 120 mg FeCl3, 50 mg H3BO3, 10 mg CuSO4.5H2O, 10 mg KI, 45 mg MnSO4.H2O, 20 mg
388
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
Na2MoO4.2H2O, 75 mg ZnSO4.7H2O, 50 mg CoCl2.6H2O, 20 mg KAl(SO4)2.12H2O, 13.25 mg CaCl2.2H2O and 10,000 mg NaCl. The original pH of that medium was 5.9. Stock cultures of the fungus were maintained on petri dishes or on slants using either Potato Dextrose Agar (PDA) or the same mineral medium as described above supplemented with 16 g agar.l-1. When using the mineral medium, the plates were incubated in a tank at 30 °C, in the presence of α-pinene vapors as sole carbon source. Stock cultures on PDA were stored in a refrigerator at 4 °C. 2.3 ENRICHMENT AND ISOLATION OF THE α-PINENE DEGRADER The α-pinene degrading fungus used in this study was obtained from the leachate of a biofilter treating toluene. 10 ml of the liquid was suspended in 90 ml mineral medium as described elsewhere (Estévez et al., 2005). α-Pinene was added as the only source of carbon and energy. Erlenmeyer flasks with a 5:1 headspace/liquid ratio were closed with Teflon wrapped rubber stoppers and were incubated in a rotary shaker (150 rpm) at 35 °C. The flasks were aerated daily, and α-pinene was added as needed. After several serial transfers, a stable microbial consortium developed (Jin et al., 2006). Individual members of the consortium were isolated by streaking on mineral agar medium and incubation under solvent vapor. The isolated strain was identified as Ophiostoma stenoceras by the Centraal Bureau voor Schimmelcultures (The Netherlands). For the preparation of cell suspensions, the fungus was cultured for 10-12 days in 100 mL mineral medium in 500 mL flask at 35 °C with shaking at 150 rpm. The bottle was sealed with a Teflon-lined screw cap, and 15 ìL α-pinene was added to the medium. After the culture had degraded six additions of α-pinene, it was transferred to a 5 L bottle containing 2 L nutrient medium. After the culture had degraded three 0.5 mL additions of α-pinene, it was recirculated through the packed bed bioreactor using a peristaltic pump (model 323E/D, Watson-Marlow Ltd, Falmouth, Cornwall, UK) at a rate of 0.5 L.min-1 for 24 hours in order to allow the biomass to attach to the support material. 2.4 EXPERIMENTAL SETUP The schematic of the biofilters used in this study is shown in Figure 1. All the bioreactors are cylindrical packed bed reactors made of glass, with different dimensions. Bioreactor 1 and 2 were 75 mm in diameter and 700 mm in height. The active height of packing column, filled with different packing materials were 25cm and 65cm for lava rock (Bioreactor 1) and Pall rings (Bioreactor 2), respectively. The working volumes of Bioreactor 1 and Bioreactor 2 were 1.25 L and 2.78 L. Biofilter 3, packed with the mixture of Pall rings and perlite, consisted of a cylindrical glass column with an inner diameter of 10 cm and a total height of 70 cm. The length of the biofilter bed
PERFORMANCE EVALUATION OF FUNGAL BIOFILTERS PACKED WITH PALL RING
389
was 60 cm, leading to a working volume of approximately 4.71 L. All fittings, connections and tubings were made of teflon. A large stream of compressed air was humidified up to 97% relative humidity by passing it through a tower humidified with water. Another smaller stream of air was bubbled through a vial containing pure α-pinene and was mixed with the larger humidified gas stream. Gas phase α-pinene concentrations ranging from 0 to 460 ppm were obtained by changing the relative flow rates of the gas streams. The resulting synthetic waste gas was introduced through the top of the column (co-current flow). An aqueous mineral medium was recirculated over the packed bed once a week in order to add fresh nutrients and remove the accumulated metabolites. The pH of the leachate was measured on a regular basis.
Figure 1. Schematic of the laboratory scale fungal biofilter.
2.5 ANALYTICAL METHODS Gas phase concentrations of α-pinene in the biofilters were measured by gas chromatography using a Hewlett-Packard 5890 series II chromatograph. The GC was equipped with a flame ionization detector (FID). The flow rates were 30 mL.min-1 for H2 and 300 mL.min-1 for air. The inlet and outlet streams were sampled, as well as air aliquots taken at different reactor heights. The GC was equipped with a 50 m TRACER column (TR-WAX, internal diameter 0.32 mm, film thickness 1.2 μm) and Helium was used as the carrier gas (flow rate 2.0 mL.min-1). The α-pinene concentration was determined at an oven temperature of 120 °C and using a FID at 250 °C. Similarly, CO2 concentrations were measured on another Hewlett-Packard 5890 series II GC
390
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
equipped with a thermal conductivity detector (TCD). The CO2 concentrations were determined at an injection temperature of 90 °C, an oven temperature of 25 °C and using a TCD at 100 °C.
3 RESULTS AND DISCUSSION 3.1 START-UP PERIOD During the start-up period, the inlet loading rate of α-pinene was maintained around 15 g.m-3 .h-1 in all reactors. It took several weeks before Ophiostoma stenoceras had grown and attached enough to the packing material. Almost 28 days were needed before complete removal of α-pinene took place in bioreactors 1 and 3, as shown in Figure 2. However, the removal efficiency of bioreactor 2 only reached 45% in the same period. On day 0, before inoculation, no degradation of α-pinene was observed, indicating the absence of any abiotic removal. During the next 7 days, α-pinene degradation was observed only in the section of the reactor closest to the inlet in both reactors 1 and 3. Subsequently, α-pinene degradation moved progressively toward the outlet section of the biofilter bed. The data indicate that the inoculated Ophiostoma stenoceras first grew near the inlet section of the biofilter column and then spread out to the sections of the bed closer to the outlet over a period of 4 weeks. The regular addition of the nutrient solution probably helped fungal spreading. However, the removal of α-pinene in Bioreactor 2 only increased from 0 to 38% in 4 weeks, which is much lower than in the other reactors. It was caused by the limited biomass growth on the surface of the packing. Probably the high void space and plastic surface of Pall rings are unfavorable for a fast and significant biomass growth and adhesion. As can be seen in Figure 2, the performance of Bioreactor 3 improved after mixing perlite with the Pall rings. 3.2 ELIMINATION CAPACITY OF Á-PINENE Of the three support materials, lava rock performed best, achieving a α-pinene removal rate in excess of 143 g.m-3.h-1, followed closely by the mixture of Pall rings and perlite. Pall rings alone gave less favorable results. The relationships between the inlet load and the elimination capacity are shown in Figure 3 for each packing material. The values for the critical load and maximum elimination capacity are summarized in Table 2.
PERFORMANCE EVALUATION OF FUNGAL BIOFILTERS PACKED WITH PALL RING
Figure 2. Removal efficiency of α-pinene during the start-up period.
Figure 3. Relationship between load of α-pinene and elimination capacity. (a) Pall rings. (b) Lava rock, and Pall rings+perlite.
391
392
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
Table 2. Performance characteristics of the filter bed materials used in the experiment. Packing Pall rings Lava rock Perlite+Pall rings
Critical load (g.m-3.h-1) 18 125 130
Maximum elimination capacity (g.m-3.h-1) 17.5 143 135
The poor performance of Bioreactor 2 is due to the low biomass growth on the surface of Pall rings. Although there is no big difference between each other, for Bioreactor 1 and 3, the performance could be improved since most of the pollutant removal took place in only part of the packing. As much as 90% α-pinene removal took place in the inlet section of the reactor, corresponding to the upper one-third filter bed layer. The rest of the packing material of the reactor only played a minor role in the removal. This was mainly due to the non homogeneous distribution of the biomass in the filter bed, as easily confirmed by visual observation and observations under the microscope. Recent studies have shown that biomass distribution and performance could be improved with a directionally switching operation, in which the contaminant inlet feed is periodically reversed between the top and bottom of the bioreactor column or using a split-feed operation mode (Song and Kinney, 2001; Mendoza et al., 2003). 3.3 PRESSURE DROP Pressure drop is a key aspect of biofilter performance. It affects the energy consumption of the blower, which contributes most to the operation costs. In order to reduce such problem the design of high filter beds fed with large air flow rates should be avoided. The pressure drop depends on the nature of the filter bed and its moisture content. Adequate selection and design of the carrier is a means of reducing pressure drop. Compared to bacterial systems, the filamentous fungi may cause some higher head losses due to the fact that fungal biomass fills the pore spaces of the packing media. This may eventually lead to channeling and clogging problems in the biofilter, which ends up in a reduced efficiency. In the fungal biofilters treating α-pinene, no significant pressure drop was detected for any of the three packings even after 12 months operation (Figure 4). The pressure drops remained stable at low values (<2.4 cm H2O.m-1) over all the experimental period. Initially, it increased very slowly in Bioreactors 1 and 3, while the pressure drop of Bioreactor 2 remained at the same value (4 mm H2O.m-1) during that operational period. After 6 months, the pressure drop remained essentially constant in all three reactors. One can speculate that at the end of the experiment, the surface of Pall rings in Bioreactor 2 was covered with the
PERFORMANCE EVALUATION OF FUNGAL BIOFILTERS PACKED WITH PALL RING
393
biomass, but the whole matrix was still empty. The main problem in Biofilter 2 is insufficient biomass. Bioreactor 1 experienced a backwashing for removal of excess biomass on day 424. It was repacked and the pressure drop decreased to 1.4 cm H2O.m-1. Apparently, the porous structure of perlite allowed easy biomass growth and attachment. However, our experience with the perlite packing shows that it may need to be repacked after several months of operation, because heavy overgrowth of biomass causes compaction or channeling phenomena that can adversely affect the performance (Prado et al., 2002). It is well known that packing materials with high void fractions can limit pressure drop. In this study, the addition of Pall rings assured minimization of the pressure drop and avoided channeling, resulting in the superior performance of the mixture. This demonstrated that the mixture of Pall rings and perlite is very suitable for use in this fungal biofilter.
Figure 4. Variations of the pressure drop along the three biofilters.
3.4 SEM PHOTOS Figure 5 shows all three packing surfaces grown with filamentous fungi. The filamentous fungal structure enhances the mass transfer of hydrophobic pollutants from the gas phase to the biocatalyst, thereby improving the performance of biofilters.
394
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
Figure 5. SEM pictures of a filter-bed samples from three bioreactors colonized by a culture of Ophiostoma stenoceras. Left: lava rock, middle: Pall rings, right: perlite.
4 CONCLUSIONS The following conclusions can be drawn from the results presented in this study: (1) The biofilters packed with lava rock and the mixture of Pall rings and perlite reached a similar maximum elimination capacity around 143 g.m-3.h-1. This value is much higher than for the Pall rings-packed biofilter. (2) Pressure drop of all three bioreactors is below 2.4 cm H2O.m-1 over a one and half year operation. The addition of Pall rings to the perlite packing improved not only the performance of the biofilter, but did also stabilized the pressure drop. (3) Overall, the mixture of Pall rings and perlite is the best choice for this fungal biofilter treating α-pinene.
5 ACKNOWLEDGEMENTS The present research was funded by the Spanish Ministry of Education and Science (Project CTM2004-00427/TECNO) and through European FEDER funds. Yaomin Jin was financially supported by the Xunta de Galicia (Project PGIDIT05PCIC10304PN).
PERFORMANCE EVALUATION OF FUNGAL BIOFILTERS PACKED WITH PALL RING
395
REFERENCES Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. Boca Raton: Lewis Publishers. 299 p. Estévez, E., Veiga, M.C. and Kennes, C. (2005) Biodegradation of toluene by the new fungal isolates Paecilomyces variotii and Exophiala oligosperma. J. Ind. Microbiol. Biotechnol. 32(1): 33-37. Jin, Y., Veiga, M.C. and Kennes, C. (2006) Performance optimization of the fungal biodegradation of alpha-pinene in gas-phase biofilter. Proc. Biochem. 41(8): 1722-1728. Kennes, C. and Thalasso, F. (1998) Waste gas biotreatment technology. J. Chem. Technol. Biotechnol. 72(4): 303-319. Kennes, C. and Veiga, M.C. (2001) Bioreactors for Waste Gas Treatment. Kluwer Academic Publishers, Dordrecht, The Netherlands, 312 p. Kennes, C. and Veiga, M.C. (2002) Inert filter media for the biofiltration of waste gasescharacteristics and biomass control. Rev. Environ. Sci. Biotechnol. 1(3): 201-214. Kennes, C. and Veiga, M.C. (2004) Fungal biocatalysts in the biofiltration of VOC-polluted air. J. Biotechnol. 113(1-3): 305-319. Mendoza, J.A., Veiga, M.C. and Kennes, C. (2003) Biofiltration of waste gases in a reactor with a split-feed. J. Chem. Technol. Biotechnol. 78: 703-708. Ottengraf, S.P.P. and van den Oever, A.H.C. (1983) Kinetics of organic-compound removal from waste gases with a biological filter. Biotechnol. Bioeng. 25(12): 3089-3102. Prado, O.J., Mendoza J., Veiga, M.C. and Kennes, C. (2002) Optimization of nutrient supply in a downflow gas-phase biofilter packed with an inert carrier. Appl. Microbiol. Biotechnol. 59(4-5): 567-573. Song, J. and Kinney, K.A. (2001) Effect of directional switching frequency on toluene degradation in a vapor-phase bioreactor. Appl. Microbiol. Biotechnol. 56: 108-113.
This page intentionally left blank
397
Styrene degradation in perlite biofilter: The overall performance characteristics and dynamic response HALECKY MARTIN1, PACA JAN1, GERRARD A. MARK2 AND SOCCOL C. RICCARDO3 1
Institute of Chemical Technology, Department of Fermentation Chemistry and Bioengineering, Technicka 5, 160 28 Prague, Czech Republic 2 University of Teesside, Middlesbrough, TS1 3BA, England 3 Universidade Federal do Paraná, Centro Politécnico/Jardim das Américas, 19011, 81531-970 Curitiba-PR, Brasil
ABSTRACT Styrene’s degradation in a perlite biofilter including the long-term operation, dynamic response to step-changes in inlet concentration and non-use periods were tested. The study was performed in a bench-scale biofilter with ID 100 mm and a bed height of 1 m. Perlite with a particle size of 2 – 4 mm was used as a packing material. An enrichment mixed culture was immobilized on the packing. The inoculum was obtained from a styrene biofilter. Two different loading conditions were tested: (1) Loading with a high inlet concentration and a high residence time. (2) Loading with a low inlet concentration and the low residence time. Both conditions are common in industrial practice. The dynamic response to a repeated step-change in the inlet concentration (from 50 to 200 mg.m-3) was tested. The dynamic behaviour of the restarting period after varying periods of non-use was also investigated. The results demonstrate a high biofilter stability under extreme loading conditions and also during the step-changes of the inlet concentration. The non-use periods tested had almost no effect on the biofilter performance. The maximum outlet concentration after the restarting of the load was 4 mg.m-3, when a 95 hours idle period was used. After shorter idle periods, the outlet styrene concentrations did not exceed 0.6 mg.m-3.
397
398
M. HALECKY, J. PACA, A. M. GERRARD AND C. R. SOCCOL
1 INTRODUCTION Styrene is a toxic volatile liquid with a specific odor witch is commonly used in industry. Styrene monomer is used for the production of polystyrene, styrene copolymers and polystyrene resins (Paca and Koutsky 2000; Bina et al., 2004). Production and processing of styrene are main sources of styrene pollution. Another source of contamination is the incineration of styrene polymers. Since styrene is a toxic compound, it is necessary to reduce its emissions. In addition, by reason of very low odor threshold (0.1 ppm) and characteristic uncomfortable odor, styrene can be defined like odor emission. Styrene-containing gases may be biologically treated in biofilters (Paca and Koutsky, 2000; Bina et al., 2004; Cox et al., 1997) of trickling filters (Webster et al., 1999; Sorial et al., 1998). The object of the study is to test long-term biofilter stability under conditions which are common in industrial usage, such as sudden changes of inlet concentration, periods of non-use and different loading conditions.
2 MATERIALS AND METHODS 2.1 MICROORGANISMS AND CULTIVATION A biofilter was inoculated with enrichment mixed culture which was isolated from biofilter degrading styrene. The cells were cultivated in a mineral medium of the following composition (g.L-1): K2HPO4 4.3; KH2PO4 3.4; KNO2 2; MgCl2 0.34; MgCl2 0.7.10-3; FeSO4 0.6.10-3; Na2MoO4 1.7.10-3. The inoculum was achieved after 7 days cultivation in modified Erlenmayer flasks on a rotary shaker at 26 °C with gasoline vapor as the sole carbon and energy source. pH of the medium was 7.0. 2.2 EXPERIMENTAL SETUP A schematic diagram of the biofilter system is shown in Figure 1. The biofilter was made of glass with an internal diameter of 100 mm and a bed height of 1 m. The packing material was perlite with a grain size 1-3 mm and a porosity of 0.1027. Biofilter operating conditions were: up flow mode, temperature 21 – 23 °C, mineral medium added two times a week. 2.3 ANALYTICAL METHODS The styrene concentration in the gas phase was determined using an Agilent 6890 N gas chromatograph equipped with an ultra alloy-5 (5 % phenylmethylsilicone) capillary column 15 m in length, inner diameter of 0.53 mm, and film thickness of 1.5 μm (Quadrex Corp., UA5 - 30V - 1.5 F, New Haven, CT). The carrier gas was argon at a flow rate of 5 ml.min-1. The detection was carried out with a flame ionization
STYRENE DEGRADATION IN PERLITE BIOFILTER
399
Figure 1. 1 - blower, 2 – needle valve for flow rate control, 3 – valve for leachate, 4 – humidification vessel, 5 – styren suplying to air, 6 – rotameter, 7 – inlet sampling ports, 8 – packing, 9 – outlet sampling port
detector (FID) with hydrogen and air at flow rates of 30 ml.min-1 and 320 ml.min-1, respectively. Operating conditions were: inlet temperature, 200 °C; oven temperature, 150 °C; FID temperature, 250 °C. 2.4 CALCULATIONS Performance parameters of the biofilter - empty bed residence time (EBRT), removal efficiency (RE), elimination capacity (EC) and organic load (OL) - were calculated as follows:
400
M. HALECKY, J. PACA, A. M. GERRARD AND C. R. SOCCOL
where Cin, Cout are inlet and outlet concentrations (g.m-3), Q is air flow rate (m .h ) and Vb is bed volume (m3). 3
-1
3 RESULTS AND DISCUSSION 3.1 LONG-TERM BIOFILTER OPERATION The EBRT was 3.9 min during the start-up period. The increase of the inlet concentration during the first three days caused a drop of removal efficiency of styrene (Figure 2). When the inlet concentration was kept stable at 650 mg.m-3, the RE increased from 50 % to 98 % in 5 days. Neither the slow increasing of inlet concentration between days 10 and 22 nor the step change of the inlet concentration from 700 to 1300 mg.m-3 had no negative effect on RE which remained above 97 %. From this, it can be concluded that the start up-period took 15 days. The similar duration of start-up period was reported by Kraakman et al. (1997) (14 days), Juneson et al. (2001) and Arnold et al. (1997) (12 days) in biofilters. A much longer start up period was observed by Cox et al. (1993) for styrene removing biofilter inoculated with fungus Exophiala jeanselmei (4 – 5 weeks). Values of the inlet concentration from 50 to 1520 mg.m-3 and the EBRT from 0.4 to 3.9 min were used to test the biofilter stability under different loading manners (Figure 2a). The organic load ranged from 5.6 g.m-3.h-1 to 23.2 g.m-3.h-1. The removal efficiency did not drop below 95 % during the entire biofilter operation with the exception of the start-up phase (cf. Figure 2b, days 1-10) and the step change of the inlet concentration (day 104, Figure 4). It can be concluded that the degradation ability was very high and stable during a long-term operation, despite the very different loading conditions have been used. 3.2 RESPONSES OF NON-USE PERIODS Figure 3 shows the dynamic behavior of the re-starting period after varying periods of non-use. The data are not included in long-term performance plots. 8, 69 and 95 h of interruptions were tested. The inlet concentrations before these periods were 50 mg.m-3, the outlet concentrations were 0 mg.m-3 (RE = 100 %) and EBRT was 0.39 min. It is evident that the non-use periods had minimal effect on the biofilter
401
STYRENE DEGRADATION IN PERLITE BIOFILTER
Figure 2. Overall performance characteristics.
– Cin; Δ – EBRT;
– OL;
- RE
Figure 3. The dynamic behaviour of the re-starting periods after 8 hours ( ) day 127, 69 hours ( ) - day 132, 95 hours ( ) - day 140 of non-use periods.
402
M. HALECKY, J. PACA, A. M. GERRARD AND C. R. SOCCOL
performance. The maximum outlet concentration after the re-starting of the loading rate of 50 mg.m-3 was 4 mg.m-3, when a 95 h idle period was used. After shorter idle periods, the outlet styrene concentrations did not exceed 0.6 mg.m-3. The recovery of RE (to 100%) took several minutes only in cases of the nonuse periods lasting 8 and 69 h, while 2.8 h was necessary in a case of the longest nonuse period. Martin and Loehr (1996) for biofilter treating benzene reported two times longer response (4.2 h) to non use periods (40 and 90 h). When the non-use period lasted several days or several weeks, the biofilter recovery ranged from several days to even several weeks (Martin and Loehr, 1996; Bastos et al., 2003). 3.2 DYNAMIC RESPONSES TO STEP CHANGES OF INLET CONCENTRATION A dynamic response on a double step-change of the inlet concentration from 50 to 200 mg.m-3 was carried out on day 104 of operation. The step changes and the intermission lasted the same time (40 min). The EBRT value was 0.39 min.
Figure 4. Dynamic response on a double step-change of inlet concentration.
– RE;
– OL
The first step-change of the styrene inlet concentration resulted in a drop of the RE (from 98 % to 62 %). As it can be seen from Figure 4, a mild recovery of removal efficiency occurred during that step-change (from 62 % to 68 %). The second stepchange of the inlet concentration caused a smaller decrease of the RE than that of the first one (only to 77 %). However, the recovery of the biofilter was longer than in case of the first change. The original RE was achieved in 20 min before decreasing the inlet concentration to 50 mg.m-3 again.
STYRENE DEGRADATION IN PERLITE BIOFILTER
403
4 CONCLUSIONS • • •
The biofilter showed a stable degradation ability under very different loading patterns. The tested non-use periods had almost no effect on the biofilter performance. Step changes of the inlet concentration caused drop of the RE (to 62 %). Nevertheless, they had no effect on the long-term degradation rate of styrene.
5 ACKNOWLEDGEMENT The work was financially supported by the Czech Science Foundation, Join Project 104/05/0194 and by the Ministry of Education of the Czech Republic, Project MSM 6046137305.
REFERENCES Bastos, F.S.C., Castro P.M.L. and Jorge, R.F. (2003) Biological treatment of a contaminated gaseous emission from a paint and varnish plant – from laboratory studies to pilot-scale operation. J. Chem. Technol.Biotechnol. 78: 1201-1207. Bina, B., Dehghanzadeh, R., Pourmoghadas, H., Kalantary, A. and Torkian, A. (2004) Removal of styrene from waste gas stream using a biofilter. Journal of Research in Medical Sciences 6: 31-39 . Cox, H.H.J., Moerman, R.E., van Baalen, S., van Heiningen, W.N.M., Doddema, H.J., Harder, W. (1997) Performance of a styrene-degrading biofilter containing the yeast Exophiala jeanselmei. Biotechnol. Bioeng. 53: 259-266. Martin, F.J. and Loehr, R.C. (1996) Effect of periods of non-use on biofilter performance. J. Air Waste Manage. Assoc. 46: 539-546. Paca, J. and Koutsky, B. (2000) Effect of the packing materials on styrene removal in the biofilter, Proc. 2000 USC-TRG Conference on Biofiltration, Los Angeles, CA, USA, p. 21-29. Sorial, G.A., Smith, F.L., Suidan, A.P., Biswas, P. and Brenner, R.C. (1998) Evaluation of tricklebed air biofilter performance for styrene removal. Wat. Res. 32: 1593-1603. Webster, T.S., Cox, H.H.J. and Deshusses, M.A. (1999) Resolving operational and performance problems encountered in the use of a pilot/full scale biotrickling filter reactor. Environ. Prog. 18: 162-172.
This page intentionally left blank
405
Biodegradation of methyl ethyl ketone and methyl isopropyl ketone in a composite bead biofilter WU-CHUNG CHAN
AND
KANG-HONG PENG
Civil Engineering Department, Chung-Hua University, Hsinchu, Taiwan 30067, R. O.C.
ABSTRACT Biodegradation of methyl ethyl ketone (MEK) and methyl isopropyl ketone (MIPK) in a composite bead biofilter was investigated. Both microbial growth rate kg and biochemical reaction rate kd would be inhibited at higher inlet concentration. The kg and kd values of MEK were greater than those of MIPK in the average inlet concentration of 100-300 ppm. For the microbial growth process, the degree of inhibitive effect was almost the same sensitivity for two ketone compounds. Zeroorder kinetic with the diffusion rate limitation could be regarded as the most adequate biochemical reaction model. For the biochemical reaction process, the inhibitive effect was more pronounced for MEK in the average inlet concentration of 100-150 ppm and it was more pronounced for MIPK in the average inlet concentration of 150-300 ppm. The maximum elimination capacity of MEK and MIPK were 0.127 and 0.101 g-C h-1 kg-1 packed material.
1 INTRODUCTION The removal of volatile organic compounds (VOCs) from a polluted air stream using a biological process is highly efficient and has low installation and operation/ maintenance costs. Biofiltration technology offers environmental advantages: it does not generate undesirable byproducts by converting many organic and inorganic compounds into harmless oxidation products (e.g., water and carbon dioxide). Contaminants are then used as carbon and/or energy sources for the microorganisms within the biofilm. The solid filter material provides a nutrient source and matrix for the attachment of microorganisms in the biofiltration process. Therefore, the filter
405
406
WU-CHUNG CHAN
AND
KANG-HONG PENG
material property is an important factor in obtaining optimal pollutant removal. The optimal filter material should have the following characteristics: high moisture holding capacity, porosity, available nutrients, compression strength, and pH buffer capacity (Deviney et al., 1999). A spherical polyvinyl alcohol (PVA)/peat/KNO3/GAC composite bead was prepared and was proven suitable as a filter material in the biofiltration process in our previous works (Chan and Lin, 2006). The diffusivity of nutrient within the filter material was an important control factor for achieving good biofilter performance. Methyl ethyl ketone (MEK), methyl isobutyl ketone (MIBK) and methyl isopropyl ketone (MIPK) are widely used industrial chemicals. These ketone compounds were designed high-priority toxic chemicals. Large volumes of these ketone compounds are released into the atmosphere during manufacturing processes every year, leading to endanger the air quality and public health. Reports on the biodegradation of MEK and MIBK in a biofilter are limited. (Deshusses and Hamer, 1993; Deshusses et al., 1996; Deshusses, 1997; Lee et al., 2006) Recently, we had indicated that the process for degradation of VOCs in a composite bead biofilter could be divided into: lag, log growth and maximum stationary three phases; and the log growth and maximum stationary phases were important for controlling the removal efficiency of biofilter (Chan and Lin, 2006; Chan and Chang, 2006). However, details of the biodegradation kinetic in these two phases in biofilter are scant in the relevant literature. This article investigates the biochemical kinetic behaviors of MEK and MIPK in a composite bead biofilter. The composite bead is the spherical PVA/peat/ KNO3/GAC composite bead. The effect of inlet concentration and type of ketone compounds on the microbial growth rate and biochemical reaction rate are studied.
2 MATERIALS AND METHODS Peat (industrial grade from KekkilaOyj, Tuusula, Finland) was dried at 105oC before use. It has a dry density of 90 kg/m3, a pH of 5.5, a pore volume of 96%, and an organic substance content of 91%. Boric acid, sodium monobasic phosphate, sodium dibasic phosphate, potassium nitrate, methyl ethyl ketone and methyl isopropyl ketone (extra pure grade from Union Chemical, Hsinchu, Taiwan) were used as received. Poly (vinyl alcohol) (PVA) powder (industrial grade from Chung Chun Petrochemical, Hsinchu, Taiwan) and granular activated carbon (GAC) (industrial grade from Taipei Chemical, Hsinchu, Taiwan) were also used as received. The procedures for preparing PVA/peat/GAC/KNO3 composite beads and the apparatus and operation of the biofilter system were described in our previous work (Chan and Chang, 2006). Methyl ethyl ketone and methyl isopropyl ketone were used as VOCs.
BIODEGRADATION OF METHYL ETHYL KETONE AND METHYL ISOPROPYL KETONE
407
3 RESULTS AND DISCUSSION The variation in the percentage of removed VOCs over operation time in an air flow rate 0.102 m3 h-1, two ketone compounds average inlet concentration in the range of 100-300 ppm, an operation temperature 30oC and a relative humidity of more than 95 % through a composite bead filter bed is shown in Figure 1. (only the average inlet concentration of 200 ppm is shown because the data for the other concentrations were visually similar). It was found that the variation in the percentage of removed VOCs over operation time appeared in three phases: lag phase (phase I), log growth phase (phase II) and maximum stationary phase (phase III). (Chan and Lin, 2006) Only the biochemical kinetic behaviors in the log growth and maximum stationary phases was studied in this work.
Figure 1. Percentage of removed VOCs over operation time (t) at the average inlet concentration 200 ppm: ( ) methyl ethyl ketone, ( ) methyl isopropyl ketone.
3.1 MICROBIAL GROWTH In the log growth phase (phase II), the microbial growth rate increased exponentially and could be represented by the following equation (Chan and Lin, 2006) ln(C/C0) = -kg t
(1)
408
WU-CHUNG CHAN
AND
KANG-HONG PENG
where C0 and C are the concentration of VOCs in the inlet and exit air stream, respectively. A plot of ln(C/C0) versus t should correspond to a straight line and kg can be determined. The microbial growth rate kg of two ketone compounds at various inlet concentrations was calculated from the data in phase II and Eq. 1. The variations of the kg values with average inlet concentration C0 for two ketone compounds are shown in Figure 2. The kg value of MEK was greater than that of MIPK in the average inlet concentration range of 100-300 ppm. The reason was that the amount of MEK dissolved in the biofilm was greater than that of MIPK because the Henry’s law constant of MEK and MIPK was 6.6361x10-5 and 9.2561x10-5 atm-m3 mol-1, respectively (Yaws et al., 1998). Therefore, more microorganisms participated in the MEK biodegradation activity. The kg value decreased with increasing average inlet concentration in the concentration range of 100-300 ppm. An increase in the inlet concentration generally would enhance the transfer rate of the VOCs from the gas phase to the biofilm. This phenomenon leads more microorganisms participating in the biodegradation activity. However, high concentrations of some recalcitrant VOCs may produce inhibitive effects on the metabolic activity of the microbial population (Leson and Winer, 1991). Therefore, the result indicated that the microbial growth rate would be inhibited at higher inlet concentration.
Figure 2. The variations of kg with average inlet concentration (C0) for two ketone compounds: ( ) methyl ethyl ketone, ( ) methyl isopropyl ketone.
BIODEGRADATION OF METHYL ETHYL KETONE AND METHYL ISOPROPYL KETONE
409
The slope of the linear profiles in the concentration range of 100 to 150 ppm for MEK and MIPK were 2.83x10-3 and 2.27x10-3 h-1ppm-1, respectively; these in the concentration range of 150 to 200 ppm for MEK and MIPK were 7.56x10-4 and 7.28x10-4 h-1ppm-1, respectively; and these in concentration range of 200 to 300 ppm for MEK and MIPK were 0.89x10-4 and 1.10x10-4 h-1ppm-1, respectively. These results indicated that the inhibitive effect, resulting from increased inlet concentration, was almost the same sensitivity for two ketone compounds in the concentration range of 100-300 ppm. 3.2 BIOCHEMICAL REACTION In the maximum stationary phase, the population of viable cells was at a relatively constant value. Under the steady-state conditions, the substrate utilization rate by microbial was proposed by Ottengraf (Deviney et al., 1999). Three situations may be encountered in a biochemical reaction system. These corresponding equations could be derived from the Michaelis-Menten relationship to express the rates of biochemical reaction for each situation as follows: 1. First-order kinetic ln(C/ C0) = – k1θ
(2)
2. Zero-order kinetic with reaction rate limitation C0-C = k0θ
(3)
3. Zero-order kinetic with diffusion rate limitation 1-(C/C0)1/2 = kdθ
(4)
The plots of ln(C/C0) versusθ, C0-C versusθ and 1-(C/C0)1/2 versusθ calculated from the data in phase III. It was found that all three plots had high correlation coefficient and almost follow a linear relationship. To verify the biochemical reaction kinetic model, assume there was a plug air flow in the biofilter column and the following equation was derived from the MichaelisMenten equation (Valsaraj, 1995) (C0-C)/ln(C0/C) = Vm (θ/ln(C0/C))-Ks
(5)
where Ks is half-saturation constant and Vm is maximum reaction rate. A plot of (C0-C)/ln(C0/C) versus θ/ln(C0/C) should correspond to a straight line, and Ks and Vm can be determined. The plot of (C0-C)/ln(C0/C) versus θ/ln(C0/C) for two ketone
410
WU-CHUNG CHAN
AND
KANG-HONG PENG
compounds is shown Figure 3. The calculated Ks for MEK and MIPK were 21.56 and 22.96 ppm, respectively. The calculated Vm for MEK and MIPK were 9.06 and 7.55 g-C h-1 kg-1 packed material, respectively. The ratio of C0/Ks for MEK and MIPK were 4.64-13.92 and 4.36-13.07, respectively. Since the values of Ks and C0 were comparable for two ketone compounds in this study, zero-order kinetic with diffusion rate limitation could be regarded as the most adequate biochemical reaction kinetic model (Yang and Allen, 1994). The kd value of two ketone compounds at various inlet concentrations was calculated from Eq. 4. The variation of kd with average inlet concentration C0 for two ketone compounds is shown in Figure 4. The kd value of MEK was larger than that of MIPK in the average inlet concentration range of 100-300 ppm. The result indicated that the biodegraded rate of MEK was faster than that of MIPK in this concentration range. The kd value decreased with increasing average inlet concentration in the concentration range of 100-300 ppm. The result indicated that the biochemical reaction rate would also be inhibited at higher inlet concentration. The slope of the linear profiles in this concentration range of 100 to 150 ppm for MEK and MIPK were 8.18x10-4 and 2.34x10-4 h-1ppm-1, respectively. The result indicated that the inhibitive effect, resulting from increased inlet concentration, for MEK was more sensitive than that for MIPK in this concentration range. The slope of the linear profiles in the concentration range of 150 to 200 ppm for MEK and MIPK were 3.72x10-4 and 4.26x10-4 h-1ppm-1, respectively; and these in the concentration range of 200 to 300 ppm for MEK and MIPK were 8.2x10-5 and 9.5x10-5 h-1ppm-1, respectively. These results indicated that the inhibitive effect, resulting from increased inlet concentration, for MIPK was more sensitive than that for MEK in the concentration range of 150-300 ppm. 3.3 ELIMINATION CAPACITY The relationship of elimination capacity (EC) of biofilter versus load for two ketone compounds is shown in Figure 5. The maximum elimination capacity of MEK and MIPK were 0.127 and 0.101 g-C h-1 kg-1 packed material, respectively. The result indicated that the maximum elimination capacity of MEK was greater than that of MIPK. Thus, the compound with less number of carbons or no side group in the main chain would be easier biodegraded by the microbial. The result was closely corresponding to the result reported in the relevant literature (Deshusses and Hamer, 1993; Dehusses, 1997), which reported that the maximum elimination capacity of MEK was greater than that of MIBK.
BIODEGRADATION OF METHYL ETHYL KETONE AND METHYL ISOPROPYL KETONE
411
Figure 3. Plot of (C0-C)/ln(C0/C) versus θ/ln(C0/C) for two ketone compounds: ( ) methyl ethyl ketone, ( ) methyl isopropyl ketone.
Figure 4. The variations of kd with average inlet concentration (C0) for two ketone compounds: ( ) methyl ethyl ketone, ( ) methyl isopropyl ketone.
412
WU-CHUNG CHAN
AND
KANG-HONG PENG
Figure 5. The variations of elimination capacity (EC) with load for two ketone compounds biofilters: ( ) methyl ethyl ketone, ( ) methyl isopropyl ketone, (—) 100% removal.
4 CONCLUSIONS The biochemical kinetic behaviors of methyl ethyl ketone (MEK) and methyl isopropyl ketone (MIPK) in a composite bead biofilter were investigated. Both microbial growth rate kg and biochemical reaction rate kd would be inhibited at higher inlet concentration. For the microbial growth process, the inhibitive effect almost the same sensitivity for two ketone compounds and the kg value of MEK was larger than that of MIPK in the average inlet concentration range of 100-300 ppm. Zero-order kinetic with the diffusion rate limitation could be regarded as the most adequate biochemical reaction model. For the biochemical reaction process, the inhibitive effect for MEK was more sensitive than that for MIPK in the average inlet concentration of 100-150 ppm. The inhibitive effect for MIPK was more sensitive than that for MEK in the average inlet concentration range of 150-300 ppm. The kd value of MEK was greater than that of MIPK in the average inlet concentration range of 100-300 ppm. The maximum elimination capacity of MEK and MIPK were 0.127 and 0.101 g-C h-1 kg-1 packed material. MEK degraded by microbial was easier than MIPK did.
BIODEGRADATION OF METHYL ETHYL KETONE AND METHYL ISOPROPYL KETONE
413
5 ACKNOWLEDGEMENTS The authors wish to thank the National Science Council of the Republic of China for financial aid through the project, NSC 95-2211-E-216-019.
REFERENCES Chan, W.C. and Lin, Z.Y. (2006) A process to prepare a synthetic filter material containing nutrients for biofiltration. Bioresour. Technol. 26: 223-230. Chan, W.C. and Chang, L.Y. (2006) Kinetic behaviors between acetone and composite bead in biofilter. Appl. Microbiol. Biotechnol. 72: 190-196. Deshusses, M.A. and Hamer, G. (1993) The removal of volatile ketone mixtures from air in biofilters. Bioprocess Eng. 9: 141-146. Deshusses, M.A., Hamer, G. and Dunn, I.J. (1996) Transient-state behaviors of a biofilter removing mixtures of vapors of MEK and MIBK from air. Biotechnol. Bioeng. 49: 587-598. Deshusses, M.A. (1997) Transient-state behaviors of a biofilter: start-up, carbon balance, and interactions between pollutions. J. Environ Eng. June: 563-568. Deviney, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. New York: Lewis publishing Inc. Lee, T.H., Kim, J., Kim, M.J., and Ryu, H.W. (2006) Degradation characteristic of methyl ethyl ketone by Pseudomonas sp. KT-3 in liquid culture and biofilter. Chemosphere. 3: 315-322. Leson, G. and Winer, A.M. (1991) Biofiltration: an innovative air pollution control technology for VOC emissions. J. Air Waste Manage. Assoc. 41: 1045-1054. Valsaraj, K.T. (1995) Elements of environmental engineering: thermodynamics and kinetics. New York: Lewis publishing Inc. Yaws, C.L., Sheth, S.D. and Han, M. (1998) Using solubility and Henry’s law constant data for ketones in water. Pollut. Eng. 44: 44-46. Yang, Y. and Allen, E.R. (1994) Biofiltration control of hydrogen sulfide. 2. Kinetics, biofilter performance, and maintenance. J. Air Waste Manage. Assoc. 44: 1315-1321.
This page intentionally left blank
415
Removal of a mixture of oxygenated VOCs in a biotrickling filter F. J. ÁLVAREZ-HORNOS, C. GABALDÓN*, V. MARTÍNEZ-SORIA, P. MARZAL, J.M. PENYA-ROJA AND F. SEMPERE Department of Chemical Engineering, University of Valencia, Dr. Moliner, 50, 46100 Burjassot, Spain
ABSTRACT Laboratory scale-studies on the biodegradation of a 1:1:1 wt mixture of three oxygenated volatile organic compounds (VOCs), ethanol, ethyl acetate and methyl-ethyl ketone (MEK) in a biotrickling filter were carried out using two identically sized columns, filled with different polypropylene rings. The reactors were seeded with a two-month preconditioned culture from activated sludge. The performance of the biotrickling filters was examined for a continuous period of 4 months at VOC concentration from 125 mg-C/m3 to 550 mg-C/m3 and at gas flow rates of around 1.0 m3/h, 2.0 m3/h and 4.6 m3/h, which correspond to gas empty bed residence times (EBRT) of 68 s, 33 s and 16 s, respectively. Similar performance was obtained for both supports. Intermittent flow rate of trickling liquid was shown as beneficial to improve the removal efficiency of the system. A stratification in the substrate consumption was observed from gas composition profiles, with MEK % in the emission greater than 78%. Continuous VOC feeding resulted in an excessive accumulation of biomass and high pressure drop was developed in less than 20-30 days of operation. Intermittent VOC loading with night and weekend feed cut-off periods passing dried air, but without water addition, was shown as a successful operational mode to control the biofilm thickness. In this case, operation at high inlet loads was extended for more than 50 days maintaining high removal efficiencies and low pressure drops.
1 INTRODUCTION Biotrickling filters for air pollution control use a well-specified inorganic packing material and involves a liquid phase, which trickles through the bed. The liquid phase provides nutrients to the biofilm and allows for pH control, yielding a more stable
415
416
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL, J. M. PENYA-ROJA and F. SEMPERE
operation in comparison with biofilters. These characteristics along with a larger air/ liquid interfacial area lead to removal rates which are substantially higher than those obtained with conventional biofilters (Koutinas et al., 2005), implying potentially lower sizes and lower capital expenditure for industrial applications. These factors have caused a shift in interest from conventional biofilters to biotrickling filters in the recent years. A major disadvantage limiting its use for VOC removal is the clogging that occurs as a result of excessive microbial accumulation. Excessive biomass formation leads to progressive bed obstruction and is accompanied by an increase in pressure drop and channelling (Iliuta et al., 2005). To overcome this problem, the inert packing material can be cleaned by regular backwashing or by using chemicals. The number of studies concerning VOC removal by biotrickling filters has increased in the past decade (Iranpour et al., 2005). However, few studies have been focused on treating mixtures of pollutants (Kim et al., 2005; Paca et al., 2006) and on intermittent operation (Cai et al., 2005; Cox and Deshusses, 2002). By other side, although the effect of biomass accumulation on pressure drop has been reported (Cox and Deshusses, 1999; Smith et al., 1998; Weber and Hartmans, 1996), the optimization of operational parameters to avoid this problem still needs further research. The purpose of the present research was to investigate the removal of a mixture of three easy biodegradable oxygenated compounds: ethanol, ethyl acetate and methylethyl ketone (MEK) by two lab-biotrickling filters filled with different inert supports, taking into consideration the following objectives: (1) to study the interactions among the degradation rates of the three compounds, (2) to evaluate the clogging problems under different operational conditions, (3) and to show the benefits of short-term starvation periods in clogging control at high VOC loadings.
2 MATERIALS AND METHODS 2.1 BIOTRICKLING FILTER SYSTEM The experiments were performed using two identical laboratory-scale biotrickling filters, BTF1 and BTF2, constructed each of them of three cylindrical modules, with a total length of 120 cm and internal diameter of 14.4 cm. The packed bed contained four equidistant gas sampling ports and three equidistant bed sampling ports. Both reactors were filled with different polypropylene rings. The rings of reactor BTF1 had a nominal diameter of 40 mm and 97% of initial porosity. The reactor BTF2 was filled with Pall rings: first module was packed with 25 mm nominal diameter rings (initial porosity of 92%) and the other two modules were filled with 15 mm rings (initial porosity of 87%). The reactors were also provided with 20-cm of top and bottom spaces. The 1:1:1 wt mixture of ethanol, ethyl acetate, and MEK was introduced into the compressed, filtered and dried air by using a syringe pump (New Era, infusion/
REMOVAL OF A MIXTURE OF OXYGENATED VOCs IN A BIOTRICKLING FILTER
417
withdraw NE 1000 model, USA). Contaminated air was introduced through the bottom of the column; flowrate was adjusted by using mass flow controllers (Bronkhorst HiTec, Nederlands). The setup was completed with a 14-L recirculation tank. The recirculation solution, renewed every week, was fed into the reactor in a countercurrent flow with respect to air flow by using a centrifugal pump at 2 - 4 L/min. The nutrient solution (composition in g L-1: KH2PO4 5.8; K2HPO4 10.1; KNO3 50.3; MgSO4·7H2O 1.9, Ca, Fe, Zn, Co, Mn, Mo, Ni and B at trace doses, pH = 7.0) was supplied once per day (on working days) into the recirculation tank. The volume of nutrient solution was adjusted depending on the inlet load applied to assure N-NO3 concentrations greater than 10 mg L-1. Total concentration of VOC was measured by using a total hydrocarbon analyzer (Nira Mercury 901 model, Spirax-Sarco, Spain). The inlet and outlet gas streams were monitored daily and the intermediate ports were monitored once a week. The composition of the gas streams were periodically determined by using a gas chromatograph (GC 8000 model, CE Instruments, Spain) equipped with a 0.86 mL automated gas valve injection system and a flame ionization detector. Pressure drop was monitored daily when water was trickled through the beds. Soluble COD (Chemical oxygen demand), suspended solids and N-NO3 concentrations along with conductivity and pH of the recirculation solution were also monitored daily. The void fraction at the first one-third of the beds was determined once a week. BTF performance was evaluated in terms of inlet load (IL), elimination capacity (EC) and removal efficiency (RE). 2.2 OPERATIONAL CONDITIONS The two parallel reactors were operated in three consecutive phases: – Phase I (day 1 to 44): In this phase, continuous VOC feed was applied. After inoculation with a two-month preconditioned culture from activated sludge, the start-up was performed working at 33 g-C m-3h-1 of IL and 1.0 m3h-1 of gas flow rate until high removal efficiency was reached. Five days later, the gas flow rate was increased to 2.0 m3h-1 and two increasing inlet loads were applied: 33 and 63 g-C m-3h-1. Washing was applied as clogging control technique as follows: two-thirds of the BTF was filled with tap water, and then air was flown in pulses for 5 min. – Phase II (day 45 to 79): After disassembling the reactors on day 44 to remove the excessive accumulation of biomass, a continuous IL of 36 g-C m-3h-1 was applied at a gas flow rate of 4.7 m3h-1 to evaluate the total time until clogging occurred without washing. This strategy finally resulted in high pressure drops, so from day 79 to 89, two sequential starvation strategies were tested as possible procedures for clogging control: first, dried air without
418
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL, J. M. PENYA-ROJA and F. SEMPERE
VOC was supplied combined with water trickling for 15 minutes each 8 hours. Second, uncontaminated dried air was fed without water trickling. – Phase III (days 89 to 140): Intermittent loading simulating shift work of many industrial facilities (16 h/day, 5 days/week) was applied. Uncontaminated dried air was supplied during night and weekend periods as a clogging control procedure. Water trickling was off during night closures. During weekends, water trickling was varied between 15 min/8 h to off depending on the biofilm thickness. Three increasing inlet loads were applied: 37, 63 and 141 g-C m-3h-1 at a constant gas flow rate of 4.7 m3h-1.
3 RESULTS AND DISCUSSION 3.1 PERFORMANCE OF BIOTRICKLING FILTERS The results of the monitoring of the biotrickling filters are presented in Figures 1 and 2 for BTF1 and BTF2, respectively. In part a), the total RE and inlet and outlet concentrations are plotted. In part b), the pressure drop and the void fraction at the first one-third of the reactors are presented. Performance parameters are resumed in Table 1 for the different phases. Figures 1 and 2 show a similar performance of both biotrickling filters during the three phases. The start-up was successfully carried out, high RE was reached in 1-2 days for both reactors working at moderate conditions (IL of 33 g-C m-3h-1 and 68 s of EBRT). The decrease on RE observed on day 5 (Monday) was related to nitrogen exhaustion on the first weekend of operation. In phase I, operational conditions as nutrient dose or water trickling regime were regulated to improve the RE of the reactors, so this phase is characterized by some variability on RE. During the first two weeks of Phase I, high RE and low pressure drop was maintained in both BTFs. When inlet load was increased to 63 g-C m-3h-1, the continuous water trickling caused noticeable water retention, with partial flooding of the beds. Then, trickling regime was changed on day 22 from continuous to an intermittent mode (15 min/4 h). Besides, the intermittent water trickling was shown beneficial to improve the RE of the process in comparison with continuous trickling. The pressure drop progressively increased due to the accumulation of biomass until 350-450 Pa m-1 were reached on day 22. By applying a specific washing cleaning to partially remove the excess biomass, pressure drop was restored in previous low values, but the systems were not able to keep on low pressure drop more over than 5-10 days. At the end of the phase I, pressure drop was in 443 Pa m-1 for BTF1 and 1384 Pa m-1 for BTF2. As pressure drop increased, void fraction decreased, reaching values lower than 60% at the first one-third of the column for pressure drop higher than 300 Pa m-1.
REMOVAL OF A MIXTURE OF OXYGENATED VOCs IN A BIOTRICKLING FILTER
419
Figure 1. Monitoring the performance of biotrickling filter BTF1. a) VOC removal with time b) Evolution of the pressure drop and void fraction at the first one-third of the packed bed.
Figure 2. Monitoring the performance of biotrickling filter BTF2. a) VOC removal with time b) Evolution of the pressure drop and void fraction at the first one-third of the packed bed.
420
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL, J. M. PENYA-ROJA and F. SEMPERE
Table 1. Performance parameters for both biotrickling filters on the different phases. Days Phase I 5-15 15-44 Phase II 48-79 Phase III* 91-104 110-122 125-141 *
EBRT, s BTF1 BTF2
ILgC m-3h-1 BTF1 BTF2
ECgC m-3h-1 BTF1 BTF2
RE, % BTF1
BTF2
33 33
35 35
32.3±1.4 62.9±3.8
29.2±0.9 58.4±2.1
28.1±2.8 52.3±4.7
27.8±6.5 53.6±5.1
86.7±6.7 83.2±7.3
95.1±2.4 91.7±7.1
16
15
36.4±5.1
33.4±4.8
34.2±5.2
31.9±4.6
93.7±4.0
95.5±0.8
16 16 16
15 15 15
36.9±2.7 62.8±3.4 140.8±4.0
33.7±2.7 58.4±2.4 133.0±3.8
33.6±1.9 52.2±3.7 96.2±4.4
32.0±2.5 45.1±1.7 96.1±2.2
91.0±1.9 83.1±2.1 68.3±3.2
95.0±0.9 77.2±2.8 72.3±2.9
Data corresponding to monday (after weekend closure) were not considered.
In sight of the difficulties to adequately control the clogging at an IL of 63 gC m-3h-1, the reactors were disassembled on day 44 to remove the excess biomass prior to start the Phase II. In Phase II, IL was decreased to 36 g-C m-3h-1 to evaluate the total time needed to experience clogging problems working at 16 s of EBRT without washing. A 15 min/8h water trickling program was applied in this stage. From day 48, uniform and stable operation with high RE (> 94%) was observed in both biotrickling filters, even when high pressure drops were reached. On day 72, washing was conducted for both reactors, being the removal of accumulated biomass more effective on BTF2 than in BTF1. The systems were kept on high RE for another week more, but the presence of thick biofilms would have derived in clogging problems. So, a new strategy for clogging control was adopted: uncontaminated dried air was fed from day 79 to day 85 to promote the biological stabilization of the biomass; water trickling was maintained in 15 min/8 h. On day 85, although pressure drop remained in low values, no benefits in void fraction were attained. Thus, a more severe strategy was tested; water trickling was suppressed in order to facilitate the dehydration of the biofilm. On day 89, thin biofilms were observed and measures of void fraction indicated a restoration to values close to the initial ones. This result showed that dehydration periods could be beneficial to control the clogging problem derived from biofilm development. Considering this observation and taking into account that common industrial application usually involves non-use periods associated to shift work, an intermittent loading pattern was applied to the reactors in Phase III. During night and weekend non-fed periods, uncontaminated dried air was supplied to promote the partial
REMOVAL OF A MIXTURE OF OXYGENATED VOCs IN A BIOTRICKLING FILTER
421
dehydration of the biofilm. Three increasing inlet loads were applied at an EBRT of 16 s, and in all cases, stable operation was achieved, even for the higher IL. Water trickling was intermittently fed during 15 min/8 h for the IL of 37 g-C m-3h-1, and during 15 min/4 h for higher ILs. Average maximum ECs of 96.2 g-C m-3h-1 and 96.1 g-C m-3h-1 were reached at ILs of 141 g-C m-3h-1 and 133 g-C m-3h-1 for BTF1 and BTF2, respectively. The re-acclimation response after weekend closures indicated a nearly full recovery after feed resumption. Only data measured on monday mornings (days 96, 103, 110, 117, 124, 131, 138) showed a 5-10 % decrease on typical RE. Pressure drop was kept in low values for more than 50 days without further cleaning. These results demonstrate the feasibility of the process working at high loading and short EBRT conditions: the intermittent loading allows the application of short-term starvation combined with dehydration periods to control the biofilm thickness that prevents from clogging. 3.2 COMPOUND ABATEMENT This section deals with the evaluation of the efficiency of the process for the individual components. Results from Phase III, corresponding to operation under controlled pressure drop, are discussed. Similar general tendencies were observed for both biotrickling filters. Data for BTF1 are presented herein as an example. Figure 3 shows the removal efficiency for each pollutant, ethanol, ethyl acetate and MEK versus the total VOC concentration in the mixture feed for biotrickling filter BTF1 at an air flow rate of 4.7 m3h-1 (EBRT = 16 s). Ethanol and ethyl acetate were degraded more
Figure 3. Removal efficiency of each pollutant in the biotrickling filter BTF1 as a function of the total inlet concentration of oxygenated VOCs at an EBRT = 16 s. Data obtained under intermittent loading conditions corresponding to Phase III.
422
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL, J. M. PENYA-ROJA and F. SEMPERE
efficiently than MEK. These compounds were fully removed for VOC concentration in the mixture feed up to 300 mg-C m-3. For the highest inlet concentration tested, removal efficiencies slightly decreased to values of 91 and 95% for ethanol and ethyl acetate, respectively. Nevertheless, it was not possible to obtain so high MEK removal efficiencies in presence of ethyl acetate and ethanol, and MEK % in the emission remained always greater than 78%. For MEK, the removal efficiency varied as follows: from 175 to 644 mg-C m-3 of total VOC at the inlet, RE almost linearly decreased from 76% to 29%. Analyzing the total RE, it can be pointed out that biotrickling filter presented a competitive performance for treating the mixture of the three oxygenated compounds up to inlet concentrations as high as 300 mg-C m-3 by using an EBRT as short as 16 s. The adverse effect of ethyl acetate and ethanol on the removal of MEK is shown in Figure 4, in which pollutant concentration profiles along the length of the biotrickling filter BTF1 are shown for the three total ILs applied in Phase III. Results show that at ILs of 40 and 64 g-C m-3h-1, ethanol and ethyl acetate were completely eliminated in the first two-thirds of the column, with most of removal (between 50 to 80%) taking place over the first one-third. For the IL of 148 g-C m-3h-1, a greater breakthrough along the bed was observed, with removal mainly taking place in the first two-thirds of the column. MEK profiles indicated a competition between substrates so that more easily biodegradable ethanol and ethyl acetate were used prior. For ILs of 40 and 64 g-C m-3h-1, MEK was used at a nearly constant rate throughout the total height. At an IL of 148 g-C m-3h-1, negligible MEK removal was observed in the first one-third of the column; MEK started to be used in the last zones of the column, where ethanol and ethyl acetate have been mostly degraded.
Figure 4. Pollutant concentration profiles along the length of biotrickling filter BTF1. Data obtained under intermittent loading conditions corresponding to Phase III, EBRT = 16 s.
REMOVAL OF A MIXTURE OF OXYGENATED VOCs IN A BIOTRICKLING FILTER
423
4 CONCLUSIONS The removal of a mixture of three oxygenated VOCs, ethanol, ethyl acetate and MEK, in two biotrickling filters filled with different inert packing materials was investigated. No significant differences between both biotrickling filters were observed: similar removals and problems related to clogging were observed by using packing materials of different shape and initial void fraction. Under continuous loading, both in a continuous or intermittent trickling of water, biofilm growth caused an increase in pressure drop resulting in clogging problems in less than 10-30 days of operation for total ILs of 37 and 63 g-C m-3h-1 and EBRTs of 16 and 33 s. Weekly washing could improve the short-term performance, but it was unable to assure a medium-term operation. Intermittent loading operated 16 h/day and 5 days/week leads to short-term starvation periods that can be used to dehydrate the biofilm by combining inlet dried air with water trickling off. This strategy included intermittent water trickling and allowed achieving a stable operation at high loads and at short EBRT conditions with low pressure drop for more than 50 days; demonstrating its capability to adequately control the biofilm thickness and preventing clogging. Ethyl acetate and ethanol presented similar removal profiles, they were used mainly in the first one-third of the column; but a greater stratification of the MEK degradation was observed. Competition among the pollutants defers the MEK consumption to the last zones of the column bed, where the easiest biodegradable compounds have been degraded.
5 ACKNOWLEDGEMENTS Financial support by Pure Air Solutions b.v. (Netherlands), AIDIMA (Spain) and Ministerio de Educación y Ciencia (Spain, research project CTM 2004-05714-C0201/TECNO with FEDER funds) is acknowledged.
REFERENCES Cai, Z., Kim, D. and Sorial, G.A. (2005) Removal of methyl isobuthyl ketone from contaminated air by trickle-bed air biofilter. J. Environ. Eng. 131(9): 1322-1329. Cox, H.H.J. and Deshusses, M.A. (1999) Chemical removal of biomass from waste air biotrickling filters: screening of chemicals of potential interest. Wat. Res. 33(10): 2383-2391.
424
F. J. ÁLVAREZ, C. GABALDÓN, V. MARTÍNEZ, P. MARZAL, J. M. PENYA-ROJA and F. SEMPERE
Cox, H.H.J. and Deshusses, M.A. (2002) Effect of starvation on the performance and re-acclimation of biotrickling filters for air pollution control. Environ. Sci. Technol. 36(14): 3069-3073. Iliuta, I., Iliuta, M.C. and Larachi, F. (2005) Hydrodynamics modeling of bioclogging in waste gas treating trickle-bed bioreactors. Ind. Eng. Chem. Res. 44(14): 5044-5052. Iranpour, R., Cox, H.H.J., Deshusses, M.A. and Schroeder, E.D. (2005) Literature review of air pollution control biofilters and biotrickling filters for odor and volatile organic compound removal. Environ. Progress 24(3): 254-267. Kim, D., Cai, Z. and Sorial, G.A. (2005) Impact of interchanging VOCs on the performance of trickle bed air biofilter. Chem. Eng, J. 113(2-3): 153-160. Koutinas, M., Peeva, L.G. and Livingston, A.G. (2005) An attempt to compare the performance of bioscrubbers and biotrickling filters for degradation of ethyl acetate in gas streams. J. Chem. Technol. Biotechnol. 80(11): 1252-1260. Paca, J., Klapkova, E., Halecky, M., Jones, K. and Webster, T.S. (2006) Interactions of hydrophobic and hydrophilic solvent component degradation in an air-phase biotrickling filter reactor. Environ. Progress 25(4): 365-372. Smith, F.L., Sorial, G.A., Suidan, M.T., Pandit, A., Biswas, P. and Brenner, R.C. (1998) Evaluation of trickle bed air biofilter performance as a function of inlet VOC concentration and loading, and biomass control. J. Air Waste Manage. Assoc. 48(7): 627-636. Weber F.J. and Hartmans, S. (1996) Prevention of clogging in a biological trickle-bed reactor removing toluene from contaminated air. Biotechnol. Bioeng. 50(1): 91-97.
425
EMERGING BIOREACTOR TECHNOLOGIES
This page intentionally left blank
427
Solid-Liquid two-phase partitioning bioreactors for the treatment of gas-phase VOCs ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS AND NEAL G. BOUDREAU Department of Chemical Engineering, Queen’s University, Kingston Ontario Canada K7L3N6
ABSTRACT Two-Phase Partitioning Bioreactors (TPPBs) consist of a cell-containing aqueous phase and a separate, biocompatible and immiscible phase that partitions toxic substrates to the cells based on their metabolic demand and on maintaining the thermodynamic equilibrium of the system. TPPBs have traditionally used immiscible liquid organic solvents as the substrate delivery phase, however, one of the limitations of organic solvents is their potential bioavailability as substrates, and therefore these TPPB systems have generally been limited to the use of pure strains of organisms incapable of metabolizing the solvent. We have replaced the organic solvent phase in TPPBs with inert polymers (plastic beads). A TPPB employing styrene-butadiene beads as the sequestering phase was used to treat high step change loadings of BTEX in a contaminated air stream. The presence of the polymers allowed the system to effectively capture the incoming VOCs, buffer the cells from high VOC levels and release the VOCs to the cells for biodegradation. The polymer TPPB system demonstrated substantially higher performance than an aqueous phase bioscrubber and comparable performance to a solvent-aqueous TPPB. Also of great interest was the increase in oxygen transfer provided to the system by the addition of polymer beads, which have significant affinity for oxygen. The presence of polymer beads, which are biocompatible and non-bioavailable, provides a simple and effective means of enhancing the bioremediation of toxic organics present in gas streams, and potentially other phases.
1 INTRODUCTION Biological treatment of contaminated air streams can be an effective and economical means for the degradation of volatile contaminants in airstreams. Generally, the pollutants are absorbed from the gas phase to an aqueous phase in which the active
427
428
ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS
AND
NEAL G. BOUDREAU
microbial culture degrades the target pollutants. The most common types of biological treatment systems are biofilters, biotrickling filters, and bioscrubbers. The mechanisms for removal are similar for all reactor types, however, differences are found in the phases in which the microbial population is located, and the state of the liquid phase. Two-phase partitioning bioreactors (TPPBs) are a relatively new method of dealing with VOCs, with the inherent features of these devices allowing them to buffer what could possibly be toxic loadings of VOCs. TPPBs contain an aqueous, cell containing phase, as well as an immiscible organic phase that acts as a reservoir for toxic or inhibitory substrates. The organic phase absorbs the substrate as it enters the reactor and, based on equilibrium partitioning between the two phases, releases it to the cell containing aqueous phase at low concentrations. We have recently shown (Amsden et al., 2003; Prpich and Daugulis, 2004, 2005) that the second phase, traditionally an organic solvent, can be replaced by solid polymer (plastic) beads. One advantage of using polymer beads is that they are non-bioavailable to microorganisms, and thus, a consortium of bacteria, rather than a pure species, can be used for the degradation of pollutants. Solid polymers absorb (rather than adsorb) organic molecules, and are characterized by a partition coefficient analogous to organic solvents. In a somewhat related fashion to that employed in TPPBs, some researchers (Aizpuru et al., 2003; Tang et al., 2005; Weber and Hartmans, 1995) have used granular activated carbon (GAC) as part of their biofilter matrix to absorb VOCs, thus mitigating their potentially toxic effects on the microbial community present. Such systems, however, rely on adsorption rather than absorption (which is the VOC uptake mechanism in TPPBs) and are therefore limited by the GAC surface area. Full-scale industrial air treatment devices are exposed to changes in operating conditions and it is important to determine how effectively treatment processes will be able to handle these influent fluctuations. Biological treatment options are particularly sensitive to such variations as many pollutants in air streams can be toxic to microorganisms past a certain threshold concentration. The design of TPPBs provides the potential to handle fluctuations with the second immiscible phase acting as a buffer, or a «sponge», and absorbs the high concentrations of pollutants. This work was conducted in order to compare the performance of an organic solvent with polymer beads as second phases in a TPPB treating a continuous air stream contaminated with toluene, while at the same time testing the ability of TPPBs to handle transient VOC loadings. In addition, the positive effect that the presence of a polymer second phase can have on oxygen transfer was investigated.
SOLID-LIQUID TWO-PHASE PARTITIONING BIOREACTORS
429
2 MATERIALS AND METHODS 2.1 CHEMICALS In the two-liquid TPPB arrangement, n-hexadecane was used as the toluenesequestering phase. n-hexadecane is a suitable solvent for the TPPB treatment of VOCs and the long-term (> 1 month) use of hexadecane as a benzene delivery phase for A. xylosoxidans Y234 has been amply demonstrated (Nielsen et al., 2005a). Styrenebutadiene (28% styrene) ABA co-polymer beads (with dimensions of approximately L=4.25 mm, D=3.75 mm and density of 0.94 g/mL) were used as toluene-sequestering phase in the solid-liquid TPPB. For oxygen transfer experiments nylon 6,6, glass beads and silicone rubber beads were used in addition to the styrene-butadiene (SB) beads. 2.2 MICROORGANISM, MEDIUM, AND CULTURE CONDITIONS Achromobacter xylosoxidans Y234 is known to have the ability to degrade toluene, benzene and m-xylene. The growth medium (Davidson and Daugulis, 2003) was: 7 g/L (NH4)2SO4, 0.75 g/L MgSO4•7H2O, 6.6 g/L K2HPO4, 8.42 g/L KH2PO4, 2 g/L sodium benzoate, and 1 mL/L trace elements. Eight 125 mL Erlenmeyer shake flasks containing 50 mL of medium were inoculated from frozen stock prior to incubation at 30oC and 150 rpm for 24 hours in preparation for their inoculation in the bioreactor. 2.3 REACTOR SET-UP AND OPERATION A 5 L New Brunswick Scientific BioFlo III was set to operate at 30oC, a pH of 6.6, an agitation speed of 800 rpm and a total working volume of 3 liters. The aqueous medium consisted of 14 g/L (NH4)2SO4, 1.5 g/L MgSO4×7H20, 13.2 g/L K2HPO4, 16.84 g/L KH2PO4 and 0.16mL/L trace elements. For the fermentations conducted using liquid n-hexadecane as the second phase, an organic fraction of 0.33 was used with the remainder being the aqueous, cell-containing phase. The two phases were maintained as a dispersion through agitation. For the fermentations conducted using SB beads as a second phase, 500g of polymer beads were used with 2.518L of aqueous medium added for a final total volume of 3L. Higher bead fractions were found to result in excessive build up of beads behind baffles and other reactor internals, due to the relatively large size of the beads. The toluene delivery system consisted of an Erlenmeyer flask with 2L of toluene and a regulated amount of compressed air being sparged through it that continued into the reactor. A water bath kept the flask at the 30 oC. This air stream was mixed with air for bioreactor aeration and this combined stream was delivered into the reactor through a sparger. Dissolved oxygen levels were measured with a polarographic-membrane electrode. Concentrated nutrient boluses were added periodically (every 2-3 days) to
430
ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS
AND
NEAL G. BOUDREAU
the reactor to ensure that the system was not nutrient limited. A small amount of Sigma Antifoam 289 (~0.5mL) was added as required. 2.4 ANALYTICS Liquid samples, to measure biomass concentration, were centrifuged, the liquid supernatant was discarded and the biomass was then re-suspended in deionized water. Appropriate dilutions were then performed and measured at 650 nm and compared to a previously determined calibration curve. Inlet and outlet gas samples were taken by means of a gas tight 250 μL syringe. A Perkin Elmer AutoSystem Gas Chromatograph fitted with a flame-ionizing detector and a fused silica capillary column (DB-5, 0.53mm I.D., 30m length, 1mm film thickness, Model 125-503J, J & W Scientific, Inc., Folsom, CA) was used to analyze toluene concentrations. The aqueous phase toluene concentration was calculated based on Henry’s Law between air and the aqueous medium previously found to be 0.247 (mg/L)gas/(mg/L) aq. Toluene concentrations in n-hexadecane and SB beads were determined based on partition coefficients relative to the aqueous medium as determined previously. 2.5 STEADY STATE AND TRANSIENT OPERATION Immediately after inoculation a total flow rate of 1.71L/min air (0.57vvm) at a toluene concentration of 10 mg/L was established for a loading rate of 343 g/m3.h. This loading rate was maintained during the biomass growth phase and between dynamic step experiments. The cell growth reached a steady state in each case within 5 to 7 days at a cell mass in the bioreactor of between 20 and 25 grams (CDW). Achieving a steady state biomass concentration even with continued addition of substrate is due to the use of the consumed substrate for cell maintenance purposes only, rather than cell growth. All transient experiments were performed once the biomass levels had stabilized after the initial 5 to 7 day growth period. Inlet toluene steps were introduced for periods of 60 minutes by varying the proportions of air passing through the toluene flask and the aeration air, after which the toluene loading was reduced to its initial level. The size of the step was normalized with respect to total cell mass present to ensure that the performance of each system was not affected by the amount of biomass present. The step was imposed for each bioreactor configuration above the steady state feeding condition at a loading of approximately 110 (gToluene/m3reactor.h)/(g-cells). Alternatively, from a stable loading of 343 g/m3.h, steps of approximately 2400 g/m3.h were performed, after which the flows were readjusted to their original set-points and the inlet and outlet toluene levels were monitored until the instantaneous removal efficiency of the system returned to its original steady state value.
SOLID-LIQUID TWO-PHASE PARTITIONING BIOREACTORS
431
2.6 OXYGEN TRANSFER EXPERIMENTS The same New Brunswick Bioflo III operating at 30 °C was used for oxygen transfer experiments. All systems, which consisted of either tap water or 500 g of either glass beads, nylon 6,6, styrene–butadiene copolymer (SB) or silicone rubber in tap water to a total working volume of 3 L, were operated at aeration rates of 0.5 L/min, 0.75 L/min and 1 L/min and agitation rates of 100-800 rpm. Mass transfer measurements were taken in duplicate with the average value being reported, and the effect of probe response was also incorporated in the analysis. All solid particles used were approximately spherical in shape. The unsteady-state method described by Shuler and Kargi (2002) was used to determine mass transfer coefficients that reflected either the presence of inert particles (e.g. glass beads and nylon 6,6) alone, or the presence of particles that also possessed oxygen affinity (e.g styrene–butadiene or silicone rubber). The effect of the presence of inert particles was «separated» from the combined effect of «presence + affinity» using mathematical analysis as previously described (Littlejohns and Daugulis, 2007) to isolate the oxygen transfer enhancement that can be obtained in the presence of polymer beads possessing oxygen affinity.
3 RESULTS AND DISCUSSION 3.1 TPPB OPERATION Shortly (< two days) after inoculation, the removal efficiencies increased to greater than 95% and at a toluene loading of 343 g/m3.h, the cells reached steady state masses ranging between 19.4 and 26.2 grams (CDW) within 7 days. The removal efficiencies of the systems remained greater than 95% for the entirety of the experiments except during transients. Recent work by Nielsen et al. (2005a) has shown that a constant cell concentration will eventually be established due to cellular maintenance requirements, which are responsible for all of the substrate consumed. Figures 1-3 show the transient responses when the loading (~110 (g/m3.h)/ (g-cells)) was increased from its nominal rate of 343 g/m3.h to approximately 2400 g/ m3.h for a period of 60 minutes. The instantaneous removal efficiencies of the singlephase system and the polymer phase system both dropped immediately upon the onset of the step reaching minimum values of 57 and 87%, respectively before the end of the 60 minute step. This is also reflected in the outlet toluene concentrations which reached 20 and 10 mg/L, respectively. The instantaneous removal efficiencies of the n-hexadecane as a second phase system remained above 95% for the entirety of the 60 minute step, and in fact increased at the initial stages of the transient, reflecting absorption of the higher toluene loading. Outlet toluene concentrations remained low for this system, reaching just 2 mg/L, or about one-tenth of the single-phase system.
432
ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS
AND
NEAL G. BOUDREAU
Figure 1. Removal efficiency (triangles) and exit toluene concentration (squares) for single-phase system.
The solvent as a second phase system removed 97% of the toluene fed to the system over the course of the 200 minute experiment (Table 1), with the polymer phase system removing 90% of the toluene, and the single phase clearly performing the worst of the three systems removing only 69%. The performance comparison between the systems with second phases may not be entirely fair, however, given the different masses of second phases that were used. It can be anticipated that as more polymer phase is used (approaching the mass of n-hexadecane) the performance of this system would be closer to the two-liquid phase system. A comparison of the DO traces of the three systems (Figure 4) shows that DO for the solvent system remained the highest reaching a minimum DO value of 48% of saturation, while that of the polymer system was intermediate (33% of saturation), and the DO for the system with no second phase dropped to the lowest level (10% of saturation). The higher level of oxygen in the n-hexadecane case may be expected due to the greater capacity for oxygen by this solvent (Nielsen et al., 2003), and it is also interesting to see that the SB beads had a similar effect, albeit to a lesser degree with the mass of beads used in this case. Thus the presence of a second phase, originally intended to absorb and sequester toxic VOC substrates, has the added beneficial effect of enhanced oxygen absorption and release.
SOLID-LIQUID TWO-PHASE PARTITIONING BIOREACTORS
Figure 2. Removal efficiency (triangles) and exit toluene concentration (squares) for solvent-phase system.
Figure 3. Removal efficiency (triangles) and exit toluene concentration (squares) for polymer-phase system.
433
434
ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS
AND
NEAL G. BOUDREAU
Figure 4. DO profiles for single (circles), polymer (triangles), and solvent (squares) systems.
Table 1. Performance summary during imposed step transients. Reactor Type
95% Recovery Time (Minutes)
DO Recovery Time (Minutes)
Toluene Released During Step (mg)
Toluene Released After Step (mg)
Total Toluene Released (mg)
Overall Removal Efficiency (%)
Single -phase Solvent Polymer
30
60
3002
292
3294
69
63 48
115 162
116 864
198 165
314 1030
97 90
3.2 OXYGEN TRANSFER In light of the enhanced oxygen transfer seen in the polymer TPPB, the effect of the presence of polymer beads on O2 transfer was examined in more detail as described elsewhere (Littlejohns and Daugulis, 2007). The presence of the beads could have 2 effects on O2 uptake: a physical effect arising from their mere presence, and an absorptive effect in which O2 is actually taken up by the polymer. In order to examine this in detail, polymers with negligible O2 affinity (e.g. nylon 6,6 with a diffusivity of 1.6 x 10-9 cm2/s), and polymers with substantial O2 affinity (SB with a diffusivity of
SOLID-LIQUID TWO-PHASE PARTITIONING BIOREACTORS
435
Figure 5. Mass transfer coefficients at 400 rpm for nylon 6,6 (square), glass beads (circle), water (triangle and line), silicone rubber (square) and styrene–butadiene copolymer (star).
Figure 6. Oxygen transfer rate between 30% and 80% of liquid saturation by a system of water with silicone beads (circles) and water without particles (squares).
436
ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS
AND
NEAL G. BOUDREAU
1.4 x 10-6 cm2/s, and silicone rubber with a diffusivity of 3.4 x 10-5 cm2/s) were tested along with glass beads with essentially zero O2 affinity as a control. The presence of beads on both the kLa and on the oxygen transfer rate (OTR) was examined. The volumetric mass transfer coefficients (kLa) for aqueous systems with and without particles are shown in Figure 5 for different aeration rates at 400 rpm. The kLa values are up to 55% lower for the system with SB beads relative to the system without particles and up to 63% lower for the system containing silicone rubber beads, which at first seems counter-intuitive, given the earlier TPPB results. For systems containing SB and silicone rubber, the measurement of mass transfer coefficient contains the effect that the solid polymer particles may have on kLa, as well as any effects that they may have on absorbing oxygen. The system containing nylon 6,6 shows up to a 268% increase in kLa. Due to the low oxygen diffusion coefficient of nylon 6,6, the effect of the nylon particles on the kLa is isolated, and mass transfer enhancement due to the mere presence of particles alone is clearly observed. In a similar manner to nylon 6,6, glass beads are inert and enhance the kLa up to 159%. Both nylon 6,6 and silicone rubber have very similar dimensions and densities, which have been identified earlier as critical factors for the effect of particles on kLa. Nylon 6,6, can therefore be used to approximate the effect of the presence of silicone rubber beads on kLa, as both the effects of oxygen absorption by the silicone rubber and the effects on the gas–liquid mass transfer are contained within the measured kLa for the silicone rubber system and cannot be separated. In order to demonstrate a larger overall uptake of oxygen into a TPPB system relative to a system without a second phase, the instantaneous oxygen transfer rate (OTR) as a function of time at 400 rpm agitation and 1 L/min aeration is shown in Fig. 5. This plot, which was generated by mathematically «separating» the kLa effect from the overall observed measurements, clearly shows that between 30% and 80% of liquid saturation the system containing silicone rubber beads has a much larger OTR during the progression to liquid saturation than the system without a second phase. As well, the system with a second phase reaches 80% liquid saturation much later than the system without a second phase. This is due to the polymers acting as an oxygen sink within the system, in turn causing the liquid oxygen concentration to be lower relative to the system without polymers, at any given time. This decrease in the liquid concentration causes an increased driving force for oxygen between the gas and liquid phases, which causes a larger oxygen transfer rate for an extended period of time. Therefore, although the kLa is measured as lower for the reasons explained above, the overall oxygen transfer rate into the solid–liquid system is larger for systems containing particles with oxygen affinity (e.g. silicone rubber or SB). This is due to the oxygen transfer rate not only being proportional to the volumetric mass transfer coefficient, but also to the increased instantaneous concentration driving force. The SB and silicone polymers have a large uptake of oxygen and therefore more oxygen can ultimately be contained within the
SOLID-LIQUID TWO-PHASE PARTITIONING BIOREACTORS
437
system. These results are comparable to those for liquid–liquid systems that have found that oxygen is transferred at a higher rate due to an increased driving force (Nielsen et al., 2005b). However liquid–liquid systems increase the working volume oxygen saturation concentration, whereas the solid–liquid system increases the driving force by decreasing the liquid concentration at any given time, as well as by enhancing gas–liquid mass transfer. Nevertheless, both liquid–liquid and solid–liquid systems can increase the overall amount of oxygen that can be contained within a working volume.
4 CONCLUSION The presence of a second immiscible phase, whether a liquid, or a carefully selected polymer, has been shown to significantly improve the ability of bioreactors to enhance VOC removal. Although data from the use of a pure strain have confirmed this in the present work, similar results have also been obtained with microbial consortia operating in polymer-based TPPBs. In addition, the presence of these materials can have a significant positive effect on O2 transfer, which can be critical during dynamic periods of operation treating VOC surges. Adding such polymeric materials to more conventional biotreatment devices (e.g. biofilters) may also provide similar positive benefits on VOC buffering and removal, and on O2 transfer.
REFERENCES Aizpuru, A., Khammar, N., Malhautier, L. and Fanlo, J.L. (2003) Biofiltration for the treatment of complex mixtures of voc influence of the packing material. Acta Biotechnology. 23: 211-226. Amsden, B.G., Bochanysz, J. and Daugulis, A.J. (2003) Degradation of xenobiotics in a partitioning bioreactor in which the partitioning phase is a polymer. Biotechnol. Bioeng. 84: 399-405. Davidson, C.T. and Daugulis, A.J. (2003) Addressing biofilter limitations: A two-phase partitioning bioreactor process for the treatment of benzene and toluene contaminated gas streams. Biodegradation. 14: 415-421. Littlejohns, J.V. and Daugulis, A.J. (2007) Oxygen transfer in a gas-liquid system containing solids of varying oxygen affinity. Chem. Eng. J. 129: 67-74. Nielsen, D.R., Daugulis, A.J. and McLellan, P.J. (2003) A novel method of simulating oxygen mass transfer in two-phase partitioning bioreactors. Biotechnol. Bioeng. 83: 735-742.
438
ANDREW J. DAUGULIS, JENNIFER V. LITTLEJOHNS
AND
NEAL G. BOUDREAU
Nielsen, D.R., Daugulis, A.J. and McLellan, P.J. (2005a) Quantifying maintenance requirements from the steady-state operation of a two-phase partitioning bioscrubber. Biotechnol. Bioeng. 90: 248-258. Nielsen, D.R., Daugulis, A.J. and McLellan, P.J. (2005b) A Restructured framework for modeling oxygen transfer in two-phase partitioning bioreactors. Biotechnol. Bioeng. 91: 773-777. Prpich, G.P. and Daugulis, A.J. (2004) Polymer development for enhanced delivery of phenol in a solid-liquid two-phase partitioning bioreactor. Biotechnol. Progr. 20: 1725-1732. Prpich, G.P. and Daugulis, A.J. (2005) Enhanced biodegradation of phenol by a microbial consortium in a solid-liquid two phase partitioning bioreactor. Biodegradation. 16: 329-339. Shuler, M.L. and Kargi, F. (2002) Bioprocess Engineering, 2nd ed., Prentice Hall, New Jersey. Tang, H.M., Hwang, S-J and Hwang, S-C. (1995) Dynamics of toluene degradation in biofilters. Hazard Waste Hazard Mater. 12: 207-219. Weber, F.J. and Hartmans, S. (1995) Use of activated carbon as a buffer in biofiltration of waste gases with fluctuating concentrations of toluene. Appl. Microbiol. Biotechnol. 43: 365-369.
439
Microbial solvent regeneration in biotreatment of air contaminated by styrene ERIC DUMONT AND YVES ANDRÈS UMR CNRS 6144 GEPEA, Ecole des Mines de Nantes, 4 rue Alfred Kastler, BP 20722, 44307 Nantes Cedex 03, France
ABSTRACT In this study, the biodegradation of a styrene-polluted waste gas in a reactor containing 5 L of a biphasic mixture (10:90% v/v) of organic solvent (silicone oil) – water – biomass was investigated to establish the ability of a microbial solvent regeneration. Reproducible microbial solvent regenerations have been observed. The regeneration time, which increases with the increase of the styrene load (varying from 543 to 1800 mg), leads to elimination capacity up to 48 gstyrene.m-3mixture.h-1. The solvent regeneration requires roughly 1.2 molecule of oxygen per molecule of styrene and corresponds to the first steps of the biodegradation of the styrene.
1 INTRODUCTION Styrene is of major importance in the petrochemical and polymer-processing industries, which can contribute to the pollution of natural resources via the release of styrenecontaminated effluents and off-gases. Generally the produced polluted air flows are high in volume with low styrene concentrations (around 200 mg.m-3) corresponding to the application area of bioprocesses and some biofilter utilizations are described (Cox et al., 1996; Arnold et al., 1997; Jorio et al., 2000; Zilli et al., 2001; Dehghanzadeh et al., 2005) but with the needing of high bed volume. Another possibility in biological air treatment is the use of a bioscrubber, combination of a column for the pollutant air to liquid transfer and a biological reactor for the solvent recycling. However, for waste gases containing hydrophobic compounds having low solubility in water (like styrene), it is necessary to use a mixture of a non-biodegradable organic solvent and water. The pollutant is preferentially transferred from the gas phase to the organic solvent, and
439
440
ERIC DUMONT and YVES ANDRÈS
diffuses to the aqueous phase in order to be degraded by the microorganisms. Thus complete regeneration of the mixture of organic solvent – water is possible. There are only few studies devoted to the regeneration of the organic solvent for styrene biodegradation (see Dumont et al., 2006a). Consequently, the aim of this study is to present preliminary results about a solvent regeneration due to the degradation of the styrene by a mixed culture able to use this molecule as the unique source of carbon. Silicone oil is specifically used as solvent and the regeneration is followed by the measure of the oxygen variation in the gas phase in a batch reactor.
2. MATERIALS AND METHOD 2.1 MICROORGANISMS AND CHEMICAL The mixed culture was obtained from the Nantes (France) wastewater treatment plant. Styrene was used as the sole source of carbon and energy and a nutrient solution consisting in an aqueous solution of H8N2O4S and H2KO4P was used according to the quantity of carbon to keep the C/N/P ratio around 100/5/1. The biomass was progressively acclimated to styrene. Silicone oil (Rhodorsil fluid 47V5, dimethylpolysiloxane) was obtained from Rhodia Company. The physical properties at 25°C are: viscosity, 5 mPa.s; density, 910 kg.m-3; styrene solubility, 38g/L (Dumont et al., 2006b); oxygen solubility, 7 higher than in water (Dumont et al., 2006a). The styrene solubility in water at 25°C is 320 mg/L (Kirk-Othmer, 1983). 2.2. EXPERIMENTAL SETUP The description of the reactor used in thus study is shown in Figure 1. The reactor has an 11.5 L total volume (height 0.33 m, diameter 0.21 m). In the experiments, air was supplied from a compressor and sparged through an elliptical distributor (75x150mm) with 50 holes (1mm diameter). All experiments were carried out at a constant temperature of 25°C maintained by a thermostatic bath. The total volume of the mixture was 5 L (silicone oil volume fraction: 10%) allowing to absorb roughly 3g of styrene (Dumont et al., 2006b). Each experiment was carried out according to the following sequential procedure: 1) A synthetic waste gas stream was prepared by passing compressed air through a styrene generator filled with liquid styrene (flowmeters 3 and 4 opened; valves 1 and 5 opened; valve 2 closed). The known styrene stream bubbled through the mixture silicone oil – water – biomass during 1 hour in order to be absorbed by the liquids. The styrene concentrations in the gas phase were measured simultaneously at the inlet and outlet of the reactor using a
MICROBIAL SOLVENT REGENERATION IN BIOTREATMENT OF AIR CONTAMINATED
441
Figure 1. Scheme of the laboratory scale multiphase reactor.
Flame Ionisation Detector (Combustion HFR 400 FFID) calibrated before each experiments with standards. From these measures the amount of styrene absorbed by liquids is known. 2) The system was then completely isolated from the outside air by closing the valves 1 and 5. By closing the flowmeter 4 and opening the valve 2, a finite volume of air (6.7 L) was continuously flowed through the mixture in order to supply the biomass in oxygen (gas flow rate 0.9 10-4 m3/s). The decrease in oxygen concentration in air due to the biodegradation of the styrene by microorganisms was monitored and recorded as a function of time during 23h for further analysis. The oxygen fraction in the gas phase was determined using a paramagnetic oxygen analyser (Cosma Cristal 300). It was assumed that: (i) in the reactor the ideal gas law is applicable to calculate the number of moles of oxygen absorbed by the liquids (temperature and pressure of the gas phase was low), (ii) the presence of silicone oil in the emulsion does not change Henry constant for oxygen in water, (iii) the response time for the paramagnetic oxygen analyser is <5 s which is smaller than the mass transfer response time of system.
442
ERIC DUMONT and YVES ANDRÈS
3 RESULTS AND DISCUSSION Tests of sequential styrene degradation were conducted over a continuous period of 40 days. During these operations input styrene concentrations in the gas phase were adjusted in order to change the styrene load in the mixture which varied from 543 mg to 1800 mg. Several tests were carried out for the same styrene load. Typical examples of oxygen decrease in gas phase versus time are presented in Figure 2 (absorbed styrene around 1150 mg). According to the data shown in Figure 2, it appears that results can be reproduced. Firstly, the oxygen fraction dramatically decreases due to styrene degradation and secondly increases very slowly to reach an equilibrium plateau. This slight increase in probably due to the oxygen balance between the three phases air – silicone oil – water. At the end of each experiment, the oxygen concentration in the water phase was measured and compared with the theoretical value which should verify the oxygen balance between air and water. Experimental and theoretical results are in good agreement. For a styrene load of 1150 mg, the reproducible results presented in Figure 2 indicate that the whole mass of pollutant is degraded during the first 5 hours of the experiments. It can be assumed that this period represents the regeneration time of the solvent by the microorganisms. Figure 3 presents experimental data obtained for different styrene loads. As it can be observed, the time of degradation logically increases with the styrene load and the oxygen fraction decreases by about the same proportion as the styrene load increases. From these data, it can be possible to estimate the elimination capacity of the mixture and the mole number of atomic oxygen used to degrade one mole of styrene (table 1). The obtained elimination capacities are in the same order of magnitude of elimination capacities measured in biofilters: Cox et al. (1996), using perlite-packed filters to enrich styrene-degrading fungi, reported styrene elimination capacities of 70 g.m-3filter bed.h-1. Arnold et al. (1997), using peat as filter material, reported 30 g.m-3filter bed.h-1. Zilli et al. (2001) using a filter packed with a mixture of peat and glass beads (4:1) inoculated with the styrene degrader Rhocococcus rhodochrous AL NCIMB 13259, recorded an elimination capacity of 63 g.m-3filter bed.h-1 and Dehghanzadeh et al. (2005), using yard waste compost mixed with shredded hard plastics in a 75:25 v/v ratio of plastics inoculated with thickened municipal activated sludge, obtained 45 g.m-3filter bed.h-1. In contrast, Jorio et al. (2000) recorded maximum elimination capacities up to 141 g.m-3.h-1. In the specific case of biodegradation due to microorganisms in a biphasic aqueous-organic mixture, Osswald et al. (1996) measured in a stirred reactor similar elimination capacities (ranged from 23 to 63 g.m-3.h-1).
MICROBIAL SOLVENT REGENERATION IN BIOTREATMENT OF AIR CONTAMINATED
443
Figure 2. Typical examples of oxygen decrease in the gas phase versus time due to styrene elimination (styrene load: around 1150 mg in 5 L mixture).
Figure 3. Oxygen fraction decrease in the gas phase versus time for different styrene loads (total liquid mixture: 5 L; silicone oil volume fraction 10%).
444
ERIC DUMONT and YVES ANDRÈS
On average the biodegradation of one molecule of styrene implies 1.2 moles of atomic oxygen. According to O’Leary et al. (2002), two main pathways for the aerobic degradation of styrene exist: an initial oxidation of the vinyl side-chain and a direct attack on the aromatic nucleus (Figure 4). For theses two pathways, the conversion of styrene in either 3-vinylcatechol or phenylacetic acid uses one mole of atomic oxygen. Then it can be assumed that the first steps of the styrene conversion to soluble intermediates in water are sufficient to regenerate the organic solvent for further applications without needing the complete biodegradation in CO2, water and biomass. Obviously, it should be necessary to carry out consecutively a large number of sequential procedures in order to verify that the mixture is not polluted by degradation products, i.e. the solvent regeneration time remains constant for a given styrene load. Table 1. Elimination capacity of the biphasic mixture and ratio O2/styrene according to the styrene load. Styrene load in the mixture (g) 0.543 1.150 1.522 1.800
mole O2 mole styrene removed
Elimination capacity (gstyrene.m-3mixture.h-1)
1.7 1.1 1.0 1.2
20 42 48 27
4 CONCLUSIONS The following conclusions can be drawn from the preliminary experimental results presented in this study: (i) reproducible microbial solvent regeneration has been observed. The regeneration time, which increases with the increase of the styrene load, leads to elimination capacity ranged from 20 to 48 gstyrene.m-3mixture.h-1. (ii) the solvent regeneration requires roughly 1.2 molecule of oxygen per molecule of styrene which corresponds to the first steps of the biodegradation of the styrene. Obviously, further experiments will be necessary to complete this preliminary results including the continuous measurement of the pH of the emulsion, the dissolved oxygen in the aqueous phase, the styrene concentration in the gas phase, the identification of the microorganisms present in the mixture and if possible the identification of the intermediary products of the styrene conversion by the microorganisms.
MICROBIAL SOLVENT REGENERATION IN BIOTREATMENT OF AIR CONTAMINATED
445
Figure 4. Two major pathways of bacterial styrene degradation: oxidation of the vinyl side-chain and a direct attack on the aromatic nucleus.
446
ERIC DUMONT and YVES ANDRÈS
REFERENCES Arnold, M., Reittu, A.M., von Wright, A., Martikainen, P.J. and Suikho, M.L. (1997) Bacterial degradation of styrene in waste gases using a peat filter. Appl. Microbiol. Biotechnol. 48: 738-744. Cox, H.H.J., Faber, B.W., van Heiningen, W.N.M., Radhoe, H., Doddema H.J. and Harder, W. (1996) Styrene metabolism in Exophiala jeanselmei and involvement of a cytochrome P450-dependent styrene monooxygenase. Appl. Environ. Microbiol. 62(4): 1471-1474. Dehghanzadeh, R., Torkian, A., Bina, B., Poormoghaddas, H. and Kalantary, A. (2005) Biodegradation of styrene laden waste gas stream using a compost-based biofilter. Chemosphere. 60: 434-439. Dumont, E., Andrès, Y. and Le Cloirec, P. (2006a) Effect of organic solvents on oxygen mass transfer in multiphase systems: Application to bioreactors in environmental protection. Biochem. Eng. J. 30(3): 245-252. Dumont, E., Andrès, Y. and Le Cloirec, P. (2006b) Mass transfer coefficients of styrene and oxygen into silicone oil emulsions in a bubble reactor. Chem. Engin. Sci. 61(17): 5612-5619. Jorio, H., Bibeau, L. and Heitz, M. (2000) Biofiltration of air contaminated by styrene: effect of nitrogen supply, gas flow rate, and inlet concentration. Environ. Sci. Technol. 34: 1764-1771. Kirk-Othmer (1983) Encyclopedia of chemical technology. New York, Wiley & Sons, 3rd, Vol 21. O’Leary, N.D., O’Connor, K.E. and Dobson, A.D.W. (2002) Biochemistry, genetics and physiology of microbial styrene degradation. FEMS Microbiol. Rev. 26: 403-417. Osswald, P., Baveye, P. and Block, J.C. (1996) Bacterial influence on partitioning rate during the biodegradation of styrene in a biphasic aqueous-organic system. Biodegradation. 7: 297-302. Zilli, M., Palazzi, E., Sene, L., Converti, A. and Del Borghi, M. (2001) Toluene and styrene removal from air in biofilter. Proc. Biochem. 37: 423-429.
447
Long-term stability of Pseudomonas putida cultures during the off-gas treatment of toluene RAÚL MUÑOZ TORRE, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL SANTIAGO VILLAVERDE
AND
Department of Chemical Engineering and Environmental Technology University of Valladolid, Paseo del Prado de la Magdalena s/n, Valladolid, Spain
ABSTRACT The long term stability of Pseudomonas putida mt-2 and Pseudomonas putida F1, bacteria harbouring the genes responsible for toluene biodegradation in the plasmid pWW0 and in the chromosome, respectively, was investigated in a chemostat under high toluene loadings. When inoculated with P. putida mt-2 process collapse occurred after approx 3.5 days regardless the toluene gas inlet concentration (9.1 and 3.2 g m-3). Maximum bacterial activity (evaluated based on toluene elimination capacity (EC), CO2 production rate, and specific respiration rate was achieved at the first day of cultivation for both concentrations. Afterwards, both toluene EC and CO2 production decreased concomitantly with a decrease in the biomass concentration. Likewise, the specific respiration rate on toluene steadily decreased after 1 day of cultivation suggesting a toluene mediated accumulative microbial damage. In addition, the fraction of toluene degrading cells decreased during the second day of cultivation from 100 to 73 ± 4 %, and remained constant until the end of the experimentation. PCR and plasmid restriction analyses of non toluene degrading isolates indicated deletion of a significant portion of the pWW0 plasmid. P. putida F1 exhibited however a stable performance as shown by the stability of the EC, CO2 production rate, and biomass concentration within each steady state when exposed to gas toluene concentrations of 3.3, 6.2, 11.2 and 20.2 g m-3. In addition, no mutant population was established in the chemostat as shown by a constant fraction of toluene degrading cells of approx 100 %.
1 INTRODUCTION Despite being more cost-effective and environmental friendly than their physical chemical counterpartners, biological off-gas technologies for the treatment of Volatile Organic Compounds (VOCs) can be seriously limited by the toxic and mutagenic
447
448
RAÚL MUÑOZ, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL AND
SANTIAGO VILLAVERDE
nature of some of these VOCs and their excreted metabolic intermediates. This toxicity and mutagenic nature might impose irreversible or temporary looses of the degradation capacity of microorganisms during the long-term exposure at low concentration mutagenic contaminants, or as a result of short-term episodes of high inlet pollutant concentration (Jones et al., 1997; Mirpuri et al., 1997; Oliveira and Livingston, 2003). Thus, operational problems derived from microbial instability in processes treating toluene have been recently reported in literature. For instance, Song and Kinney (2005) reported a decline in the elimination capacity (EC) of biofilters subjected to high toluene loadings, likely due to the deterioration of the toluene degrading community. In their study, the process was maintained stable during 3 weeks at sub-critical conditions (RE = 100 %) and then overloaded. The EC immediately increased (concomitantly with a decrease in RE) reaching a pseudo-stationary state and started to decrease approx. 10-12 days after process overload. However, the mechanisms responsible of microbial deterioration were not identified. Likewise, Jones et al. (1997) and Villaverde et al. (1997) observed a decreasing fraction of toluene degrading Pseudomonas putida 54G in both suspended and biofilm-based bioreactors during the off-gas treatment of 0.5-3 g toluene m-3. This reduction, measured as cellular culturability on toluene, correlated with decreasing toluene degradation rates, being more pronounced the higher the toluene inlet concentration was. In addition, the fraction of respiring non-toluene culturable cells (supposed to grow at the expenses of leakage and lysis products) increased up to 95.5 %. Practical implications of this reduction in the fraction of toluene degrading cells might be the increase in culture oxygen requirements, which itself is problematic when treating high toluene loads (Villaverde et al., 1997). In this context, Leddy et al. (1995) and Brinkmann et al. (1994) reported that when present at high concentrations, toluene can induce irreversible defects in both plasmid and chromosomally encoded toluene degradation pathways in P. putida strains. Therefore, further research on the implications that toluene-mediated mutations in Pseudomonas culture impose on process performance, i.e EC, is needed in order to overcome operational problems derived from microbial instability in the long-term operation. In this context, Pseudomonas putida mt-2, harbouring the genes for toluene degradation in the TOL plasmid pWW0, and Pseudomonas putida F1, harbouring the chromosomally encoded TOD degradation pathway, were compared in terms of process stability under high toluene loadings. Culture stability was evaluated using process parameters such as EC, CO2 production, biomass concentration, fraction of toluene degrading cells, and specific toluene respiration rates.
LONG-TERM STABILITY OF PSEUDOMONAS PUTIDA
449
2 MATERIALS AND METHODS 2.1 MICROORGANISMS AND CULTIVATION Pseudomonas putida mt-2 [DSM 3931] and P. putida F1 [DSM 6899] were selected for its well known capability to mineralize toluene using the TOL and TOD pathways, respectively (Worsey and Williams, 1975; Bordel et al., 2007). The MSM used for P. putida cultivation was composed of (g·l-1): NaHPO4· 12H2O, 6.15; KH2PO4, 1.52; (NH4)2SO4, 1.0; MgSO4·7H2O, 0.2; CaCl2, 0.038; and 10 ml·l-1 of a trace element solution containing (g·l -1): EDTA, 0.5; FeSO 4·7H 2O, 0.2; ZnSO 4·7H 2O, 0.01; MnCl2·4H2O, 0.003; H3BO3, 0.03; CoCl2·6H2O, 0.02; CuCl2·2H2O, 0.001; NiCl2·6H2O, 0.002; NaMoO4·2H2O, 0.003. The final pH of medium was 7.0. Pseudomonas putida mt-2 was cultivated in MSM where trace elements were increased by 100 %. 2.2 EXPERIMENTAL SETUP The influence of toluene concentration on process stability was investigated under aseptic conditions in a magnetically stirred 1-L glass bioreactor (Afora S.A, Spain) operated as a chemostat. Temperature and agitation rate were maintained constant at 25 °C and 500 rpm, respectively. The reactor was filled with 900 ml of sterile MSM and inoculated with 40 ml of the tested P. putida strain to attain an initial biomass concentration of approx. 4-7 mg Dry Weight l-1. Toluene was supplied in the gas phase through the aeration (1100 ml min-1 of synthetic air filtered through a 0.2 m Millex® –FG membrane filter) by mixing a toluene-saturated stream with a toluenefree air stream at different proportions (Fig. 1). When inoculating with P. putida mt-2, two series of continuous experiments were carried out during 4 days. In the first series of experiments (D = 0.2 h-1) toluene inlet concentration was set at 9.1 ± 0.9 g m-3. In the second series (D = 0.1 h-1 in order to avoid bacterial washout) the suspended growth reactor was fed with 3.2 ± 0.3 g toluene m-3. When testing P. putida F1, two series of continuous experiments were carried out during 32 days at D of 0.1 and 0.3 h-1, respectively. In the first series of experiments (D = 0.1 h -1) toluene inlet concentration was set at 3.3 ± 0.2 g m-3 during 21 days and thereafter increased up to 6.2 ± 0.4 g m-3. In the second series (D = 0.3 h-1) the chemostat was fed with 11.7 ± 0.7 g toluene m-3 during the first 21 days and thereafter with 20.2 ± 1.9 g toluene m-3. Gaseous toluene and CO2/O2 concentrations were daily monitored through valves A and C (Figure 1). Excreted metabolites, dissolved total organic carbon (TOC) content, pH, absorbance at 650 nm, and CFU plate counts in selective medium (toluene) and non-selective medium (peptone) were also daily recorded by withdrawing a 10 ml liquid sample through valve B under sterile conditions (Figure. 1). Dissolved Oxygen Concentration (DOC) and Temperature (T) were monitored on line. In addition, the specific O 2 consumption and the presence of catechol-2-3-dioxygenase were periodically monitored during toluene biodegradation by P. putida mt-2.
450
RAÚL MUÑOZ, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL AND
SANTIAGO VILLAVERDE
Figure. 1. Schematic representation of the experimental set up. 1, Compressed air; 2, Mass flow controller; 3, toluene evaporator; 4, mass flow controller; 5, air-filter; 6, stirred tank reactor, 7 MSM reservoir, 8 MSM pump. PC, pressure control; PI, pressure indicator.
2.3 ANALYTICAL METHODS Toluene, CO 2 /O 2 , and benzyl alcohol (BA) and benzoic acid (BAc) concentrations were analyzed by GC-MS, GC-TCD, and HPLC, respectively, according to Bordel et al. (2007). The determination of dissolved Total Organic Carbon (TOC), pH, DOC and dry weight were also performed out according to Bordel et al. (2007). Respirometric assays were carried out to determine the time course of the specific O2 consumption during toluene biodegradation. Reaction vessels were filled with 15.6 ml of MSM, 2 ml of P. putida mt-2 (previously centrifuged at 6000 rpm during 15 min and resuspended in fresh MSM) and supplied with toluene at 25 mg l-1 (from a toluene saturated MSM stock solution). The total reaction volume was 18.5 ml. Tests in the absence of metabolites and toluene were also carried out under similar conditions to serve as controls for endogenous respiration. In P. putida F1 culture the fraction of viable bacteria degrading toluene was determined by standard plate count of colony-forming units (CFU) on non-selective medium (casein peptone 15 g l-1, soymeal peptone 5 g l-1, NaCl 5 g l-1, and agar 15 g l-1) and selective medium (MSM supplied with toluene vapors). Cells were incubated in
LONG-TERM STABILITY OF PSEUDOMONAS PUTIDA
451
sealed containers for 24 h in triplicate at 30ºC. In P. putida mt-2 the fraction of toluene degrading cells was determined by periodically withdrawing samples from the bioreactor, diluting and spreading on agar plates with glucose as the sole carbon and energy source. After 1 day of growth, colonies were transferred and grid-arranged in triplicate to agar plates under non- selective (2 plates of glucose) and selective (1 plate grown on saturated toluene atmosphere) conditions. After 1 or 2 days of incubation at 30 ºC, colonies grown under non-selective conditions were subjected to the catechol2,3-dioxygenase test (see below) Plates grown under toluene atmosphere showed the TOL+ colonies. Absence of growth in selective medium plates indicated TOL- cells, which also corresponded with white colonies obtained in the catechol-2,3-dioxygenase (C23O) test. The presence of C23O was assayed according to Duetz et al. (1991). Between 450 and 750 colonies (after 1 day of growth at 30 ºC) on agar plates with glucose as the sole carbon and energy source) were screened by spraying the plates with a 250 mM catechol solution. Positive colonies (those with an intact TOL plasmid) turned bright yellow almost immediately, while those that had lost C23O activity remained white. Polymerase chain reaction (PCR) was used to search for the gene responsible for catechol-2,3-dioxygenase activity (xylE). The following specific oligonucleotides were designed: xylE-forward: 5’-TACTGGACATGAGCAAGGC-3’ and xylE-reverse: 5’-GATAGATGTGTCGGTCATGG-3’ (670-bp fragment). Nucleotide sequence was retrieved from Genebank, accession number AJ344068.1 (region 50914-51837) (Greated et al., 2002). 25-μL reactions had the following composition: 1 mM oligonucleotides (Roche-TIB MOLBIOL, Berlín, Germany); 80 μM dNTPs (Q-Biogen, Austin, Texas, USA); 0.1 U/25 μL Taq-polymerase (Roche, Barcelona, Spain); 1:10 dilution of Taq-polymerase buffer (Roche, Barcelona, Spain); volume up to 25 μL was completed with extra pure water. Temperature cycles were: initial denaturing step of 94ºC, 3 min; 30 cycles of 94ºC (40 sec), 62 ºC (40 sec), 72 ºC (40 sec); final elongation step 72ºC, 3 min. Cleaner PCR results were obtained when purified plasmid DNA was used as template (0.2 ng per reaction), than with direct colony-PCR. The 10-minute protocol for plasmid DNA isolation (Zou et al., 1990) was carried out for C23O- and C23O+ colonies. PCR results were visualized in 1.2%-agarose gels with 1xTAE buffer, according to standard protocols (Sambrook and Russel, 2002). Size changes on TOL plasmid was confirmed by restriction analysis of purified plasmid DNA. HindIII -fragment pattern of C23O- and C23O+ colonies was visualized on 0.6% agarose gels on 1xTAE (Sambrook and Russel, 2002).
452
RAÚL MUÑOZ, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL AND
SANTIAGO VILLAVERDE
3 RESULTS AND DISCUSSION When inoculated with P. putida mt-2 process collapse occurred after approx 3.5 days, regardless the toluene gas inlet concentration (9.1 and 3.2 g m-3). In both cases, bacterial activity (evaluated based on toluene elimination capacity (EC), CO2 production rate and specific respiration rate on toluene) increased during the first day of cultivation followed by a gradual decrease in both toluene EC and CO2 production (Figure 2a). This decay of bacterial activity corresponded with a steady decrease in biomass, TOC and benzoic acid concentrations, and an increase in the pH of the cultivation medium (Figure 2b,c). Likewise, the fraction of toluene degrading P. putida mt-2 started to decrease during the second day of cultivation reaching a steady state value of 73 ± 4 % until the end of the experimentation, which was independent of the loading rate scenario (Figure 2d). Non-toluene degrading variants of P. putida mt-2 did not become yellow after spraying with catechol, indicating no basal activity of the catechol 2,3 dioxygenase (Brinkmann et al., 1994). PCR and plasmid restriction analyses of non toluene degrading isolates indicated deletion of a significant portion of the pWW0 plasmid (Data not shown).
Figure 2. Time course of EC ( ), CO2 production ( ), biomass concentration (Δ), pH ( ), benzoic acid ( ), TOC ( ), and fraction of toluene degrading cells ( ) during toluene biodegradation by P. putida mt-2 operated at 9.1 g m-3 and D = 0.2 h-1.
LONG-TERM STABILITY OF PSEUDOMONAS PUTIDA
453
These results are in agreement with previous observations reported in literature. For instance, Villaverde et al. (1997) reported a decrease in the toluene culturable cell fraction down to 2 % after 3 month of operation in flat plate biofilm reactor exposed to toluene liquid concentrations of 7 mg l-1. In a similar study, Jones et al. (1997) observed that the number of toluene-mediated injured cells correlated negatively with the time of exposure and toluene concentration, although in both cases the reason of this significant bacterial deterioration were not identified. Several authors have pointed out that some intermediates from toluene biodegradation might induce severe irreversible defects in both plasmid and chromosomal-associated pathways in Pseudomonas cultures. Partial deletions or curing of the TOL plasmid have been already reported in P. putida mt-2 during growth on toluene, BA or BAc (Brinkmann et al., 1994; Leddy et al., 1995). Likewise, Duetz et al. (1991) recorded a total loss of the C23O activity in P. putida mt-2 grown on benzoic acid in a 120-h phauxostat culture. In our study, both benzoic acid and benzyl alcohol were always detected when operating at very low concentration as a result of toluene biodegradation via the TOL pathway, which might have promoted the loss of toluene catabolic capacity likely due to the deterioration of both plasmidic and chromosomal pathways (Brinkmann et al., 1994; Leddy et al., 1995). However, the decrease in the fraction of toluene degrading P. putida mt-2 cells observed in our study can not explain itself the severe decrease in process performance recorded in the process. Indeed, a constant rate of segregation of P. putida variants would finally render to a stable steady state, but never to bacterial washout. Further analysis of the physiological state of the culture based on respirometric tests suggested an accumulative toluene-mediated microbial damage, which could explain the deterioration of the microbial culture and therefore process instability. Thus, the specific respiration rate on toluene steadily decreased after 1 days of cultivation regardless the toluene inlet concentration (Figure 3). Aromatic compounds such as toluene can act as uncoupplers diminishing the efficacy of the respiratory chain and therefore reducing microbial metabolic activity (de Smet et al., 1978). These aromatic pollutants are also known to increase cell permeability by damaging the cytoplasmatic membrane, which contributes to the loss of ions, cytoplasmatic material and proton gradients. In this regard, Dominguez-Cuevas et al. (2006) have recently suggested that membrane damage in a pWW0 bearing P. putida KT2440 (similar to P. putida mt-2) strain induced high levels of oxidative stress and inhibition of the most productive steps of the aerobic metabolism. On the other hand, P. putida F1 exhibited stable performance at all tested toluene inlet concentrations (up to 20.2 g m-3) as shown by the constant values of EC, CO2 production, pH and biomass concentration within each steady state (Figure 4a,b). Thus, no mutant population established in the culture at the expense of the organic carbon excreted by the wild type P. putida F1 cells. This was confirmed by two facts:
454
RAÚL MUÑOZ, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL AND
SANTIAGO VILLAVERDE
Figure 3. Time course of the specific toluene respiration rate in P. putida mt-2 cultures operated at high ( ) and low ( ) toluene loadings.
Figure 4. Time course of EC ( ), CO2 production ( ), biomass concentration (Δ), pH ( ), benzoic acid ( ), TOC ( ), and fraction of toluene degrading cells ( ) during toluene biodegradation by P. putida F1 operated at D = 0.1 h-1 and 3.3 g m-3 and 6.2 g m-3.
LONG-TERM STABILITY OF PSEUDOMONAS PUTIDA
455
Firstly, TOC and BA remained constant within each steady state (Figure 4c), and secondly, no significant differences in the number of CFU growing on selective and non-selective medium was observed throughout the experimentation (Figure 4d). Benzyl alcohol, a curing agent causing TOL plasmid deletion and damage at chromosomal level in Pseudomonas strains, was present in the cultivation medium as a result of the monooxygenation of toluene by the dioxygenase system of P. putida F1 (Bordel et al., 2007). However, despite being exposed to BA and inlet gas toluene concentrations of up to 60 mg l-1 and 20.2 g m-3, respectively, there was no significant effect on the overall process performance over 1 months of operation. Both P. putida strains exhibited similar performances in terms of EC and CO2 production. Thus, when the process was operated at D = 0.1 h-1 and 3.3 g toluene m-3, toluene was degraded at rates of approx. 200 g m-3 h-1. However, toluene is sensed as a stressor rather than a nutrient by P. putida mt-2. Based on expression profiles, Dominguez-Cuevas et al. (2006) reported that in the presence of toluene, the bulk of the available transcriptional machinery of a pWW0-bearing P. putida is reassigned to endure general stress, while only a small share is redirected to the degradation of the aromatic compound.
4 CONCLUSIONS Despite the numerous studies on the effect of toluene-mediated mutations in Pseudomonas putida cultures, no systematic studies of the influence of these mutations on process performance (EC, CO2 production) have been, to the best of our knowledge carried out. This work demonstrated that under high toluene loadings P. putida F1 exhibited a much more stable performance than P. putida mt-2. Process collapse in P. putida mt-2 cultures occurred in approx 4 days as a result of membrane damage and further inhibition of the aerobic metabolism as shown by a decaying toluene respiratory activity. No significant decrease in the overall performance of the biodegradation process was however recorded over 32 days of operation when using P. putida F1. Therefore, the selection of highly resistant bacterial strain is crucial in order to ensure not only the viability but also the long term stability of the biodegradation process specially when operating at high loading rates.
5 ACKNOWLEDGEMENTS This research was supported by the Spanish Ministry of Education and Science (PPQ2006-08230 and JCI-2005-1881-5 contracts).
456
RAÚL MUÑOZ, ANTONIA ROJAS, LUIS FELIPE DÍAZ, SERGIO BORDEL AND
SANTIAGO VILLAVERDE
REFERENCES Bordel, S., Diaz, L.F., Muñoz, R. and Villaverde, S. (2007) New insights on toluene biodegradation by Pseudomonas putida F1: Influence of pollutant concentration and excreted metabolites. Appl. Microbiol. Biotechnol. 74: 857-866. Brinkmann, U., Ramos, J.L. and Reineke, W. (1994) Loss of the Tol meta-cleavage pathway functions of Pseudomonas putida strain PaW1 (pWW0) during growth on toluene. J. Basic Microbiol. 34 (5): 303-309. de Smet, M.J., Kingma, J. and Witholt, B. (1978) The effect of toluene on the structure and permeability of the outer and cytoplasmic membranes of Escherichia coli. Biochim. Biophys. Acta 506: 64-80. Dominguez-Cuevas, P., Gonzalez-Pastor, J.E., Marques, S., Ramos, J.L. and de Lorenzo, V. (2006) Transcriptional tradeoff between metabolic and stress-response programs in Pseudomonas putida KT2440 Cells exposed to toluene. J. Biol. Chem. 281(17): 11981-11991. Duetz, W.A., Winson, M.K., van Andel, J.G. and Williams, P.A. (1991) Mathematical analysis of catabolic function loss in a population of Pseudomonas putida mt-2 during non-limited growth on benzoate. J. Gen. Microbiol. 137: 1363-1368. Greated, A., Lambertsen, L., Williams, P.A. and Thomas, C.M. (2002) Complete sequence of the IncP-9 Tol plasmid pWW0 from Pseudomonas putida. Environ. Microbiol. 4(12): 856-871. Jones, W.L., Mirpuri, R.G., Villaverde, S., Lewandowski, Z. and Cunningham, A.B. (1997) The effect of bacterial injury on toluene degradation and respiration rates in vapor phase bioreactor. Water Sci. Technol. 36(1): 85-92. Leddy, M.B., Phipps, D.W. and Ridgway, H.F. (1995) Catabolite-mediated mutations in alternate toluene degradative pathways in Pseudomonas putida. J. Bacteriol. 177(16): 4713-4720. Mirpuri, R., Jones, W. and Bryers, J.D. (1997) Toluene degrading kinetics for planktonic and biofilm grown cells of Pseudomonas putida 54G. Biotechnol. Bioeng. 53: 535-46. Oliveira, T.A.C. and Livingston, A.G. (2003) Bioscrubbing of waste gas-substrate absorber to avoid instability induced by inhibition kinetics. Biotechnol. Bioeng. 84(5): 552-563. Sambrook, J. and Russel D.W. (2002) Molecular cloning: a laboratory manual. Third edition. Cold Spring Harbour aboratory Press, Cold Spring Harbor, New York. Song, J.H. and Kinney, K.A. (2005) Microbial response and elimination capacity in biofilters subjected to high toluene loadings. Appl. Microbiol. Biotechnol. 68: 554-559. Villaverde, S., Mirpuri, R.G., Lewandowski, Z. and Jones, W.L. (1997) Physiological and chemical gradients in a Pseudomonas putida 54G biofilm degrading toluene in a flat plate vapor phase bioreactor. Biotechnol. Bioeng. 56: 361-71. Worsey, M.J. and Williams, P.A. (1975) Metabolism of toluene and the xylenes by Pseudomonas putida arvilla (mt-2): evidence for a new function of the TOL plasmid. J. Bacteriol. 124: 7-13. Zou, C., Yang, Y. and Y. Jong A.Y. (1990). Miniprep in Ten minutes. Biotechniques 8 (2): 172-173.
457
Development of a novel bioscrubbing process for complete treatment of NOX from flue gases SANJEEV S.R. ARJUNAGI AND LIGY PHILIP* Department of Civil Engineering, Indian Institute of Technology, Madras, Chennai-600 036, India
ABSTRACT In this study, a novel bioscrubber was developed for the complete treatment of NOx from flue gases. As a first step, an autotrophic ANNAMOX system was developed using ammonia as the electron donor and nitrate as electron acceptor generating nitrogen gas as the reaction product. Once an efficient ANNAMOX culture was developed, nitrate was replaced by NOx. Initially synthetic flue gas with a composition of 80% N2, 19% CO2 and 100 ppmv was fed to the bioscrubber with an EBRT of 60 sec. The system showed a NOx removal of 20-25 %. The synthetic flue gas with 3-5% oxygen also showed the same NOx removal efficiency. In the latter case, more ammonia consumption in the system was noted. However, there was no nitrate accumulation in the system in both the cases. Inorder to improve the NOX removal efficiency, NO in the flue gas was partially oxidized to NO2 with the help of ozone (one mole of ozone per mole of NO with an EBRT of 10 sec) and fed to the bioscrubber. The NOx removal efficiency in the system was improved to 75-80%. The bioscrubber was able to remove more than 90% of the generated NO3-. Performance of the reactor is being monitored at different EBRTs. The new system developed seems to be a promising alternative for the complete treatment of NOx from flue gases in an environmentally friendly way.
1. INTRODUCTION The presence of oxides of nitrogen (NOx) in the ambient air has been, and still is, of great concern because of the toxicity of individual compounds or the secondary pollutants produced by the reaction of NOx with hydrocarbons and other chemicals such as ozone in presence of sunlight (Wark and Warner, 1981). Nitrogen oxides (NOx) is a collective name of six compounds namely nitrous oxide (N2O), nitrogen
457
458
SANJEEV S. R. ARJUNAGI AND LIGY PHILIP
dioxide (NO2), nitrogen trioxide (N2O3), nitrogen tetra oxide (N2O4) and nitrogen pentoxide (N2O5). Among these, NO and NO2 are of major concerns in air pollution control. From the air pollution point of view, NO and NO2 are of major concern. NO exists in very low concentrations in the atmosphere and is converted to NO2 at higher concentrations. About 44% of NOx pollution is contributed by mobile sources, 55% by stationary sources and the remaining by solid waste disposal and miscellaneous processes. NOX is responsible for troposphere ozone and urban smog through photochemical reactions. Further, NOx together with SOx is the major contributor to the «acid rain» that harms forest crops, buildings, as well as aquatic life. Since 1970, The U.S. Environmental Protection Agency (EPA) has tracked emissions of the six principal air pollutants – carbon monoxide, lead, nitrogen oxides, particulate matter, sulfur dioxide, and volatile organic compounds. Emissions of all of these pollutants have decreased significantly except for NOx, which has increased approximately 10%, over this period and nitrogen oxides are facing increasingly stringent regulations due to Clean Air Act Amendments of 1990. The oxides of nitrogen are formed by the direct combination of oxygen and nitrogen during a variety of thermal processes. Various stationary sources that emit NO2 are power plants, utility boilers, steel industries, ceramic industry, nitric acid manufacturing industry, oil refineries, ammonia manufacturing industry, fertilizer manufacturers, pickling operations in anodizing plants and nylon intermediate plants. Concern for environmental and health issues coupled with stringent NOx emission standards indicate a need for the development of efficient low-cost NOx removal technologies. Conventional NOx control technologies include combustion modifications, dry processes and wet processes. The combustion modification technologies such as reburning, flue gas recirculation, and low NOx burners, are considered to be suitable for operations with single digit NOx emission levels (Devahasdin et al., 2003). Dry processes include selective catalytic/non-catalytic reduction (SCR/SNCR) of NOX to N2 with ammonia, urea, and hydrocarbons. Wet processes include absorption with liquid phase oxidation, absorption with liquid phase reduction and gas phase oxidation followed by absorption. Among the above mentioned NOx emission control technologies, the combustion modification (e.g. low NOx burners) and SCR are the most popular methods (Fujishima et al., 1999). Combustion modification or pretreatment, a subsidiary of a NOx treatment system, aims at keeping NOx production at low. For dry systems such as selective catalytic reduction (SCR) and selective non catalytic reduction (SCNR), elevated temperatures (900-1000 0C for SNCR and 270-400 0C for SCR) are needed for reasonably good process efficiency. Wet systems have the drawbacks of expensive chemical additives, high water usage, and safety risks from handling of chemicals.
DEVELOPMENT OF A NOVEL BIOSCRUBBING PROCESS FOR COMPLETE TREATMENT OF NOX
459
However, most of these control technologies are converting the pollutant from one phase to another or they are highly sensitive and costly. Application of biological process for the flue gas control seems, to be environmentally friendly, economical and sustainable technology. Biological removal of NOx from contaminated gas streams is emerging as a novel treatment method. Recently, biofiltration / biotrickling filtration is tried as an attractive alternative for both organic and inorganic air pollutants treatment. It is reported that NO2 and SO2 could be removed effectively using biotrickling filter / scrubber (Flanagan et al., 2002; Philip and Deshusses, 2003) due to their high solubility in water. However, not many studies have been carried out on the treatment of acid gases like SOx and NOx. Moreover, the tried technologies have many disadvantages like inhibition in presence of oxygen and need high detention times to achieve the necessary degree of treatment. As the solubility of NO in water is very low, biotrickling filter or scrubber is not able to remove NO effectively at a short contact time. Thus, conversion of NO to NO2 or any other soluble form using suitable technical method and subsequent scrubbing followed by biological denitrification seems to be a viable alternative Barman and Philip (2005) reported a novel and effective system for the complete treatment of NOx from flue gases. The system consisted of photocatalytic or ozone oxidation of NOx, followed by scrubbing and biological denitrification. Maximum photocatalytic oxidation of NOx was achieved while using powdered TiO2 at a catalytic loading rate of 10 g/ h, relative humidity of 50%, and a space time of 10 s. The used catalyst was regenerated and reused. A total of 72% of oxidized NO was recovered as HNO3/HNO2 in the regeneration process. Stochiometrically, 10% excess ozone was able to affect 100% oxidation of NO to NO2. Presence of SO2 adversely influenced the oxidation of NO by ozone. The scrubbing of NO was effective with distilled water. Heterotrophic denitrifiers were able to denitrify the leachate with an efficiency of 90%, using sewage (COD 450 mg/L) as electron donor. In this system, for the complete treatment, one needs to have three units. It is always advisable to have minimum number of units for the ease in operation and maintenance. Hence, it is essential to try whether the scrubbing and biological denitrification can be achieved in one system. In this paper, performance of a biosrubber employing anaerobic ammonia oxidizing bacteria for the removal of NOx from flue gas was evaluated. Effect of operational parameters such as empty bed contact time, and presence of oxygen on the performance of the system was also studied.
460
SANJEEV S. R. ARJUNAGI AND LIGY PHILIP
2. MATERIALS AND METHODS 2.1 CHEMICALS The chemicals used in this study were procured from Ranbaxy Fine Chemicals Limited. A-3, Okhla Industrial Area, Phase-I New Delhi-110 020, India. Nitric Oxide gas was procured from Bhoruka Gas Limited. Mahadevpua Road, White Field, Bangalore, India. 2.2 SEED SLUDGE The initial seed sludge was collected from the Sewage treatment plant of IIT Madras Chennai. Cow dung sludge acclimatized for ANNAMOX process in Environmental Engineering Laboratory, Department of Civil Engineering, IIT Madras, Chenani was also used for the study. Settled sludge was added to the reactor (50% of reactor volume, MLSS = 1150 mg/L) for acclimatization. Acclimatization was carried out in twelve separate batch reactors (each having a volume of 100 mL, fitted with air tight septum), under anaerobic environments. During acclimatization, nutrients were added to the reactors for the growth of bacteria. The composition of mineral media is KH2PO4, 0.0572g; NaHCO3, 2.1g; NaNO2, 0.5915g; CaCl2.2H2O, 0.3g; MgSO4.7H2O, 0.2g; FeSO4, 0.00625g; EDTA, 0.00625g; (NH4)2SO4, 0.455g; Trace element solution, 1mL in one liter distilled water. 2.3 ANALYTICAL METHODS The reactor influent and effluent were analyzed as per Standard Methods (APHA, 1998) in order to monitor the performance of the biological systems. COD of liquid samples was estimated as per standard methods (Ref. No. 5220 Chemical Oxygen Demand, APHA, 1998). Closed reflex method was followed. To estimate the NH3 –N, NO2--N and NO3--N concentration, Dionex LC-20 Ion Chromatograph with ED-50 electrochemical detector and IP-25 isocratic pump was used. 2.4 ESTIMATION OF NITRIC OXIDE (NO), NITROGEN DIOXIDE (NO2) NO and NO2 concentration in inlet and outlet gas stream were measured by flue gas analyzer (Quitnox KM 9106). Range NO: 0-5000 PPM, NO2: 0- 1000 PPM. Accuracy: +/- 1%. The samples were analysed for the pH, MLSS, MLVSS, and Dissolved Oxygen as per the methods described in APHA, 1998. 2.5 EXPERIMENTAL METHODS 2.5.1 DEVELOPMENT OF ANAMMOX CULTURE FOR NOX REMOVAL The experiments were carried out in batch systems. A sample volume of 100 mL was taken in a number of serum bottles. Each sample contained 100 mL of minimal medium with 140mg/L of Ammonium-Nitrogen and 84.5 mg/L Nitrate-Nitrogen. The
DEVELOPMENT OF A NOVEL BIOSCRUBBING PROCESS FOR COMPLETE TREATMENT OF NOX
461
study was carried out at 25oC in an orbital shaker at 150 rpm under anaerobic environment. Samples were collected at different time intervals and were analyzed for NH4+, NO3-, COD and biomass concentrations. In the second set of experiments, all other conditions were kept the same except the electron acceptor. Instead of NO3-, 152.1 mg/L NO2 --N was supplied to the system. 2.5.2 FED BATCH SYSTEM The ANAMMOX reactors were operated until the removal of nitrite, nitrate and ammonia reached 80% and the concentrations were brought back to initial concentrations to study the performance of the system in the subsequent removal of nitrite, nitrate and ammonia. Based on the above study, cow dung sludge performed better in presence of nitrite as electron acceptor. Hence, cow dung sludge with nitrite as electron acceptor was used for continuous system. 2.5.3 CONTINUOUS REACTOR STUDIES The reactor was made out of clear schedule 40 PVC pipe (ID=4cm). The total length of the reactor was 60 cm and bed height was 50 cm. The gas inlet and outlet ports were located at the bottom and top lids of the reactor. The liquid inside the column consisted of mineral medium with the composition specified earlier. There was no external carbon source supplied to the column except for CO2 and carbonate. The reactor was seeded with about 300 mL of digested cow dung sludge containing a MLSS concentration of 1000 mg/L and remaining 700 mL contained mineral media. During the start-up of the continuous experiment, 150 mg/L of NO2- -N and 230 mg/ L of NH4+-N was maintained in the system throughout, since 150-200ppm of NOx gas was supposed to enter the system daily, with a gas flow rate of 1L/min with a NOx concentration of 150ppm. Based on the stochiometry the above concentrations were chosen. The system was operated with a Hydraulic retention time of 24 h. Once the system was stabilized, liquid nitrate feed was replaced by gaseous NO feed. Simulated flue gas prepared by mixing a metered flow of approximately 1% NOx (5000 ppm), 10% CO2, 89% N2. The gases were homogenized in a mixing chamber and fed to the bioreactor at a flow rate of 1000 mL/min with the help of rotameter. Initial studies were carried out in the absence of oxygen in the system with an EBRT of 60 seconds. The EBRT was gradually reduced upto 10sec. Once, satisfactory performance was observed in the system, 5% of oxygen was introduced in the simulated gas stream and the reactor was operated with a detention time of 60 seconds. The inlet and outlet NO and NO 2 concentrations were measured, ammonia and nitrite concentrations were also measured from the liquid effluent. The mineral media was added to the system and effluent was removed from the system from bottom and allowed to settle in a closed reactor. The supernatant was thrown and the settled biomass
462
SANJEEV S. R. ARJUNAGI AND LIGY PHILIP
was again added to the system. Further, in order to increase the efficiency of the system, NO was oxidizesd to NO2 with the help of ozone. 2.5.4 OXIDATION OF NO BY OZONE The experiment was conducted in a 5 L borosil bottle at room temperature 0 (27 C). Inlet NOx concentration was maintained at 150 ppm (NO: 145 ppm and NO2: 5 ppm) and space time (10sec) was adjusted by controlling the inlet stream flow rate. Ozone was supplied from ozone generator (Vortex, India) and flow was controlled by controller.
3 RESULTS AND DISCUSSION 3.1 PERFORMANCE OF REACTORS DURING BATCH STUDY One of the objectives of the study was to check whether NO can be used as an electron acceptor by ANAMMOX bacteria for the generation of di-nitrogen gas. If this route is possible, control of NOx from flue gases through biological systems can be achieved easily. To check this concept, reactors were developed for the removal of different concentration of nitrite and nitrate in presence of ammonia under anaerobic conditions. Once a strong ANAMOX culture is developed, it may be easy to replace nitrite and nitrate with NOx. The performances of these reactors were monitored in terms of COD, nitrite, nitrate and ammonia removal. The MLSS concentrations in the reactors were also monitored periodically. 3.2 PERFORMANCE OF ANAMMOX REACTOR Ammonia was one of the critical parameters to be monitored to determine the efficiency of ANAMOX process. Here in this study, ammonia removal in presence of both nitrite and nitrate as electron acceptors were monitored. Two reactors were operated with different seed sludge. Figure 1 a and 1b shows the result of ammonia oxidation in presence of nitrite and nitrate as electron acceptor by both the cultures. Ammonia removal was more in presence of cow dung sludge with nitrite as electron acceptor. This may be due to the prior acclimatization of the sludge for ANAMMOX process. In case of NO3- as electron acceptor, both the sludges performed almost in the same manner. The sludge collected from IIT Madras lagoon may be having considerable population of nitrifiers and denitrifiers. The initial concentration of NH4+-N was 140mg/L and after the test run of 40 days the concentration was reduced to 42 mg/L. NO2- - N concentration changed from 182 mg/L to 38 mg/L. The stochiometry of the reaction is as follows NH4+-N / NO2- - N = 0.69
DEVELOPMENT OF A NOVEL BIOSCRUBBING PROCESS FOR COMPLETE TREATMENT OF NOX
463
(a) (b) Figure 1. Ammonia oxidation with a) Nitrite b)Nitrate as electron acceptor.
As per reported ANAMMOX Studies (Strous et al., 1998), the ratio of NH4+-N / NO2 - N is 1.3. Stochiometrically, the reported value and the one obtained in the present study are not matching well. This may be due to the presence of nitrifiers and denitrifiers in more number in the system. A significant quantity of ammonia might be getting oxidized rto nitrite/nitrate in the system. It is reported that anoxic nitrification is possible by certain group of microbes under oxygen stress conditions. When nitrate was employed as electron acceptor, the initial concentration of NH4+-N was 140mg/L and after the test run of 40 days the concentration reduced to 52 mg/L. NO3- - N concentration changed from 88 mg/L to 21 mg/L The stochiometry of the reaction is as follows -
NH4+-N / NO3- - N =1.31 As per reported ANAMMOX Studies (Mulder et al., 1998), the ratio of NH4+N / NO3- - N is 0.6. Stochiometrically, the reported value and the one obtained in the present study are not matching well. From this it may be inferred that, the developed sludge has more quantity of denitrifiers. They might have reduced the nitrate to nitrogen gas utilizing the available COD in the system. There was a significant quantity of COD reduction in the system. The initial MLSS concentration of reactor with cow dung sludge was 1500mg/ L. MLSS concentration gradually increased and after the end of batch study (40 days) the MLSS concentration was 1760mg/L Reactor with IIT lagoon sludge was operated with an initial MLSS concentration of 2100mg/L. At the end of the study, MLSS concentration in the system was 2250mg/L. Though, the reactor with IIT lagoon sludge had more MLSS, the performance was better in the reactor with cow dung sludge. This must be due to the prior acclimatization of this sludge for ammonia oxidation using nitrite/nitrate as electron acceptors.
464
SANJEEV S. R. ARJUNAGI AND LIGY PHILIP
3.3 PERFORMANCE OF REACTORS DURING CONTINUOUS STUDY The batch study clearly showed that cow dung sludge is very effective in the removal of nitrite and ammonia (89% of nitrite and 69% of ammonia) and hence, cow dung sludge was used for the continuous studies. The system was run continuously on a sequential batch mode with a detention time of 24 h and with initial concentration of 150mg/L of NO2. Initially the study was carried out without introducing NOx gas. The significance of this study was to make the system get acclimatized to nitrate removal with ammonia oxidation. The amount of nitrate was chosen in such a way that, that will be the total nitrite/nitrate generated in the system by the scrubbing of NOx from flue gas for a flow rate of 1L/min with a concentration of 100 ppmv during 24 h. The removal of nitrite and ammonia soon after 24 h was monitored. Every 24 h interval, the concentration of nitrate and ammonia was made up to the initial concentration. 3.4 AMMONIA REMOVAL WITH NITRITE AS ELECTRON ACCEPTOR IN THE CONTINUOUS SYSTEM The study of ammonia oxidation with nitrite as electron acceptor was carried out in batch systems with a longer hydraulic retention time. To assess the ability of developed ANAMOX culture to remove nitrate, a continuous system was operated with a high loading. The results are presented in Figures 2a and 2b.
Figure 2a. Ammonia removal in the system.
Figure 2b. Nitrite removal in the system.
After 25 days of acclimatization, the continuous system showed 71% of ammonia removal and 76% of nitrite removal with a detention time of 24 h. The initial concentration of NH4+-N was 233 mg/L and after 23 days of continuous operation the concentration reduced to 131.4 mg/L. The corresponding nitrite removal was from 152 mg/L to 35 mg/L. The stoichiometry is as follows NH4+-N / NO2- - N = 0.87. Hence from the above study, it can be inferred that ammonia oxidation and nitrite removal is due to the presence of nitrifiers and denitrifiers.
DEVELOPMENT OF A NOVEL BIOSCRUBBING PROCESS FOR COMPLETE TREATMENT OF NOX
3.5 PERFORMANCE
OF REACTORS DURING CONTINUOUS STUDY AFTER INTRODUCING
465
NOX
GAS
3.5.1 STUDY IN ABSENCE OF OXYGEN Initially the study was carried out in the absence of oxygen. During the entire study period, inlet NOx concentration of 100 ppm was maintained in the system and the outlet concentration was measured regularly. The results of the study is presented in Fig 3 a. After the system was operated continuously for 11days, the outlet NOx concentration was reduced to 78 ppm. The above results indicate that, the system performance in NOx removal was very less with a percent NOx gas removal of 22%. This may be due to the very low solubility of NOx gas. 3.5.2 STUDY IN PRESENCE OF OXYGEN Flue gas always has 3-8% of oxygen. The performance of the system was evaluated in presence of 5% oxygen. Ammonium gets oxidized to nitrite in presence of oxygen with the help of autotrophic nitrifiers and ammonium reacts with nitrite to produce nitrogen gas. This is the possible mechanism in presence of oxygen. Study was carried out with an inlet NOx concentration of 100 ppm along with 5% oxygen keeping all other parameters the same. Outlet NOx concentration was monitored regularly. After 11 days the continuous operation, the outlet NOx concentration was 80ppm and the percentage removal of NOx was 20% (Figure 3b). Though presence of oxygen did not hinder the activity of the biological system, the overall removal efficiency was very low. The low solubility of NOx must be the reason for this.
Figure 3a. NOx removal in absence of oxygen. Figure 3b. NOx removal in presence of oxygen.
466
SANJEEV S. R. ARJUNAGI AND LIGY PHILIP
3.6 PERFORMANCE OF REACTORS DURING CONTINUOUS STUDY AFTER OZONE OXIDATION From earlier study results, it is clear that, using the new biosystem, it is not possible to remove NOx in the form of NO due to its low solubility. Hence, in order to increase the efficiency of NOx removal, first NO was oxidized with the help of ozone and allowed the NO2 thus formed to enter in the biosystem Inlet and outlet NO and NO2 concentrations were measured along with nitrite and ammonia removal in the aqueous phase of the biosystem. The optimization of NO oxidation by ozone was done by Barman and Philip (2005). These results were used for the present study. A NOx concentration of 200 ppm (NO: 175 ppm and NO2: 25ppm) was used. Ozone flow was gradually increased and the oxidation of NO to NO2 was monitored. It was found that almost all NO was converted (99%) to NO2 when slightly excess (10%) of stoichiometric amount of ozone was passed. 3.7 NOX REMOVAL AFTER OZONE OXIDATION NOx removal studies were carried out with an inlet NOx concentration of 150 ppm (NO2=145 ppm; NO =5 ppm) and with an EBRT of 60 sec. Gradually the contact time was reduced to 30 sec and then to 10sec. The results are presented in Figure 4 a. The results clearly indicate that the efficiency of NOx removal has increased significantly after the partial oxidation of NO to NO2. After 15days of operation, a NOx removal efficiency of 79% was achieved in the bio-system. In order to understand the system behavior at varying detention times, the study was carried out with 30sec and 10sec. when the detention time was reduced to 30 sec., the efficiency of the system was reduced to 60%. Further reduction in EBRT to 10 sec. deteriorated the performance of the system. This must be due to the less contact of the pollutant with the bioscrubbing medium. The addition of packing material or reducing the bubble size by modifying the diffuser marginally improved the efficiency of the system. 3.8 NITRATE REMOVAL IN THE SYSTEM Fig 4b shows the nitrate removal from the system after ozone oxidation. Initially the nitrate concentration in the system was increased during the study. But after 3 days, nitrate concentration gradually came down and reached a steady state in 12 days.. After reaching a steady state nitrite removal of 75%, there was no further improvement. The residual nitrate concentration may be the result of ammonia oxidation. However, as the nitrate loading to the system was increased (by reducing the EBRT and increasing the gas flow rate) the residual nitrate concentration at steady state also was increased marginally. At 30 Sec retention time, 55% of nitrate removal was observed where as 10sec retention time (gas) gave a nitrate removal efficiency of 45%.
DEVELOPMENT OF A NOVEL BIOSCRUBBING PROCESS FOR COMPLETE TREATMENT OF NOX
Figure 4b. Nitrate removal in the system.
467
Figure 4a. NOx removal in the continuous system after ozone oxidation.
3.9 NOX REMOVAL IN THE CONTROL REACTOR The system was operated at a detention time of 60 sec. The results are presented in Figure 5a. There was about 74% NOX removal initially and gradually the NOx removal efficiency was reduced to 68%. From this result, it is clear that, the major mechanism of NO2 removal is scrubbing. When microbes are present in the system, the accumulated nitrate is utilized. This gives a favorable concentration gradient for NOx removal. This may be the reason for a better performance while microbes are present in the system. 3.10 NITRATE REMOVAL IN THE CONTROL REACTOR Figure 5 b shows the abiotic nitrate removal from the system. As the time progressed, the nitrite concentration was kept on increasing. This shows the absence of autotrophic ANAMMOX culture, which was earlier responsible for the nitrite removal from the system 150 ppm of NO2 gas was allowed to enter the system for a detention time of 1 min and this was getting converted to nitrate in the system. In presence of autotrophic culture the system was responsible for removal of nitrate to 75% but in the absence of microbes nitrate was accumulated in the system.
4. CONCLUSIONS The following conclusions can be made based on the investigation carried out in the present study. Batch experiments indicated that, ammonia oxidation and simultaneous nitrite removal is possible in presence of cow dung sludge. Maximum removal efficiency of 79% of NOx gas at detention time of 60 sec, 60.1% at 30 sec, 40.33% at 10 sec was achieved after the ozone oxidation. Less removal of NOx was observed before ozone oxidation due to the low solubility of NO gas. The biological system containing ANAMMOX cultures seems to be a feasible option of treating nitrogen oxides from flue gases.
468
SANJEEV S. R. ARJUNAGI AND LIGY PHILIP
Fig. 5a. NOx removal from the system (Blank).
Fig. 5b. Nitrate accumulation in the system (Blank).
REFERENCES APHA, AWWA, APCF. (1995) Standard Methods for the Examination of Water and Wastewater. Devahasdin, S., Fan, C., Li, K. and Chen, D.H. (2003) TiO2 photocatalytic oxidation of nitric oxide. J. Photochem. Photobiol. A: Chemistry 156: 161-170. Fujishima, A., Hashimoto, K. and Watanabe, T. (1999) TiO2 Photocatalysis Fundamentals and Applications, BKC Inc., Tokyo. Flanagan, W.P., Apel, W.A., Barnes, J.M. and Lee, B.D. (2002) Development of gas phase bioreactors for the removal of nitrogen oxides from synthetic flue gas stream. Fuel. 81: 1953-1961. Ligy Philip and Deshusses, M.A. (2003) Sulfur dioxide treatment from flue gases using a biotrickling filter-bioreactor system». Environ. Sci. Technol. 37(9): 1978-1982. Mulder, A. (2003) The quest for sustainable nitrogen removal technologies. Wat. Sci. Technol. 48 (1): 67-75. Sanjoy Barman and Ligy Philip. (2006) Integrated system for the treatment of oxides of nitrogen from flue gases. Environ. Sci. Technol. 40(3): 1035 -1041. Strous, M., Heijnen, J.J., Kuenen, J.G. and Jetten, M.S.M. (1998) The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms. Appl Microbiol Biotechnol. 50: 589-596. Wark, K. and Warner, C.F. Air Pollution: Its Origin and Control; Harper and Row Publishers; New York, 1981.
469
Dichloromethane removal using mixed cultures in a biofilter and a modified rotating biological contactor – start up studies R. RAVI1, LIGY PHILIP2 AND T. SWAMINATHAN1* 1
Department of Chemical Engineering, Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600 036, India 2
ABSTRACT Dichloromethane (DCM) is a widely used organic solvent which is considered to be hazardous air pollutant. Regulatory standards in many countries require its removal from waste gas streams. Biological waste gas treatment is an attractive and environmental-friendly alternative to physicochemical methods. Volatile organic compounds (VOCs) in waste gases can serve as energy source and/or carbon source for the microbial metabolism. Biofilters and biotrickling filters, the widely used bioreactors, suffer from limitations such as control of operating parameters, pH, humidity and nutrient supply and clogging due to overgrowth of biofilm. To overcome these drawbacks, a modified rotating biological contactor (RBC) has been proposed which can retain the advantages of conventional systems. A conventional RBC system containing 20 acrylic discs 21 cm diameter and 5mm thickness with a disc spacing of 10 mm was modified by adding a leak tight cover and baffles between disks to avoid short circuiting of flow. The biofilm was formed on the discs with inoculum pre acclimatized to DCM at low concentration. The RBC was operated at an inlet concentration of 0.15 – 0.2 g/m3 at a gas flow rate of 0.06 3 m /h corresponding to an empty bed residence time (EBRT) of 2.5 min for 38 days resulting in a steady state removal of 84%. The residual DCM concentration in liquid phase was 5ppm and dissolved oxygen level was 3-4 ppm. pH decreased from 7 to 4.5 in the media, which indicated biodegradation and formation of acidic metabolites. The performance of RBC was compared with that of a biofilter packed with a mixture of garden compost and ceramic beads. The biofilter was operated at an inlet concentration of 0.15 – 0.21 g/m3 and at a gas flow rate of 0.06 m3/h corresponding to an empty bed residence time (EBRT) of 1.47 min for 90 days to reach steady state removal efficiency of 88%. Thus RBC system seems to be a potentially alternative to biofilter.
469
470
R. RAVI, LIGY PHILIP AND T. SWAMINATHAN
1 INTRODUCTION With the increasing usage of synthetic chemicals, particularly the petrochemicals, a new class of air pollutants, the Hazardous Air Pollutants (HAPs) have become a matter of concern in recent years. The list of hazardous air pollutants includes a number of volatile organic compounds (VOCs) which are commonly used as industrial solvents. Dichloromethane (DCM) is an important VOC used primarily for metal degreasing and paint removal. Due to its low boiling point (40.10 C) and high vapor pressure (47 Kpa at 200 C) significant amounts of DCM reach the environment (Wang et al., 2006). DCM is difficult to remove from contaminated water and air (Machey and Cherry, 1989) and shows high persistence in water (half life over 700 years) and atmosphere (half-life 79 days). The solvent DCM is known to be toxic to central nervous system at high exposure levels and there is suspected carcinogenicity in human liver and kidney (Green, 1991). Conventional VOC control techniques such as adsorption, absorption, scrubbing and condensation, thermal and catalytic incineration are generally energy intensive, ineffective and economically nonviable. Biological treatment systems are ecologically compatible alternatives to physicochemical methods (Kennes and Veiga, 2001). VOCs in waste gases can serve as energy source and or carbon sources for the microbial metabolism. DCM has been shown to be utilized by both aerobic and anaerobic bacteria as a sole carbon and energy source (Brunnel et al., 1980; Ritman, 1980; Krausova et al., 2003). A reasonable good number of studies have been carried on DCM biodegradation using pure culture (Dicks et al., 1994; Herbst et al., 1995). Biofiltration is a popular technique for removal of VOCs from contaminant air. It utilizes a support matrix packed in a column for microbial growth. The contaminants in the air stream are absorbed and metabolized by the microorganisms in the biofilm. Most biofilters that are in operation today can treat odors and VOCs effectively with efficiencies greater than 90%. A major limitation of biofilters is the increasing pressure drop during operation due to overgrowth of biofilm which requries periodical clean up. A potential alternative to overcome this limitation is to use rotating biological contactors (RBC) commonly used for waste water treatment. The RBC combines the advantages of low energy consumption and better control of biofilm growth. Ruediger (1999) first used RBC for waste gas treatment using DCM as the model pollutant. Vinage and von Rohr (2003) have assessed the long term performance of an RBC system for toluene removal at different loading rates, achieving an elimination capacity of 60 g/m3.h. It has been shown that microbial metabolism on the disc surface occurs through a series of phenomenological steps consisting of adsorption, absorption, diffusion and biodegradation. The present study focuses on the biodegradation of DCM in modified RBC and its comparison with a biofilter.
DICHLOROMETHANE REMOVAL USING MIXED CULTURES IN A BIOFILTER
471
2 MATERIALS AND METHODS 2.1 CHEMICALS All Chemicals used in this study were of laboratory grade: Dichloromethane (>99%), was purchased from Merck Limited, India. 2.2 MICROBIAL CULTURE AND MEDIA COMPOSITION The microbial mixed consortium was obtained from a municipal sewage treatment plant. The final inoculum was obtained by series of repeated inoculation in a mineral salt medium (MSM) that had the following composition: (NH4)2SO4 – 2.0 g/L, KH2PO4 – 2.0 g/L, NaCl – 0.5 g/L, MgSO4·7H2O – 0.025 g/L (Krausova, et al., 2003). The pH of the mineral salt media was adjusted to (7±1). 2.3 EXPERIMENTAL The schematic of the experimental setup is given in Figure 1. The VOC vapor generated by passing air at controlled rate through a VOC reservoir was mixed with humidified air in a glass mixing chamber to obtain the desired concentration of the feed gas stream. The biofilter consisted of a poly-acrylic tube (5’70 cm) having 6 sampling ports sealed with a rubber septa at 10 cm interval along the biofilter height. It was loosely packed with a mixture of compost and polystyrene spheres. A perforated plate at the bottom provided the support for the packing and also ensured uniform distribution of the vapor stream. The moisture content of the filter bed was maintained by prehumidifying the incoming vapor and by periodically sprinkling fresh media from the top of the biofilter. The modified RBC was made from two semi-cylindrical poly-acrylic parts with flanges to get a leak free operation. It contained 20 poly-acrylic discs each of 21 cm diameter, 5 mm thickness with 10 mm spacing between the discs. The discs were mounted on a stainless steel shaft and rotated slowly with variable speed motor. Baffles fixed to the top cover allowed the gas flow to pass through each chamber and increase the residence time of the gas in the reactor. The experimental runs were performed by passing the VOC vapor at different concentration and flow rates to vary the influent organic load. Samples were collected at the inlet and outlet at different intervals using gas tight syringe and analyzed for residual VOC concentration. 2.4 ANALYTICAL METHODS Residual concentration of DCM in the aqueous phase was measured in cell free samples using a Gas Chromatograph (AIMIL, India) fitted with a flame ionization detector and a (10% FFAP) packed column. Nitrogen was used as the carrier gas at a
472
R. RAVI, LIGY PHILIP AND T. SWAMINATHAN
Figure 1. Schematic of the experimental setup.
flow rate of 25 mL/min. The temperatures of injector, oven and detector were 200, 120 and 220°C respectively and retention time was 1.4 mins.
3 RESULTS AND DISCUSSION 3.1 STARTUP AND ACCLIMATIZATION OF BIOFILTER AND ROTATING BIOLOGICAL CONTACTOR (RBC) Startup studies were initiated with both biofilter and RBC in order to develop the required biofilm acclimatized to DCM as the main carbon source.
DICHLOROMETHANE REMOVAL USING MIXED CULTURES IN A BIOFILTER
473
3.1.1 BIOFILTER STARTUP Through a diverse microbial population exist in the compost, an external microbial consortium acclimatized to DCM in batch culture was added to the biofilter. Acclimatization is critical for successful operation of any biofilter and it depends on the activity of the microorganisms present in the compost.The biofilter was initially operated at an inlet concentration of 0.15 – 0.21 g/m3 and at a gas flow rate of 0.06 m3/h corresponding to an empty bed residence time (EMBRT) of 1.47 mins. The removal efficiency profiles were monitored continuously for 90 days till they achieved steady state. Figure 2 and 3 show the concentration profile of DCM and its removal efficiency in the biofilter. The removal efficiency during the first few days was high, which could be mainly due to the absorptive and adsorptive capacity of the compost. After this initial sorption phase, the biofilter efficiency decreased to 49%. However, there was a slow gain in the efficiency after 4 days. Zhu et al. (1998) have reported similar observation during the treatment of benzene vapors in biofilters containing compost and granular activated carbon as the packing material. Hodge and Devinny (1994) and Arulneyam and Swaminathan (2004) have also reported similar adsorption and biodegradation phases during startup operation in biofilters handling ethanol and methanol vapours. The removal mechanism shifted towards biodegradation and during this phase, the removal efficiency progressively increased with fluctuations, until becoming constant at about 88%.
Figure 2. Startup operation of the biofilter treating DCM vapors.
474
R. RAVI, LIGY PHILIP AND T. SWAMINATHAN
Figure 3. Removal efficiency of the biofilter treating DCM vapors.
3.1.2 STARTUP WITH RBC The RBC was initially inoculated with a seed culture acclimatized to DCM and operated at an inlet concentration of 0.25 – 0.3 g/m3 at a gas flow rate 0.06 m3/h for10 days. It was observed that fairly thick and uniform biofilm was formed on the disc surface. As the removal efficiency was observed to be low, the DCM concentration in the feed was reduced to 0.15 – 0.2 g/m3 and the study was continued. The removal efficiency profiles were monitored continuously till they achieved steady state. Figures 4 and 5 show the concentration profile and the removal efficiency in RBC. The removal efficiency increased from 19% to 54% when the RBC was operated at high concentration. After decreasing the concentration, the removal efficiency increased rapidly and reached about 85% in 32 -35 days. The shorter time to achieve the steady state may be due to the thick growth of biofilm in the RBC. The total biomass in the RBC may be more than in the biofilter. The fact that such a thick biofilm can develop in RBC with DCM as the only carbon source proves the potential application of RBC for VOC removal.
4 CONCLUSIONS The RBC system was found to be a potentially attractive configuration for gas phase biodegradation of VOC. For comparative organic load, both RBC and biofilter gave reasonably high (>80 %) removal of DCM, but RBC was able to reach it in shorter time.
DICHLOROMETHANE REMOVAL USING MIXED CULTURES IN A BIOFILTER
475
Figure 4. Startup operation of the RBC treating DCM vapors.
Figure 5. Removal efficiency RBC treating DCM vapors.
REFERENCES Arulneyam, D. and Swaminathan, T. (2004) Biodegradation of mixture of VOC’s in a biofilter. J. Environ. Sci. 16(1): 30-33. Brunner, W., Staub, D. and Leisinger, T. (1980) Bacterial degradation of dichloromethane. Appl. Environ. Microbiol. 40: 950-958.
476
R. RAVI, LIGY PHILIP AND T. SWAMINATHAN
Diks, R., Ottengraf, S.P.P. and Vanden Oever, A.H.C. (1994) The inflence of NaCl on the degradation rate of dichloromethane by Hyphomicrobium sp. Biodegradation 5: 129-141. Green, T. (1991) Human exposure. Toxicol. 16: 2981-2985. Herbst, B. and Wiseman, U. (1996) Kinetics and reaction engineering aspects of the biodegradaton of dichloromethane and dichloroethane. Wat. Res. 30(5): 1069-1076. Hodge, D.S. and Devinny, J.S. (1994) Biofilter treatment of ethanol vapours. Environ. Prog. 13: 167-173. Wang, J. and Chen, J.M. (2006) A Removal of dichloromethane from waste gases with a biocontact oxidation reactor. Chemical Eng. J. 123(3): 103-107. Krausova, V.I., Robb, F.T. and Gonzalez, J.M. (2003) Bacterial degradation of dichloromethane in cultures and natural enviroments. J. Microbiol. Methods 54: 419-422. Kennes, C. and Veiga, M.C. (2001) Conventional Biofilters. In: Bioreactors for Waste Gas Treatment, Kennes, C. and M. C. Veiga (Eds.) Kluwer Academic Publisher, Dordrecht, pp. 47-98. Machay, D.M. and Cherry, J.A. (1989) Ground water contamination pump-and-treat remediation. Environ. Sci. Technol. 23: 630-636. Rittmann, B.E. and McCarty, P.M. (1980) Utilization of dichloromethane by suspended and fixedfilm bacteria. Appl. Environ. Microbiol. 39(6): 1225-1226. Rudiger, P. (1999) Abluftreinigung in Biofilmreaktoren mit inerten Tragern. Diss. ETH No: 13229, ETH Zurich. Vinage, I. and Rudolf Von Rohr, P. (2003) Biological waste treatment with a modified rotating biological contactor. I. Control of biofilm growth and long-term performance. Bioprocess Biosyst. Engin. 26: 49-57. Zhu, L., Abumaizar, R.J. and Kocher, W.M. (1998) Biofiltration of benzene contaminated air streams using compost activated carbon filter media. Environ. Prog. 17(3): 168-172.
477
Behaviour and optimization of a novel monolith bioreactor for waste gas treatment YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES Chemical Engineering Laboratory, Faculty of Sciences, University of La Coruña, Rúa Alejandro de la Sota, 1, 15008 – La Coruña, Spain
ABSTRACT Treatment of waste gases in bioreactors is cost-effective and environmental-friendly compared to the conventional techniques used for treating large flow rates of gas streams with low concentrations of pollutants. Nowadays, significant research is dedicated at the development of new bioreactor configurations, improved biocatalysts or new packing materials, among others. In the present study, a novel bioreactor packed with ceramic monolith was developed for treating VOCs (toluene or methanol) polluted air. Operational parameters that were considered included start-up of the bioreactor, inlet loading, changes in gas flow rate, liquid feed mode, and monolith blockage and biomass growth. Preliminary data on performance and stability have been obtained showing that this system can efficiently be used for waste gas treatment.
1 INTRODUCTION Biological treatment is an established technology for air pollution control and the alternative of choice to physical and chemical treatment techniques because of its environmental friendly and cost-effective for treating waste gases characterized by high gas flow rates and low pollutant concentrations (Kennes and Thalasso, 1998; Kennes and Veiga, 2001). The most widely utilized bioreactors for air pollution control are biofilters and biotrickling filters. Biofilters are reactors in which a humid polluted air stream is passed through a porous packed bed on which a mixed culture of pollutantdegrading organisms is naturally immobilized. In biotrickling filters, a distinct free water phase containing various nutrients is trickled over a packed bed. Both biofilters and biotrickling filters have some limitations concerning performance although they
477
478
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
are currently largely used for efficiently treating air polluted with volatile organic compounds (VOCs), odorous compounds, and other air pollutants. Conventional biofilters are usually packed with natural carriers, such as compost, peat or soil. They decay over time, causing compaction, clogging, short circuiting and increased headloss across the bed. In addition, using biofiltration to control hydrophobic compounds is difficult because of mass transfer rate limitations. For biotrickling filters, packed with inert carrier materials (Kennes and Veiga, 2002), the circulating trickling liquid allows controlling the pH value and supplies the fixed biofilm with the essential inorganic nutrients. However, the trickling phase and the presence of the liquid film slow down the transfer of pollutants and oxygen from the gas to the liquid phase. Over the recent past decades, great efforts have been dedicated to the development of new bioreactor configurations in order to improve the mass transfer. Poppe and Schippert (1992) demonstrated the advantages of adding water-immiscible organic solvents to the liquid phase of bioscrubbers for the elimination of hydrophobic VOCs. By adding organic solvents with high boiling points in a range of 10-30% of the total volume, 100 to 1000 times larger amounts of hydrophobic target compounds were absorbed in the scrubber solution. In the bioreactor, the target compounds were transferred from the organic phase to the water phase driven by a concentration gradient between oil and water as biological degradation of the compounds occurred in the water phase. This new technique was demonstrated by treatment of a mixture of 13 volatile compounds in air by a two-stage scrubber. The optimization of such twoliquid-phase systems is nowadays widely being studied (Daugulis, 2001). Reij et al. (1997) used a microporous hydrophobic membrane as a support for biofilms that remove the poorly soluble propene from air. In the membrane bioreactor, the pollutant in the gas phase is transferred through a membrane to the biofilm, attached to the other side of the membrane, where nutrients and oxygen are provided (Kennes and Veiga, 2001). Vinage and von Rohr (2003) developed the rotating biological contactor for waste gas treatment. The polluted air passes through the headspace of the reactor, containing discs mounted on a rotating shaft that serve as support for a biofilm. The shaft in rotated (~ 2 rpm), and the discs are partially wetted in water containing nutrients and other additives (Kennes and Veiga, 2001). The movement of the discs favors mass transfer and the control of the fixed biomass. Kan and Deshusses (2003) developed a new vapor phase bioreactor named the foamed emulsion bioreactor (FEBR) that overcomes some of the limitations of biofilters and biotrickling filters. The FEBR consists of an emulsion of highly active pollutant-degrading microorganisms and a water-immiscible organic phase, which is made into a foam with the air being treated. The monolith, which is widely used as catalyst support for gas treatment, e.g., cleaning of automotive exhaust gases and industrial off gases, can be tailored to meet the needs of a relatively inexpensive, light weight, inert, bioreactor packing that
BEHAVIOUR AND OPTIMIZATION OF A NOVEL MONOLITH BIOREACTOR FOR WASTE
479
provides a high specific surface area (surface-to-volume) to greatly increase the mass transfer rate. Typical monoliths consist of many parallel channels separated by thin, porous ceramic walls, representing a collection of parallel microreactors. They are formed in several configurations, usually from cordierite (2MgO·2Al2O3·5SiO2) or similar silica-alumina compounds. The geometry of monolithic supports yields one major advantage over particulate packing materials since they offer very little resistance to flow. Reactors using monolithic catalyst supports may be an attractive alternative to conventional multi-phase reactors and have been used in bioconversion and fermentation processes. Nevertheless, hardly any study has been done on their application in environmental technology. In monolithic channels bubble-train or Taylor flow usually occurs. Gas bubbles and liquid slugs move with constant velocity through the monolith channels approaching plug flow behavior. The gas is separated from the catalyst by a very thin liquid film and during their travel through the channels the liquid slugs show internal recirculation. These two properties result in optimal mass transfer. The present work is related to the study and development of a novel monolith bioreactor for the treatment of waste gases containing volatile organic compounds, i.e. toluene and methanol. The feasibility of using monolith bioreactors for treating VOC-polluted air has been proved. The effect of operating conditions as the gas flow rate, liquid flow rate, and the inlet loading rate of the system have been studied.
2 MATERIALS AND METHODS 2.1 MICROBIAL CULTURES AND CULTURE MEDIUM Studies undertaken in our group on toluene-treating monolith bioreactors were performed with a microbial consortium from a conventional biofilter treating the same pollutant. Batch experiments and growth of the inocula were undertaken with an aqueous culture medium containing (per liter): 4.5 g KH2PO4, 0.5 g K2HPO4, 2.0 g NH4Cl and 0.1 g MgSO4·7H2O (Jin et al., 2005). The culture medium was autoclaved at 120 oC for 20 min before adding filter-sterilized solutions of vitamins and trace minerals. Experiments on the removal of methanol were conducted with a pure culture of Candida boidinii. The nutrient solution used both for batch assays and for bioreactor studies contained the following macronutrients (per liter): 2 g KH2PO4, 2 g K2HPO4, 0.4 g NH4Cl and 0.2 g MgCl2·6H2O, and 0.01 g FeSO4·7H2O. The culture medium was sterilized at 120 oC for 20 minutes.
480
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
2.2 MONOLITH SUPPORT The ceramic monolith packing has the following characteristics: geometry: square ducts; length: 150 mm; cross section: 100×100 mm; number of channels: 26×26; channel width: 3.0 mm; weight: 850 kg.m-3; geometric surface: 800 m2.m-3; voids fraction: 64%. Details of the composition and the preparation procedure of the monolith used in this work are proprietary information of Rauschert Verfahrenstechnik GmbH (Germany). 2.3 EXPERIMENTAL SETUP The schematic of the monolith bioreactor used in this study is shown in Figure 1 and has been described previously in detail (Jin et al., 2006). The reactors were usually maintained at room temperature. The polluted gas was fed to the bioreactors by mixing a large air stream flowing through a humidification chamber with a smaller air stream passing through a flask containing the pollutant, i.e. either toluene or methanol. The bioreactors were fed in a downflow mode.
Figure 1. Schematic of the laboratory scale monolith bioreactor.
2.4 ANALYTICAL METHODS Methanol or toluene concentrations were measured by means of a HP-6890 gas chromatograph (Agilent Technologies, Spain) equipped with a 30 m×0.53 mm HP-PLOT Q column and a flame ionization detector, operating in splitless mode. Oven temperature was 130 °C, while both the injector and detector temperature was
BEHAVIOUR AND OPTIMIZATION OF A NOVEL MONOLITH BIOREACTOR FOR WASTE
481
150 °C. Samples were injected using a 2.5 cm3 gas-tight Hamilton syringe. Under these conditions, the retention time of methanol and toluene were 3.5 min and 4.3 min, respectively (Prado et al., 2005). Similarly, CO2 concentrations were measured on another Hewlett-Packard 5890 series II GC equipped with a thermal conductivity detector (TCD). The CO2 concentrations were determined at an injection temperature of 90 °C, an oven temperature of 25 °C and using a TCD at 100 °C.
3 RESULTS AND DISCUSSION 3.1 STARTUP OF THE BIOREACTORS The microbial consortia used to inoculate the bioreactors were obtained from either a previous bioreactor or from batch enrichments. The nutrient solution containing the desired biomass was continuously recirculated over the packing material. Simultaneously, a visible biofilm developed on the surface of the square channels of the monolith. Afterwards, the monolith was transferred to the bioreactor and VOCspolluted air was fed continuously. The start-up period of the bioreactor treating toluene lasted around 24 days with removal efficiencies of 60-100% while slowly increasing the load from 0.395 to 29.5 g-toluene.m-3.h-1. These data suggest that the start-up phase is quite slow for that pollutant. However, a shorter start-up period of a few days was required for the bioreactor treating methanol. After this period, the inlet concentration was kept at 200 mg.m-3 with an EBRT of 30 s, reaching an elimination capacity of 30 g.m -3.h -1, while maintaining the removal efficiency above 95%. Acclimated biomass allowed to shorten the start-up phase, as also observed by others (Veiga and Kennes, 2001). 3.2 INFLUENCE OF THE GAS FLOW RATE The influence of the gas flow rate on the reactor´s performance was evaluated in the bioreactor treating toluene. In the range of gas flow rates of 18 to 110 l.h-1, the removal efficiency first remained constant, around 90%, while gradually increasing the gas flow rate. When the gas flow rate was further increased, the biofilm thickness decreased due to the shear force. The mass transfer limitation step was determined by the laminar film thickness between the gas and liquid phase. When the gas flow rate was increased, the turbulence of the gas increased, and the laminar film became thinner. Hence, the resistance decreased and mass transfer was enhanced. The results show that the highest elimination capacity was reached at the highest gas flow rate, although this led to a lower removal efficiency (Figure 2).
482
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
Figure 2. Effect of the gas flow rate on toluene removal.
3.3 INFLUENCE OF THE TRICKLING LIQUID PHASE The gas and liquid mixture travels through the channels of the monolith reactor. Depending on the flow rate of each phase and on the feed method, a number of different flow regimes can occur, such as dispersed bubble flow, bubble flow, aerated Taylor flow, Taylor flow, churn turbulent flow, slug flow, annular flow and mist flow. In the co-current downflow trickling operation, the gas and liquid phases travel in the same direction through the channels. In this operation mode the Taylor flow regime is preferred. In this regime the gas and liquid move through the channels as separate packages, ensuring plug flow behaviour. The gas bubbles are separated from the bio-catalytic wall, containing the attached biofilm, only by a thin liquid film. Gas adsorbed in this film can immediately be consumed by the bio-catalyst attached on the walls of the channels. Adsorbed gas that is not consumed at the film exchanges with the liquid plug. The recirculation pattern in the liquid plug facilitates a rapid exchange with the film. Because of these properties of Taylor flow in capillaries, a high gas-liquid mass transfer rate is obtained. In order to check the effect of the mode of feeding of the liquid phase, a toluenefed bioreactor was first operated in a trickling mode and later without trickling phase (no recirculation of the liquid medium) during the treatment of toluene. The flow rate of the gas and liquid were 80 and 1 l.h-1, respectively. The removal of toluene in the trickling mode was lower than without trickling phase in the experimental range of inlet concentrations used in this work (Figure 3). This seems to contradict the theory that Taylor flow could enhance the mass transfer from the gas phase to the liquid
BEHAVIOUR AND OPTIMIZATION OF A NOVEL MONOLITH BIOREACTOR FOR WASTE
483
phase, which may be due to the following reasons: (1) when no liquid is recirculated in the reactor, the gas flow is uniform, and the liquid inside the channels forms a very thin film, then the resistance between the gas phase and the liquid phase is low, (2) in the trickling mode, although the Taylor flow generated by the liquid flow could enhance the mass transfer, the liquid was not uniformly distributed in the monolith. This could cause non-homogenous mass transfer in the different channels.
Figure 3. Performance of the monolith bioreactor with or without trickling phase.
3.4 MONOLITH BLOCKAGE AND BIOMASS GROWTH The clogging of the monolith channels was first observed during the treatment of toluene. The pressure drop sharply increased from initially zero to 0.5 cm H2O. The biofilm growth made the gas flow and liquid flow regime become nonhomogenous, and the performance of the biofilter decreased dramatically. In order to remove excess biomass, a high flow rate of trickling liquid was used. The turbulence that was generated allowed to efficiently wash out part of the biofilm from the reactor. It seems that controlling the biofilm growth is a very important parameter for long term stable operation of monolith bioreactors, as also recently observed by others (Ebrahimi et al., 2006). This problem can be solved by optimizing the dimensions of the channels or by means of a high flow rate of the trickling liquid. Previous studies undertaken with smaller channels (channel width 1.27 mm instead of 3.0 mm) resulted in a still faster clogging. Studies on biomass accumulation were also performed with methanol. A low biomass growth rate acidophilic yeast, Candida boidinii, was inoculated for treatment
484
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
of methanol. The pressure drop across the bioreactor indicates that biomass accumulation was relatively insignificant until day 50 of operation. The pressure drop remained around 6 mm H2O/m after 35 days operation. On day 60 of this operational period, the inlet loading of methanol was increased from 75 to 150 g.m-3.h-1, this high loading of methanol enhanced excess biomass growth causing clogging of the channels. Finally, the accumulated biomass led to a dramatic increase in pressure drop across the bioreactor on day 75. Biomass accumulation has also been observed on the top view of the monolith packing. In order to remove excess biomass, a high liquid flow rate was used (3 l.h-1) to generate shear forces and remove some biomass attached on the channel walls allowing the pressure drop to return to its original value, around 6 mm H2O.m-1. It is also important to note that the biomass accumulation, as reflected in Figure 4, had very little effect on methanol removal even at high values of the pressure drop. Physical operational problems are, however, encountered at such high pressures, necessitating backwashing to remove excess biomass. Overall, the monolith bioreactor showed a higher elimination capacity and much lower pressure drops compared to other conventional bioreactors, which could save on operation costs when the bioreactor is scaled up for application in the field.
Figure 4. Development of pressure difference across the monolith bioreactor.
BEHAVIOUR AND OPTIMIZATION OF A NOVEL MONOLITH BIOREACTOR FOR WASTE
485
4 CONCLUSIONS According to the data available so far, it appears that monolith bioreactors are able to reach relatively high removal rates and very good performances compared to conventional systems. However, more research is still needed in order to confirm if such good results can be generalized to all monolith bioreactors and if they can be maintained over longer operation periods.
5 ACKNOWLEDGEMENTS The present research was funded by the Spanish Ministry of Education and Science (Project CTM2007-62700/TECNO) and through European FEDER funds. Yaomin Jin was financially supported by the Xunta de Galicia through Project PGIDIT05PCIC10304PN.
REFERENCES Daugulis, A.J. (2001) Two-phase partitioning bioreactors: a new technology platform for destroying xenobiotics. Trends Biotechnol. 19: 457-462. Ebrahimi, S., Kleerebezem, R., Kreutzer, M.T., Kapteijn, F., Moulijn, J.A., Heijnen, J.J. and van Loosdrecht, M.C.M. (2006) Potential application of monolith packed columns as bioreactors, control of biofilm formation. Biotechnol. Bioeng. 93: 238-245. Jin, Y., Veiga, M.C. and Kennes, C. (2005) Autotrophic deodorization of hydrogen sulfide in a biotrickling filter. J. Chem. Technol. Biotechnol. 80(9): 998-1004. Jin, Y., Veiga, M.C. and Kennes, C. (2006) Development of a novel monolith-bioreactor for the treatment of VOC-polluted air. Environ. Technol. 27(11): 1271-1277. Kan, E. and Deshusses, M.A. (2003) Development of a foamed emulsion bioreactor for air pollution control. Biotechnol. Bioeng. 84: 240-244. Kennes, C. and Thalasso, F. (1998) Waste gas biotreatment technology. J. Chem. Technol. Biotechnol. 72(4): 303-319. Kennes, C. and Veiga, M.C. (2001) Bioreactors for Waste Gas Treatment. Kluwer Academic Publishers, Dordrecht, The Netherlands. Kennes, C. and Veiga, M.C. (2002) Inert filter media for the biofiltration of waste gases – characteristics and biomass control. Rev. Environ. Sci. Bio/Technol. 1: 201-214. Poppe, W. and Schippert, E. Das KCH-biosolv-ver-fahren in kombination mit einembiowäscher herkömlicher art-eine verfahrensentwicklung zur abluftreinigung fur wasserlösliche und schwer wasserlösliche schadstoffe. In Biotechniques for Air Pollution Abatements and Odour
486
YAOMIN JIN, MARÍA C. VEIGA AND CHRISTIAN KENNES
Control Policies (Maastricht, The Netherlands, 1992), A. J. Dragt and J. van Ham, Eds., Elsevier Science Publishers B.V., pp. 71-76. Prado, O.J., Veiga, M.C. and Kennes, C. (2005) Treatment of gas-phase methanol in conventional biofilters packed with lava rock. Water Res. 39(11): 2385-2393. Reij, M.W., Hamann, E.K. and Hartmans, S. (1997) Biofiltration of air containing low concentrations of propene using a membrane bioreactor. Biotechnol. Prog., 13(4): 380-386. Veiga, M.C. and Kennes, C. (2001) Parameters affecting performance and modeling of biofilters treating alkylbenzene-polluted air. Appl. Microbiol. Biotechnol. 55: 254-258. Vinage, I. and von Rohr, R. (2003) Biological waste gas treatment with a modified rotating biological contactor. I. Control of biofilm growth and long-term performance. Bioprocess Biosyst. Eng. 26(1): 69-74.
487
Biodegradation of BTXS and substrate interactions in a Bioactive Foam Reactor JIHYEON SONG AND SHONG-GYU SHIN Dept. of Civil & Environmental Engineering, Sejong University, Kunja-Dong, Kwangjin-Gu, Seoul, Korea 143-747
ABSTRACT A bioactive foam reactor (BFR), using surfactant-driven bubbles and suspended microorganisms, has emerged as a potential alternative to packed-bed biofiltration systems for the treatment of volatile organic compounds (VOCs). The study presented herein was designed to investigate the effects of VOC mixtures (benzene, toluene, p-xylene, and styrene) on biodegradation efficiencies and substrate interactions in the BFR. Benzene, toluene p-xylene, and styrene were applied individually to the toluene-acclimated BFR at the same inlet concentration (0.78 g/m3), and then paired BTXS mixtures (BT, BX, BS, TX, TS, XS, and BTXS in the same ratio by volume) were applied but the total inlet concentration were maintained constant. The overall removal rates of each of the four VOCs were in the following order: toluene, styrene, benzene, and p-xylene in the inlet concentration range tested. However, styrene biodegradation was the highest in the presence of other VOC compounds. The removal efficiency for toluene as a single substrate was 82%, but toluene removal efficiencies dropped when the paired mixtures were applied. The removal efficiency for benzene also decreased in the presence of other TXS compounds. In contrast, the removal efficiency for p-xylene as a single substrate was only 21% in the BFR, but p-xylene removal efficiencies ranged 35 – 41% in the presence of other BTS compounds. As a result, the biodegradation of benzene and toluene was inhibited by the other carbon sources, whereas the biodegradation of styrene and p-xylene was enhanced by the others. Consequently, a careful attention needs to be given when BFR performance and biodegradation rates of mixed VOCs are utilized for system design and operational purposes.
487
488
JIHYEON SONG
AND
SHONG-GYU SHIN
1 INTRODUCTION Packed-bed bioreactors including biofilters and biotrickling filters have drawn increased interests for the treatment of volatile organic compounds (VOCs). Many studies have shown that these bioreactors can successfully treat a wide range of VOCs including benzene, toluene and xylenes (van Groenestijn and Hesselink, 1993; Devinny et al., 1999). However, operational problems such as excess biomass accumulation and biodegradation activity loss make these treatment methods less attractive (Song and Kinney, 2000), especially when subjected to high concentrations of VOC mixtures. In order to overcome those problems, several new approaches have been made using suspended microorganisms instead of the fixed biofilm (Kennes and Veiga, 2001). One of them is a bioactive foam reactor (BFR) that is operated with a surfactant bubble solution containing pollutant-degrading microorganisms (Phipps, 1998). In the BFR, a VOC-laden air stream is sparged into the surfactant solution and forms biologically active foams. The fine foams can provide a large surface area for the mass transfer of VOCs as well as enhance microbial activity without a significant accumulation of biomass over time. Kan and Deshusses (2003, 2005) demonstrated that a foamed bioreactor using an organic-phase emulsion and active microorganisms could achieve a high elimination capacity for toluene. Recently we have modified the defoamer of the BFR system from its patented prototype to make it simple, and an organic emulsion such as oleyl alcohol was not used in the liquid phase. Since BFR performance mainly relies on the mass transfer of VOCs and the subsequent microbial degradation, substrate interactions between the VOCs present in inlet streams are another important factor that must be considered to achieve successful BFR operation when it is subjected to different chemical mixtures. Emissions from various industrial sources often consist of a mixture of compounds with different chemical characteristics and biodegradability. Therefore, several bioreactor studies have examined the biodegradation of binary mixtures of VOCs, such as BTEX (Collins and Daugulis, 1999; Strauss et al., 2004), and methanol and α-pinene (Mohseni and Allen, 2000). These studies have consistently indicated that the presence of a readily degradable compound can inhibit the removal of the other recalcitrant compound. However, few attempts have been made to investigate the effects of substrate inhibition between complex VOC mixtures in bioreactors using suspended microbial cultures such as the BFR. The study presented herein was therefore designed to investigate the effects of a VOC mixture (benzene, toluene, p-xylene, and styrene) on biodegradation efficiencies and substrate interactions in the BFR.
BIODEGRADATION OF BTXS AND SUBSTRATE INTERACTIONS
489
2 MATERIALS AND METHODS 2.1 MICROBIAL CULTURE AND SURFACTANT In order to obtain the toluene-degrading culture used in this study, a mixed microbial culture was initially collected from a wastewater treatment plant in Seoul, Korea. Pseudomonas putida TDB4 was isolated from the mixed culture and cultivated in our laboratory using gaseous toluene as a sole carbon and energy source. The mineral medium used for cultivating the mixed culture and P. putida TDB4 was slightly modified from the composition described by Song and Kinney (2000), which contained 1.36 g/L KH2PO4, 1.42 g/L Na2HPO4, 3.03 g/L KNO3 and trace metals per liter of distilled water. The surfactant used in the BFR was TritonX-100 (Sigma-Aldrich, USA), which was selected based on a bottle test that showed no adverse effects on toluene degradation of P. putida TDB4. The surfactant concentration in the liquid phase was maintained at 0.013% (v/v) that was slightly lower than its critical micelle concentration (0.014%).
Figure 1. Schematic of the bioactive foam reactor (BFR) tested in this study.
2.2 BFR CONFIGURATION As shown in Figure 1, the lab-scale bioreactor used in this study consisted of a foam column (volume 1.8 L), where the mass transfer of VOCs between the gas and foams took place, and a cell reservoir (volume 2.5 L), where biologically active microorganisms degraded the VOCs in the liquid phase. An air stream, generated by a
490
JIHYEON SONG
AND
SHONG-GYU SHIN
compressor and controlled by a flow meter, was contaminated with VOC vapour by the slow injection of research-grade, pure compounds using a syringe pump (KD Scientific, USA). The VOC-contaminated air stream was introduced to the bottom of the foam column through an aeration stone. As the air stream was sparged into the nutrient solution containing the surfactant and the microbial culture, fine foams were generated and moved along with the air stream in the foam column. The foam itself broke down at the top of the cell reservoir, returned to the liquid phase, and then the air trapped in the moving foam was released and exited the reactor. The liquid phase was continuously recirculated from the cell reservoir to the foam column in a closedloop, and the total liquid volume in the entire system was 1.8 L. 2.3 BFR OPERATION AND SUBSTRATE INTERACTIONS Two sets of BFR experiments were conducted to determine the effect of various combinations and concentrations of the VOCs (BTXS) on removal efficiencies at a constant gas retention time of 40 seconds. Prior to each BFR start-up, a microbial solution consisted of 1 L of pre-grown P. putida TDB4 was mixed with 1 L of the nutrient medium containing the surfactant. And then, the BFR was acclimated to toluene biodegradation at an inlet toluene concentration of 0.38 g/m3 (i.e., 100 ppmv) until the steady state condition was obtained. First, in order to determine short-term responses of the bioreactor system at the inlet concentration of 200 ppmv, benzene, toluene, p-xylene, and styrene were applied individually to the toluene-acclimated BFR for five hours. And the short-term experiments were repeated for various VOC mixtures (BT, BX, BS, TX, TS, XS, and BTXS) at the same total concentration of 200 ppm v. Between each short-term experiment, the acclimation condition was restored to maintain the BFR at steady state. The second set of BFR experiments was conducted to observe the performance of the BFR operated continuously using BTXS that were mixed in a 1:1:1:1 ratio by volume. Throughout the 13-day operational period, the inlet VOC concentrations were changed twice. Initially the VOC mixture at the total concentration of 100 ppmv was introduced to the toluene-acclimated BFR (referred to as «Phase M1»). On day 6, the inlet concentrations of the VOC mixture were increased stepwise to 200 ppmv («Phase M2»). As a measure of substrate interaction between the compounds in the mixtures, a substrate interaction index (SII) was defined as a ratio of the changes in removal efficiency due to the other coexisting carbon source to the removal efficiency determined when a single carbon source was supplied. The SIIs were calculated for each compound in the presence of other compounds at the different concentrations tested in this study.
BIODEGRADATION OF BTXS AND SUBSTRATE INTERACTIONS
491
Substrate Interaction Index (SII) = (VOCmix – VOCsingle)/VOCsingle
(1)
Where VOCmix is the removal efficiency of the target compound in mixture, and VOCsingle is theremoval efficiency of the compound as a single substrate at the given concentration. 2.4 ANALYTICAL METHODS To determine VOC removals, gas samples were collected from three ports located at the inlet, at the middle of the connecting line between the foam and the microbial columns, and at the outlet of the BFR. The samples were collected with 0.5mL gas-tight syringes and immediately analyzed using a gas chromatograph (HP 6890, Agilent, USA) equipped with a flame ionization detector.
3 RESULTS AND DISCUSSION 3.1 SUBSTRATE INTERACTIONS IN THE SHORT-TERM RESPONSES The short-term changes in VOC removal efficiencies were determined when various combinations of different VOCs were applied to the toluene-acclimated BFR. The biodegradation of toluene in the absence of other compounds (BXS) by the toluenedegrading pure culture (P. putida TDB4) showed a high and stable removal efficiency. The removal efficiencies dropped when BXS compounds were applied individually to the BFR during the short-term period. The overall removal efficiencies of the four VOCs were in the following order: toluene, styrene, benzene, and p-xylene. Styrene is the most soluble compound among the VOCs tested, but its removal efficiency was lower than that of toluene in the toluene-acclimated microbial system. In addition, the biodegradation efficiency of pxylene was the lowest indicating that p-xylene was the most recalcitrant compound, and this finding was similar to other results reported in the biofiltration literature (Collins and Daugulis, 1999; Deshusses and Johnson, 2000; Strauss et al., 2004). The removal efficiencies of a target compound in the presence of the other compounds were monitored in the toluene-acclimated BFR. Figure 2 illustrated the SIIs calculated using the experimental data obtained in the short-term BFR operation. For instance, the removal efficiency of toluene as a single substrate was 82% at the inlet concentration of 200 ppmv, and it dropped to 55% when both toluene (100 ppmv) and styrene (100 ppmv) was applied. Therefore, the SII for toluene in the presence of styrene («Ts» in Figure 2) was -0.32, indicating that the toluene biodegradation was inhibited by the styrene addition. In comparison, the styrene removal efficiency was found to be 77% when styrene was individually supplied to the BFR, but it increased to 91% when toluene and styrene were applied together in the paired mixture, yielding
492
JIHYEON SONG
AND
SHONG-GYU SHIN
the SII of 0.18 for styrene in the presence of toluene («St» in Figure 2). These findings indicate that the styrene biodegradation was enhanced by the presence of toluene. Overall, as shown in Figure 2, the SIIs for both benzene and toluene were negative values, implying that the biodegradation of these compounds was always inhibited by the presence of the other compounds. Strauss et al. (2004) reported the results obtained from a packed-bed biofilter system that the toluene biodegradation was inhibited by the other compounds (benzene, p-xylene and ethylbenzene), and toluene had an enhancing effect on the removal efficiency of the other compounds when paired. However, in this study, the presence of toluene and benzene in the paired mixtures resulted in the mutual deterioration of the removal efficiencies of both compounds. The biodegradation of both p-xylene and styrene in the mixtures was enhanced by the presence of other compounds. The improved removal efficiencies of p-xylene and styrene occurred at the expense of toluene removal efficiency, presumably due to the similarities of the enzymatic systems used in the metabolic pathways for the aromatic compounds. Therefore, p-xylene, the most recalcitrant compound, had a greatest enhancing effect by the presence of other structurally-related aromatics. In addition, the presence of p-xylene and styrene in the paired mixture resulted in the mutual enhancement of the removal efficiencies of both compounds.
Figure 2. Substrate interaction indices for target compounds in the presence of the other compounds in paired mixtures.
BIODEGRADATION OF BTXS AND SUBSTRATE INTERACTIONS
493
3.2 SUBSTRATE INTERACTIONS IN THE CONTINUOUS BFR OPERATION The response of the BFR operated continuously using BTXS in the paired mixture of a 1:1:1:1 ratio by volume was monitored throughout the 13-day period. Figure 3 illustrates that the inlet concentration of each compound and its outlet concentration during the operational period. In the continuous experiment when subjected to the mixture of the four compounds, the removal efficiency was ranked in the following order: styrene, toluene, benzene, and p-xylene. The ranking of removal efficiencies for styrene differed from that of the short-term experiment. This finding implies that the toluene-acclimated microbial culture could become adapted to the biodegradation of styrene as the BFR operation continued, since styrene is a more soluble and readily biodegradable compound than the others. An averaged removal efficiency of each compound at pseudo-steady-state during Phase M2 was used to calculate SII values. The SII values were -0.01, -0.10, 2.02, 0.10 for Btxs, Tbxs, Xbts, Sbtx, respectively. Similar to the short-term BFR experiment, the SII values showed that the biodegradation of benzene and toluene was inhibited, but the biodegradation of p-xylene and styrene was enhanced by the presence of other compounds in the continuous BFR experiment.
Figure 3. Changes of inlet and outlet concentrations of (a) benzene, (b) toluene, (c) p-xylene, and (d) styrene in the VOC mixture during the continuous BFR operation.
494
JIHYEON SONG
AND
SHONG-GYU SHIN
It is interesting to note that all the SII values obtained in the continuous operation shifted to the positive direction from its value obtained from the previous short-term experiment (e.g., -0.20 -0.01 for Btxs). The enhancement of substrate interaction in the paired mixture was presumably due to an increase in microbial density in the cell reservoir as well as gradual adaptation of the microbial strain to the biodegradation of the BTXS mixture over time. Consequently, a careful attention needs to be given when BFR performance and biodegradation rates of mixed VOCs are utilized for system design and operational purposes.
4 ACKNOWLEDGEMENTS This work was supported by grant No. R01-2005-000-10675-0 from the Basic Research Program of the Korea Science and Engineering Foundation.
REFERENCES Collins, L.D. and Daugulis, A.J. (1999) Benzene/toluene/p-xylene degradation: solvent selection and toluene degradation in a two-phase partitioning bioreactor. Appl. Microbiol. Biotechnol. 52: 359-365. Deshusses, M.A. and Johnson, C.T. (2000) Development and validation of a simple protocol to rapidly determine the performance of biofilters for VOC treatment. Environ. Sci. Technol. 34: 461-467. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. Boca Raton: Lewis Publishers. Kan, E. and Deshusses, M.A. (2003) Development of foamed emulsion bioreactor for air pollution control. Biotechnol. Bioeng. 84: 204-244. Kan, E. and Deshusses, M.A. (2005) Continuous operation of foamed emulsion bioreactors treating toluene vapors. Biotechnol. Bioeng. 92: 364-371. Kennes, C. and Veiga, M.C. (2001) Bioreactors for waste gas treatment. Kluwer Academic Publishers. Mohseni, M. and Allen, D.G. (2000) Biofiltration of mixtures of hydrophilic and hydrophobic volatile organic compounds. Chem. Eng. Sci. 55: 1545-1558. Phipps, D.W. (1998) Biodegradation of volatile organic contaminants from air using biologically activated foam. US Patent No.5,714,379. Song, J. and Kinney, K.A. (2000) Effect of Vapor-Phase Bioreactor operation on biomass accumulation, distribution, and activity. Biotechnol. Bioeng. 68: 508-516. Strauss, J.M., Riedel, K.J. and du Plessis, C.A. (2004) Mesophillic and thermophilic BTEX substrate interactions for a toluene-acclimated biofilter. Appl. Microbiol. Biotechnol. 64: 855-861. van Groenestijn J.W. and Hesselink P.G.M. (1993) Biotechniques for air pollution control. Biodegradation. 4: 283-301.
495
Characterization and performance evaluation of a two-phase partitioning bioreactor for volatiles organic compounds treatment in off-gas JEAN-MARC ALDRIC AND PHILIPPE THONART Centre Wallon de Biologie Industrielle, Unité de Bio-industries, Faculté des Sciences Agronomiques de Gembloux, Passage des déportés, 2, B – 5030 Gembloux, Belgium
ABSTRACT The treatment of the industrials off-gas strongly evolves with change of the environmental legislation on a worldwide scale. Biotechnics existing for their treatment sometimes present limits for some volatile organic compounds (VOC) such as BTEX because of their poor water solubility. The use of two phase-partitioning bioreactors (TPPB) is an interesting alternative in this case. In this work, a laboratory scale TPPB (water / silicone oil) was monitored at high level of Isopropropylbenzene (IPB) air pollution (7g/Nm3) and a flow of 1 VVM. We focused ourselves on the inoculation with the strain Rhodococcus erythropolis T 902.1. We showed that the increase of the inoculums size to 5 g DM/l. induces a better initial abatement of the pollutant, however performances of the TPPB decrease quickly because of cellular mortality. The use of a smaller inoculum (0,2g DM /l) seems to be a good compromise to observe progressive improvement of the IPB abatement with adaptation of biomass. The TPPB was followed during 38 days in order to confirm its potentialities and characterise its evolution. We showed that the performances of the TPPB are maintained with an elimination rate near to 63 % for IPB polluted air (7 g/Nm3) and punctually reach 92 %. The biomass grows gradually and stabilizes itself around 10 g/l. With fluorescent double stain Rhodamine/ propidium iode, we also shown that cellular viability strongly evolve: cellular viability was low (30 %) in the first operating hours but is quickly increased after adaptation to IPB (80%); we also suggested an endorespiration phenomenon in the bioreactor. In this work, we could confirm the previously estimated elimination performances of the two- phase partitioning bioreactor with silicone-oil as second phase. Elimination rate of a monoaromatic compound at high concentration (7 g / Nm3) can be maintained between 240 g /m3.h and 360 g /m3.h. in the TPPB.
495
496
JEAN-MARC ALDRIC
AND
PHILIPPE THONART
1 INTRODUCTION Numerous polluting organic compounds are released by human activities and persist in the environment, because of their low solubility in water and their high concentration. Monoaromatic hydrocarbons such as benzene are produced in large amounts, and used in fuels, as solvents and as starting materials for the production of plastics, synthetic fibres and pesticides (Budavari et al., 1996). Monoaromatics have become prevalent environmental contaminants, and thirty of them are on the «EPA Priority Pollutant List» (1996). Eleven of these compounds are in the top 100 chemicals on the Priority «List of Hazardous Substances» published by the Agency for Toxic Substances and Disease Registry (ASTDR, 1997). In recent years, biological techniques have been applied more frequently to control these emissions, because they eliminate many of the drawbacks of classical physical-chemical techniques. Disadvantages of usual air treatment techniques are high-energy costs (incinerators), the use of expensive chemicals that may require special operational safety procedures (chemical scrubbers) and the generation of waste products such as spent chemical solutions or spent activated-carbon (Van Groenestijn et al., 2005; Davidson et al., 2003). Biological methods involving biofilter has been shown to be promising alternatives compared to the traditional technologies for the control of many gaseous pollutants (Rene et al., 2005). Bio filters present however several limits such as the ripening period of the bio filters during which cells proliferate to the point where the bed can be used and their restriction to the treatment of low VOC (volatile organic compounds) concentration (below 1g / Nm3). This is partly due to the poor water-solubility of gaseous pollutants. Some researches were carried out in order to improve transfer and solubility of hydrophobic pollutants during biological treatment. Budwill and Coleman (1997) showed the positive effect of silicone-oil addition on the biodegradation rate of n-hexane vapours in peat-based bio filters. Cesàrio et al. (1997) showed an enhancement of the toluene mass transfer rate by a factor of 1,1 using a dispersion containing 10% (v/v) FC40 solvent and a twofold oxygen transfer rate. More recently, many researches were devoted to the Two-Phase Partitioning Bioreactors (TPPB) as a new technology for xenobiotic degradation, The TPPB. concept has been demonstrated to be effective for the degradation oh high levels of organic compounds (Daugulis et al., 2001). Dumont and Delmas (2003) reviewed the mass transfer enhancement of gas absorption in oil-in-water systems and conclude that our understanding, of the influence of oil addition on the mass transfer parameters kL and a could be improved. By improving the oxygen and pollutant mass transfer and reduce the inhibitory substances by lowering their concentration in the aqueous phase, the elimination capacity reported for two-phase partitioning bioreactor often exceeded the performance
CHARACTERIZATION AND PERFORMANCE EVALUATION OF A TWO-PHASE
497
of bio filters. For nonexhaustive example, the elimination capacity quoted in some studies are summarized in Table 1. Table 1. Some examples of elimination capacity quoted in literature for some compounds Compounds
Microorganism
Hexane
Pseudomonas aeruginosa Sphingomonas aromaticivorans Alcaligenes xylosoxidans Y234 Alcaligenes xylosoxidans
HAP (naphtalène et phénanthrène) Benzene Toluene
Elimination capacity (EC) g/m³réact.h
Reference
135
Muñoz et al. (2006)
238
Daugulis and Janikowski (2002) Yeom and Daugulis (2001) Daugulis and Boudreau (2002)
291 727 727
To compare, a recent study (Arriaga et al., 2006) reported in a fungal biofilter supplied with silicone-oil quote EC of 100 g hexane /m3.h, however higher than those reported in classical fungal and bacterial biofilters (60g/m3react.h). The open nature of biofilters limits the control of parameters, inoculation with a single chosen species may thus fail (Devinny et al., 1999). In a TPPB, the best control of parameter permits the inoculation with an adaptated strain. In a similar way, the acclimatation in compost biofilters treating gasoline vapors, was much more rapid when they were inoculated with adaptated culture but do not affect ultimater removal efficiencies (Wright et al., 1997; Leson and Smith, 1997). In this study, we review the merit and limitation of a water / silicone oil TPPB used for treating isopropropylbenzene (IPB) gaseous vapours at high concentration. We also consider the influence of various quantities of inoculum on the initial evolution of EC.
2 MATERIALS AND METHODS 2.1 MICROORGANISMS AND CULTIVATION The Rhodococcus erythropolis strain was obtained from the collection of the Walloon Centre of Industrial Biology (C.W.B.I.; Belgium). All the substrates and other chemicals were purchased at VWR international (Leuven, Belgium) or Aldrich (Bornem, Belgium).
498
JEAN-MARC ALDRIC
AND
PHILIPPE THONART
The culture of Rhodococcus erythropolis in 868 medium (glucose 20g/l.; casein peptone 20 g/l; yeast extract 10 g/l) is harvested after 64 hours (optical density 600 nm =1,4). The inoculum for the biological reactor is obtained by centrifugation of various volume of this culture in function of experiment. The pellet obtained is washed twice and diluted in 200 ml saline water (9g/l NaCl). The inoculum is then introduced into the bioreactor where the medium for biodegradation is composed of silicone oil (10% V/V) and aqueous medium M284 (90 % V/V) whose composition is : Na2HPO4 17,7g/l; NaH2PO4 24,33 g/l (buffer pH 7); NaCl 4,68g/l ; KCl 1,49g/l; NH4Cl 1,07g/ l; Na2SO4 0,43g/l; MgCl2. 6H2O 0,20g/l ; Na2HPO4 2H2O 40mg/l ; CaCl2. 2H2O 30mg/ l Fe(III)NH4citrate 4,8mg/l; ZnSO4.7 H2O 0,144 mg/l ; MnCl2.4 H2O 0.1 mg/l ; H3BO3 0,062 mg/l ; CoCl2.6 H2O 0,19 mg/l; CuCl2.2 H2O 0,017 mg/l ; NiCl2.6 H2O 0,024 mg/l ; Na2MoO4.2 H2O 0,036 mg/l; ethanol 1g/l. 2.2 BIOREACTOR AND ASSEMBLY The stirred bioreactor used for biodegradation (LSL Bio Lafitte BL06.1, Saint German en Laye, France) described by Aldric (2005). Its reactional volume reaches 4,5l and the stirring speed was maintained at 600 rpm. The assembly is schematized in figure 1.
Figure 1. Schematic of the laboratory scale TPPB. A flow-meter (2) allows the control of the gas flow by means of a control valve (3). A bottle (5) containing pollutant (IPB) allows the generation of the polluted effluent with a temperature control (7). A valve is used to dilute the polluted effluent and to control the concentration at level of the sampling bubble preceding the bioreactor (9). A septum (10) is used to sample gas. The stirred bioreactor is the seat of the biodegradation (12). A sparger (13) is used to allow a diffusion of little bubbles of gas effluent within the bioreactor. Stirring and mixing are carried out by means of two agitation modules : a TD4 module (14) and a helicoidal module (15). Baffles (16) avoid the formation of vortex within the bioreactor. Lastly, a sampling bubble permits to measure the residual concentration at the exit of the bioreactor.
CHARACTERIZATION AND PERFORMANCE EVALUATION OF A TWO-PHASE
499
The IPB concentrated gas is generated by stripping within a thermostatized glas bottle. The gas flow is permanently controlled by a flow meter and fixed at 4,5 l/min. The concentration of IPB in the gas coming in the bioreactor is controlled by an adjustable mixture between polluted gas and air and maintained at 7 g / Nm3. 2.3 SAMPLING AND ANALYTICAL METHODS Gas samples are regularly taken from each bubble of sampling as well as in the liquid reactional medium. IPB concentration was estimated thanks to a Perkin Elmer headspace sampler HS 40 XL (for liquid samples) and a gaz chromatograph Hewlett Packard 5890 equiped with a Alltech INC. Deerfield EC-WAX column and flame ionisation detector (for gas-samples). Temperatures of the injector, column and detector were respectively 153, 150 and 250 °C. 2.4 CELLULAR VIABILITY Liquid samples are regularly taken from TPPB and diluted to approximately 106 UFC/ml. 1ml was then centrifuged (8000 rpm; 10 min), the pellet was twice washed with sterile saline water (9g/l NaCl). The pellet was double stained with 5μl of Rhoda mine 123 (1,25 mM) and 5μl of propidium iode (1,25mM), homogenized and incubated 5 minutes at ambient temperature. The cellular vialbility can be evaluated by a fluorescent microscopy, viable cells appear green and nonviable cells appear red. The proportion of viable cells were estimated by enumeration of red and green cells in 5 microscopic fields of vision / sample.
3 RESULTS AND DISCUSSION 3.1 INFLUENCE OF INOCULUM DENSITY ON THE REMOVAL EFFICIENCY In this experiment, the TPPB was inoculated with several inocula in order to evaluate impact of initial cellular biomass on the performances of bioreactor during the firsts four days. Indeed, the ratio pollutant concentration / adapted micro-organisms concentration could be an important factor when the TPPB is started. According to the data shown in Figure 2, it appeared that the use of a high size of inoculum (2,4 g DM / l) allows reaching initial removal efficiency near to 95 % but a reduction in RE is subsequently observed during the first 4 days. On the other and, for a lower size (respectively between 0,2 and 1,53 g DM/ l) the RE was maintained with a low value (around 80%) during the first 4 days. Singularly, smallest size of inoculum (0,2 g DM / l), allows to reach the highest RE after 4 operating days (84 %).
500
JEAN-MARC ALDRIC
AND
PHILIPPE THONART
Figure 2. Influence of size of inoculums on the performance of a TPPB removing IPB.
Figure 3. Evolution of biomass dry matter during the first 4 days of experimentation for various size of inoculum.
Figure 3 clearly indicates that the cellular multiplication is inversely proportional to the size of the inoculum. When the inoculum size is high (beyond 1,5 g/l), no beneficial effect can be observed, neither on the cell multiplication, nor on the removal efficiency. However, under the same conditions, a TPPB inoculated with an inoculum
CHARACTERIZATION AND PERFORMANCE EVALUATION OF A TWO-PHASE
501
size of 0,2 g/l reach 3,75 g/l and a RE maintained at 85 % after 10 operating days (results not shown). Small inoculum (0,2 g/l) seems to be the best compromise to quickly reach high RE and an adaptation of the biomass to high IPB concentration. 3.2 EVALUATION AND EVOLUTION OF THE PERFORMANCES OF THE TPPB The performances of the TPPB were evaluated while following the elimination flow of IPB (during 38 days). The TPPB was first continuously sparged during the first four days with an inlet gaz flow (7g/Nm3 IPB; 1 VVM) (phase 1). Then, the TPPB was sparged only during the day (phase 2) because of the too much low RE during the first phase. Lastly, the TPPB was again continuously sparged between day 30 and day 38 to evaluate the performances under extreme conditions (phase 3).
Figure 4. Evolution of the flows within the bioreactor Qin (rhombus) : IPB flow coming in TPPB; Qelim (triangle): IPB elimination flow (mg/min.l). Small dotted lines and large dotted lines represent respectively the average of Qin and Qelim during phase 2.
Figure 4 shows the performances evolution for a TPPB during each phase. It should be specified that le TPPB is used in the limiting predetermined conditions (Aldric et al., 2005), however, under these conditions, the removal efficiency is supported at approximately 63 % and punctualy reach 81% and 92 % with twelfth and thirteenth days, the elimination flow during phase 2 is thus estimated at 4mg/l.min (240g/m3react .h). On the other hand, when starting the third phase, the removal efficiency fall to 25 % (elimination flow below 2mg/l.min) at day 31 but follows a readadaptation at the end of experiment.
502
JEAN-MARC ALDRIC
AND
PHILIPPE THONART
Figure 5. Evolution of IPB concentration in water-silicone oil media.
IPB concentration in the two-phase media (Figure 5) increases gradually during phase 2 and stabilizes itself between 1000 and 1500 mg/l. The absorbtion of pollutant is indeed followed of a consequent biodegrdation since the accumulation of IPB in the TPPB is limited. Nevertheless, continuous sparging with IPB gas flow (night and day; phase 3) increases strongly the IPB accumulation. This leads to a reduction of driving force term (CL0– CL) resulting from well-known equation 1, this can explain the reduction of removal efficiency observed. eq. 1
3.3. EVOLUTION OF BIOMASS Growth of Rhodococcus erythropolis T 902.1 within the TPPB and viability of cells constitute significant parameters to evaluate the influence of the IPB load on the biomass. Figure 6 shows a high growth of the biomass during the first 19 days, from 0,3 g/l to 12,75 g/l. Thereafter, the biomass is stabilized between 9 and 11 g/l as from day 20 until day 38. Figure 6 show a very low viability of the cells when the TPPB is starded, only 15h after inoculation with washed biomass. Cellular viability increases then continuously to reach 84 % after 7 operating days and to be maintained between 68 % and 78 % during phase 2. When phase 3 is started (day 30), the biomass seems to undergo a shock probably induced by the increase of IPB concentration in two-
CHARACTERIZATION AND PERFORMANCE EVALUATION OF A TWO-PHASE
Figure 6. Evolution of biomass dry matter (DM)
503
Figure 6. Evolution of cellular viablity
phase media, however the biomass seems to subsequently readapt with the news operating conditions. The stabilization of the biomass as well from the quantitative point of view (dry matter) as qualitative (viability) suggest a cellular metabolism more directed towards the biodegradation and the endorespiration phenomenon that towards the cell multiplication.
4 CONCLUSIONS The following conclusions can be drawn from the results presented in this study (1) The use of a small inoculums size (0,2 g DM/l) is preferable with the use of more significant sizes because of a better adaptation of the biomass to IPB load. (2) The use of a TPPB water-silicone oil to treat hight concentration IPB flows is confirmed. For an IPB load near to 390 g.m -3.h-1, an IPB average elimination flow of 240 g.m-3.h-1 can be obtained and maintained during 31 days of noncontinuous operating conditions, in addition the biomass remains functional. (3) The limits of the TPPB seem to be reached when the TPPB is continuously fed with strongs IPB flows and concentration, although the biomass seems to relatively readapt after the shock.
5 ACKNOWLEDGEMENTS The authors wish to acknowledge Dow-Corning Society and Irs. Tiangoua Kone and Didier Mundella for their participation in this work.
504
JEAN-MARC ALDRIC
AND
PHILIPPE THONART
REFERENCES Aldric, J.M., Destain, J. and Thonart, P. (2005) The two-phase bioreactor water /silicon-oil: prospects in the off-gas treatment. Appl. Biochem. Biotechnol. 121-124: 707-720. Arriaga, S., Muñoz, R., Hernandez, S., Guieysse, B. and Revah, S. (2006) Gaseous hexane biodegradation by Fusarium solani in two liquid phase packed-bed and stirred tank bioreactors. Environ. Sci. Technol. 40: 2390-2395. ASTDR, Priority List of Hazardous Substances, Agency of Toxic Substances and Disease Registry, USA, 1997. Budavari, S., O’Neil, M.J., Smith, A. and Heckelman, P.E. (1996). http://www.amazon.com/exec/ obidos/search-handle-url/index=books&field-author-exact=Merck&rank=relevance%2C%2Bavailability%2C-daterank/102-6297388-6039339. The Merck Index: An Encyclopaedia of Chemicals, Drugs and Biological, Merck, Whitehouse Station, NJ. Budwill, K. and Coleman, R.N. (1997) Effect of silicone-oil on bio filtration of n-hexane vapours. Med. Fac. Univ. Gent. 62: 1521-1528. Cesário, M .T., Beverloo, W.A., Tramper, J. and Beeftink, H.H. (1997) Enhancement of gas-liquid mass transfer rate of polar pollutants in the biological waste gas treatment by a dispersed organic solvent. Enz. Microb. Technol. 21: 578-588. Daugulis, A.J. and Boudreau, N.G. (2003) Removal and destruction of high concentration of gaseous toluene in a two-phase partitioning bioreactor by Alcaligenes xylosoxidans. Biotechnol. Lett. 25: 1421-1424. Davidson, C.T. and Daugulis, A.J. (2003) The treatment of gaseous benzene by two-phase partitioning bioreactors: a high performance alternative to the use of bio filters. Appl. Microbiol. Biotechnol. 62: 297-301. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. Boca Raton: Lewis Publishers. 299 p. Dumont, E. and Delmas, H. (2003) Mass transfer enhancement of gas absorption in oil-in-water systems: a review. Chemical Engineering and Processing. 42: 419-438. EPA, Priority Pollutants, Code of Federal Regulations, (1996). Title 40, Part 423, Appendix A, USA, Chapter 1. Leson G.; Smith B.J., Petroleum environmental research forum field study on bio filters for control of volatile hydrocarbons. J. Environ. Eng. 123: 556. Rene E.R., Murthy D.V.S. and Swaminathan T. (2005) Performance evaluation of a compost biofilter treating toluene vapours. Proc. Biochem. 40: 2771-2779. Van Groenestijn, J.W. and Kraakman, N.J.R. (2005) Recent developments in biological waste gas purification in Europe. Chem. Engin. J. 113: 85-91. Wright, W.F., Schroeder, E.D., Chang, D.P.Y. and Romstad, K. (1997). Performance of a pilotscale compost infielder treating gasoline vapour. J. Environ. Eng.123: 547.
505
Removal of dichloromethane from waste gases using a fixed-bed biotrickling filter and a continuous stirred tank bioreactor LAURA BAILÓN1, YOLANDA DOPICO1, MARCELL NIKOLAUSZ2, MATTHIAS KÄSTNER2, MARÍA C. VEIGA1 AND CHRISTIAN KENNES1 1
Chemical Engineering Laboratory, Faculty of Sciences, University of La Coruña, Rúa Alejandro de la Sota, 1, 15008 – La Coruña, Spain 2 Department of Bioremediation, UFZ-Centre for Environmental Research Leipzig-Halle GmbH, Permoserstr. 15, 04318 – Leipzig, Germany
ABSTRACT A laboratory scale fixed bed biotrickling filter (BTF) and a continuous stirred tank bioreactor (CSTB) have been studied and compared for the elimination of dichloromethane from waste gases. The DCM removal efficiency in the trickling filter was > 85% for inlet loads up to 25 g.m-3.h-1 and a maximal removal capacity of about 170 g.m-3.h-1 was achieved at a load of 350 g.m-3.h-1. The continuous stirred tank bioreactor showed removal efficiencies > 90% for inlet loads up to 120 g.m-3.h-1. At this load the maximal removal capacity of the system was reached, i.e. about 100 g.m-3.h-1. Thus, higher maximum elimination capacities were reached in the BTF while higher removal efficiencies were obtained at high loads with the CSTB. Both systems presented good stability against overloads.
1 INTRODUCTION Dichloromethane (DCM) is produced in large amounts by the chemical industry. It is widely used as solvent in paint removers, acetate film production, pharmaceutical processes, metal degreasing and as an aerosol propellant. Due to its low boiling point (40.1 ºC) and high vapour pressure (47kPa at 20 ºC) significant amounts of DCM reach the environment via gaseous emissions. The global consumption of DCM in 2004 was about 600000 tonnes and it is still growing further (Wang and Chen, 2006).
505
506
L. BAILÓN, Y. DOPICO, M. NIKOLAUSZ, M. KÄSTNER, M. C. VEIGA AND C. KENNES
Although some DCM containing wastes are incinerated, it can be assumed that the greater part of the DCM produced is eventually lost into the environment. Besides being present in gaseous emissions, DCM is also detected in many aqueous industrial effluents (Hartmans and Tramper, 1991). As most halogenated compounds, DCM is considered a health hazard. It is an irritating compound with fragrant odour, which is harmful to the respiratory system and central nervous system. Besides, it is a potential human carcinogen. Exposure to high concentrations may cause unconsciousness and death. Its MAK-value (maximal concentration at work) is 360 mg.m-3. The emission control of DCM and other VOCs has been the subject of recent environmental regulations in several countries. Therefore, a considerable interest exists in the development of techniques for the elimination of DCM and other chemicals from waste gases. The conventional control technologies for VOCs treatment, such as thermal incineration and wet scrubbing are usually costly, mainly when the pollutant concentration is low and the air flow rate is large. Biological techniques, using biofilters, bioscrubbers, biotrickling filters, suspended-growth bioreactors, or membrane bioreactors among others, have been studied and applied successfully over the past decades to solve problems of polluted air emissions containing VOCs and odours (Kennes and Veiga, 2001). These biotechnological methods often exhibit similar or even higher efficiencies than the traditional physical-chemical processes. They avoid the need of expensive catalysts, do generally not generate secondary streams that have to be treated again, and present lower operating costs (Gadre, 1989; Groenestijn van and Kraakman, 2005; Kennes and Thalasso, 1998). DCM can be readily degraded under aerobic and anaerobic conditions by several different microbial genera using it as their sole carbon and energy source. The aerobic degradation of DCM yields 1 mole of carbon dioxide and 2 moles of hydrogen chloride, with medium acidification, according to the following reaction (Herbst and Wiesmann, 1996; Kennes et al., 2006): (1) DCM degrading bacteria are isolated quite readily from activated sludge and many water and soil samples. Strains belonging, among others, to the genera Hyphomicrobium, Pseudomonas and Methylobacterium have been described (Brunner et al., 1980; Stucki et al., 1981; Gälli and Leisinger, 1985). Free liquid phase bioreactors have proven to be superior in the removal of acid producing pollutants than biofilters, as pH control and removal of metabolites are easier. This paper presents the results of an investigation of DCM removal from air in a Biological Trickling Filter (BTF) and in a Continuous Stirred Tank Bioreactor (CSTB). Studies have very recently also been started to compare the results obtained
REMOVAL OF DICHLOROMETHANE FROM WASTE GASES USING A FIXED
507
with the BTF and CSTB of this work to those that would be reached with the same kind of reactors and conditions but working with a two-phase system, above all to check the effect of shock loads. The presence of a free liquid phase in the BTF and CSTB is interesting because of the non-negligible water solubility of DCM (Kennes et al., 2006). Besides, the addition of an organic phase makes sense when dealing with fluctuating pollutant concentrations and shock load-conditions in order to buffer load variations.
2 MATERIALS AND METHODS 2.1 BACTERIA AND MEDIUM The BTF and the CSTB were inoculated with a biomass suspension of Hyphomicrobium KDM2 and KDM4 cultivated in shake-flasks. These dichloromethane-degrading strains have been described previously (Nikolausz et al., 2006) and belong to the Department of Bioremediation of the UFZ-Centre for Environmental Research Leipzig-Halle GmbH in Germany. The mineral medium used in this studuy contained per litre of distilled water: 1.5 g KH2PO4, 4.69 g Na2HPO4.12H2O, 0.5 g (NH4)2SO4, 0.2 g MgSO4.7H2O, 1 ml of a trace mineral solution and 1 ml of a vitamins solution. The trace mineral solution contained per litre of distilled water: 5.3 mg CaCl2, 2 mg FeSO4.7H2O, 0.2 mg MnSO4.5H2O, 0.2 mg CuSO4.5H2O, 0.2 mg ZnSO4.7H2O, 0.03 mg H3BO3, 0.4 mg CoCl2 and 4 mg Na2MoO4.2H2O. The vitamins solution contained per litre of distilled water: 0.2 mg biotin, 0.2 mg folic acid, 0.5 mg riboflavin, 0.5 mg thiamine, 0.5 mg nicotinic acid, 0.001 mg vitamin B12, 0.5 mg p-aminobenzoic acid, 1 mg pyridoxamine and 0.5 mg lipoic acid. The pH of the medium was adjusted at 7. 2.2 FIXED BED BIOTRICKLING FILTER A scheme of the biotrickling filter is shown in Figure 1. The reactor consisted of a glass column of 0.09 m internal diameter, with a cone at the bottom and at the top. A total working volume of 2.1 litre was filled with lava rock and the liquid medium was distributed or collected by means of perforated plates at the top and bottom of the column. The waste gas was introduced at the top of the filter in co-current operation. The polluted gas stream was artificially created by mixing two air streams. A small stream of air was bubbled through a vaporization flask containing pure dichloromethane and was mixed afterwards in a mixing chamber with a large pure air stream. Gas phase concentrations ranging from 0.1 to 15.7 g.m-3 were obtained by changing the flow rate of the dichloromethane stream. The total gas flow rate was kept constant at 0.084 m3.h-1. The gas velocity was 13.2 m.h-1 and the empty bed retention time 90 s.
508
L. BAILÓN, Y. DOPICO, M. NIKOLAUSZ, M. KÄSTNER, M. C. VEIGA AND C. KENNES
Figure 1. Scheme of the laboratory biotrickling filter system.
The gas flow rates for air and DCM were adjusted utilizing two rotameters with valve (Brook Sho-Räte and Aalborg) and the liquid velocity was kept at 6.9 m.h-1 through a Watson Marlow peristaltic pump. The reactor was maintained at room temperature (21 ± 2 ºC). The liquid in the holding tank was gently mixed with a magnetic stirrer. The pH of the culture was kept between 6.9 and 7.05, with the aid of a pH controller (D09765T, Labprocess) coupled to an electrovalve, by dosing a 2 N NaOH alkaline solution to neutralize the HCl formed during the biological DCM degradation. This resulted in NaCl accumulation. High concentrations of this compound are expected to inhibit the biological activity (Gälli and Leisinger, 1985; Ottengraf et al., 1986; Hartmans and Tramper, 1991; Okkerse et al., 1999a, Diks et al., 1994). Thus, the conductivity was continuously measured in the holding tank (D09765T, Labprocess) and was kept below 28 mS.cm-2 by intermittent draining of liquid from the reactor. To maintain the required level of inorganic nutrients and compensate for drain and evaporation, fresh medium was added when needed. The limit of 28 mS.cm-2 was obtained from a salt tolerance test described in the Results. Air samples were taken with a Hamilton gas tight syringe at the inlet and outlet ports of the bioreactors. The DCM concentration of the samples was determined using a gas chromatograph, Hewlett Packard HP 6890 GC, equipped with a flame ionisation detector. 2.3 CONTINUOUS STIRRED TANK BIOREACTOR (CSTB) The experimental set-up is shown in Figure 2. The bioreactor consisted of an air-tight cylindrical glass vessel with a total volume of 2 litre (BioFlo 110, New Brunswick Scientific). The BioFlo-reactor was equipped with a PCU, a level control
REMOVAL OF DICHLOROMETHANE FROM WASTE GASES USING A FIXED
509
module, a dO2/pH control module, a power controller (with rotameter), a 4 peristaltic pumps unit and an injection and extraction ports for culture medium. The temperature was controlled by a water jacket with thermostated water. Mixing was achieved by a turbine stirrer.
Figure 2. Scheme of the continuous stirred tank bioreactor.
The vessel was filled with 1.5 litre of biomass suspension. The temperature was kept at 30 ºC and the stirrer at 400 rpm. As in the BTF, the pH of the culture was kept at 7 utilizing 2 N NaOH as a neutralizing agent, added automatically by means of a peristaltic pump. The conductivity was here also kept below 28 mS.cm-2. Fresh medium and biomass culture were added and removed continuously in order to maintain a constant salt concentration. The inlet and outlet liquid flow rates were controlled to keep an equilibrium between conductivity and biomass concentration; being the total reactor volume constant. The removed biomass solution was decanted and the supernatant or the settled biomass were recycled to the reactor when needed. The waste gas stream was created in the same way as for the BTF. The flow rate was also 0.084 m3.h-1 while the gas empty bed retention time was 64 s. The DCM concentration was determined at the inlet and outlet gas streams of the bioreactor in the same way as in the trickling biofilter.
510
L. BAILÓN, Y. DOPICO, M. NIKOLAUSZ, M. KÄSTNER, M. C. VEIGA AND C. KENNES
2.4 SALT TOLERANCE TEST In this test the degradation of DCM was evaluated in batch cultures at NaCl concentrations ranging from 0 to 500 mM. The experiment was carried out in 250 ml Erlenmeyer flasks containing 30 ml mineral medium inoculated with 10 ml of microbial suspension from the CSTB, previously centrifuged and diluted with fresh medium to minimize the amount NaCl that it could contain. After the addition of 10 ml pure DCM, the flasks were sealed with a screw cap containing two septa. The one in contact with the content of the flask was made of Viton, an inert material towards DCM, while the second septum, of ordinary rubber, was put on top of it in order to guarantee sealing. Subsequently, the flasks were vigorously shaken by hand until the compound had dissolved. To keep the liquid and gas phases in equilibrium during the rest of the experiment the Erlenmeyers were shaken horizontally at 120 rpm and 30ºC. Under these equilibrium conditions the degradation of DCM was followed by determining the gas phase concentration by gas chromatography as a function of time.
3 RESULTS AND DISCUSSION 3.1 BIOREACTORS PERFORMANCE The performance curves for both bioreactors are plotted in Figure 3.
Figure 3. Elimination capacity of both bioreactors versus the DCM inlet load.
The biotrickling filter had a maximal elimination capacity (EC) of around 170 g.m-3.h-1 (Figure 3). Up to inlet loads of 25 g.m-3.h-1 the BTF presented removal efficiencies (RE) between 85-100%. Between 25 and 170 g.m-3.h-1 the efficiency decreased sharply
511
REMOVAL OF DICHLOROMETHANE FROM WASTE GASES USING A FIXED
from 85 to 50%; and then more smoothly down to 45% at 350 g.m-3.h-1, which is the pseudo-critical load of the performance curve, where the maximal elimination capacity was reached. At higher loads of 600 g.m-3.h-1 the RE was only 28%. These results of RE and EC are in the same range as those obtained by other authors (Table 1) (Kennes et al., 2006). The continuous stirrer tank bioreactor shows a maximal elimination capacity of around 100 g.m-3.h-1 (Figure 3). High removal efficiencies between 90-100% are found for inlet loads up to 120 g.m-3.h-1. At this point the maximal elimination capacity was reached and the RE quickly decreased at higher loads. At 350 g.m-3.h-1 the RE was about 25%. Comparing the two different bioreactor configurations studied here, it can be concluded that for low inlet DCM loads the CSTB exhibits a better performance, while that for higher loads the BTF works better. The changing point happens at the cross point of the performance curves of both reactors, once the maximum removal efficiency of the CSTB is reached (Figure 3). To the best of our knowledge, no previous study has been published on the removal of DCM in a CSTB and on its comparison with its removal in a DCM-treating BTF. Table 1. Air biotrickling filters for the removal of DCM working at neutral pH and T ~20-30ºC. Bacteria
Packing material
Inlet range (g.m-3)
Gas / liquid velocity (m.h-1)
EBRT (s)
RE (%)
EC (g.m-3 h-1)
Hypomicrobium Sp. GJ21 Hypomicrobium DM20 Hypomicrobium Sp. GJ21 Hyphomicrobium KDM2 & KDM4
Keramic Novalox Saddles Polypropylene
References
0.5-10
160 / 36
29
32
150 (max)
0.0660.727
233 / 15.15
16
87.2 59.3
12.68 103.51
PVC
2
163 / 7.3
60
84
102
Okkerse et al., 1999a
Lava rock
0.1-15.7
13.2 / 6.9
90
90 45
13 170 (max)
This study
Diks and Ottengraf, 1991a, b Hartmans and Tramper, 1991
512
L. BAILÓN, Y. DOPICO, M. NIKOLAUSZ, M. KÄSTNER, M. C. VEIGA AND C. KENNES
After somewhat more than one year operation samples were taken from both bioreactors for identification of the dominant microorganisms by molecular techniques. It was observed that the originally inoculated Hyphomicrobium strains remained dominant in both bioreactors together with some other new species, despite working under completely non-sterile conditions as requested for practical environmental applications. The identification process was carried out at the UFZ-Centre for Environmental Research in Leipzig (non published data). To check the stability of the bioreactors both systems were subjected to 6 hours overloads. In a first experiment, the load was increased from approximately 70 g.m-3.h-1 to 250 and, later, 460 g.m-3.h-1 and, in a second experiment, it was increased from 15 g.m-3.h-1 to about the same maximum values. In all cases the recovery of both systems was almost immediate. Total recovery when the original inlet load was 70 g.m-3.h-1 took less than 1 hour; and less than 3 hours were required when the original load was 15 g.m-3.h-1. 3.2 SALT TOLERANCE TEST As DCM degradation is accompanied by HCl formation, which results in NaCl accumulation when neutralized with NaOH, the salt tolerance of our culture was tested. In Figure 4 it can be observed that at a concentration of 300 mM NaCl some limited inhibition of the bacterial activity is noticeable. At a concentration of 500 mM NaCl the inhibition becomes much more significant. Therefore, to avoid biological inhibition, the concentration of NaCl in the liquid medium of the bioreactors must be kept below 300 mM, which corresponds to a conductivity of 28 mS.cm-2 or less.
Figure 4. Degradation of DCM in batch cultures at different NaCl concentrations.
REMOVAL OF DICHLOROMETHANE FROM WASTE GASES USING A FIXED
513
Ottengraf et al. (1986) found complete inhibition at concentration levels exceeding 250-300 mM NaCl, while Hartmans and Tramper (1991) reported high inhibitory effects with 200 mM NaCl. In other bioreactor studies, the NaCl concentration was kept below 100-150 mM (Diks and Ottengraf, 1991a; Okkerse et al., 1999a). In a study of Diks et al. (1994), microbial growth of Hyphomicrobium GJ21 was shown to be strongly inhibited at NaCl concentrations exceeding 350 mM. Nevertheless, for a «TF-enrichment culture» from a laboratory trickling filter degrading DCM, initially inoculated with Hyphomicrobium GJ21, inhibition was less severe as good growth was observed up to 600 mM NaCl. So, it seems that some adaptation towards increased salt concentrations could take place in continuously operating trickling biofilters.
4 ACKNOWLEDGEMENTS The present research was partly financed by the Spanish Ministry of Science and Education (project CTM2007-62700/TECNO) and European FEDER funds, as well as the Xunta de Galicia (project PGIDIT05PCIC10304PN). We acknowledge the collaboration of Hana Simova during part of the experimental study.
REFERENCES Brunner, W.B., Staub, D. and Leisinger, T. (1980) Bacterial degradation of dichloromethane. Appl. Environ. Microbiol. 40: 950-958. Diks, R.M.M. and Ottengraf, S.P.P. (1991a) Verification studies of a simplified model for the removal of dichloromethane from waste gases using a biological trickling filter (Part I). Bioproc. Eng. 6: 93-99. Diks, R.M.M. and Ottengraf, S.P.P. (1991b) Verification studies of a simplified model for the removal of dichloromethane from waste gases using a biological trickling filter (Part II). Bioproc. Eng. 6: 131-140. Diks, R.M.M., Ottengraf, S.P.P. and Van den Oever, A.H.C. (1994) The influence of NaCl on the degradation rate of dichloromethane by Hyphomicrobium sp. Biodegradation 5: 129-141. Gadre, R.V. (1989) Removal of hydrogen sulfide from biogas by chemoautotrophic fixed-film bioreactor. Biotechnol. Bioeng. 34: 410-414. Gälli, R. and Leisinger, T. (1985) Specialized strains for the removal of dichloromethane from industrial waste. Cons. Recycling 8: 91-100. Groenestijn, J.W. van and Kraakman, N.J.R. (2005) Recent developments in biological waste gas purification in Europe. Chem. Eng. J. 113: 85-91.
514
L. BAILÓN, Y. DOPICO, M. NIKOLAUSZ, M. KÄSTNER, M. C. VEIGA AND C. KENNES
Hartmans S. and Tramper J. (1991) Dichloromethane removal from waste gases with a trickle-bed bioreactor. Bioproc. Eng. 6: 83-92. Herbst, B. and Wiesmann, U. (1996) Kinetics and reaction engineering aspects of the biodegradation of dichloromethane and dichloroethane. Water Res. 30: 1069-1076. Kennes, C. and Thalasso, F. (1998) Waste gas biotreatment technology. J. Chem. Technol. Biotechnol 72: 303-319. Kennes, C. and Veiga, M.C. (2001) Bioreactors for Waste Gas Treatment, Kluwer Academic Publishers, Dordrecht, The Netherlands. 312 pp. Kennes, C., Jin, Y. and Veiga, M.C. (2006) Fungal and dechlorinating biocatalysts in waste gas treatment. In: (Lens, P., Kennes, C., LeCloirec, P. and Deshusses, M.A., Eds), Waste Gas Treatment for Resource Recovery, IWA Publishing Co., London, UK, p. 277-301. Nikolausz, M., Nijenhuis, I., Ziller, K., Richnow, H. and Kästner, M. (2006) Stable carbon isotope fractionation during degradation of dichloromethane by methylotrophic bacteria. Environ. Microbiol. 8: 156-164. Okkerse, W.J., Ottengraf, S.P.P., Osinga-Kuipers, B. and Okkerse, M. (1999) Biomass accumulation and clogging in biotrickling filters for waste gas treatment. Evaluation of a dynamic model using dichloromethane as a model pollutant. Biotechnol. Bioeng. 63: 418-30. Ottengraf, S.P.P., Meesters, J.J.P., Van den Oever, A.H.C. and Rozema, H.R. (1986) Biological elimination of volatile xenobiotic compounds in biofilters. Bioproc. Biosys. Eng. 1: 61-69. Stucki, G., Gaelli, R., Ebershold, H.R. and Leisinger, T. (1981) Dehalogenation of dichloromethane by cell extracts of Hyphomicrobium DM». Arch. Microbiol. 130: 366-371. Wang, J. and Chen, J. (2006) Removal of dichloromethane from waste gases with a bio-contact oxidation reactor. Chem. Eng. J.123: 103-107.
515
MICROBIAL ECOLOGY AND BIOFILMS
This page intentionally left blank
517
Development of a reliable extraction method for the recovery of total genomic DNA from woodchip colonizing biofilm involved in gas biofiltration LÉA CABROL1, LUC MALHAUTIER2, JANICK ROCHER2, FRANCK POLY3, XAVIER LE ROUX3, MARC JOVIC1, ANNE-SOPHIE LEPEUPLE1 AND JEAN-LOUIS FANLO2 1
Anjou Recherche-Veolia Water, Chemin de la Digue, BP76, 78600, Maisons Laffitte, France Laboratoire Génie de l’Environnement Industriel, Ecole des Mines d’Alès, Avenue de Clavières 6, 30319, Alès Cedex, France 3 Laboratoire Ecologie Microbienne du Sol, UMR-CNRS 5557, Bâtiment Gregor Mendel, Université Claude Bernard Lyon I, 43, boulevard du 11 Novembre 1918, 69622, Villeurbanne Cedex, France 2
ABSTRACT This preliminary study focused on a critical step for the characterization of microbial ecosystem involved in biofiltration. Two aspects of nucleic acid recovery were explored: (i) cell dispersion (three methods tested) and (ii) total DNA extraction (four methods tested). The objective is to select the optimal combination of desorption/extraction methods, allowing subsequent molecular investigations to be reliable. Three relevant criteria are used to assess extraction efficiency: DNA amount and purity, and subsequent amplification feasibility.
1 INTRODUCTION During the past two decades, most studies concerning biofiltration concentrated on two main objectives: (i) assessment of operating parameter impact (e.g. packing material impact, pH effect: Kim et al., 2000; Prado et al., 2006); and (ii) definition of the system limits under different loading conditions (Aizpuru et al., 2001; VergaraFernandez et al., 2007). Both problematic were based on elimination performance evaluation, without regard for intrinsic biological phenomena. It was noticeable that a
517
518
L. CABROL, L. MALHAUTIER, J. ROCHER, F. POLY, X. LE ROUX, M. JOVIC, A. S. LEPEUPLE AND J. L. FANLO
number of studies tended to consider the system according to a «black-box» approach, while in a biofilter the pollutant removal is mainly due to the microbial component, whose structure and activity still remain unclear or even unknown. Microbial ecology aims to characterize microbial communities by their structure (i.e. their diversity, stability, spatial and temporal dynamics, occurrence of specific groups), as well as their interactions with the environment. Along the last decade, staggering progress in molecular biotechnologies offered powerful tools which made possible the fine characterization of microbial communities, granting access to uncultivable microorganisms. Among these molecular tools are quantitative PCR, fingerprintings, clone libraries sequencing, and, more recently, metagenome shotgun sequencing. They have been applied in various ecosystems, such as soil (Patra et al., 2006), sea sediments (Venter et al., 2004), anaerobic sludge (Godon et al., 1997), wastewater treatment biofilters (Ahn et al., 2004). More and more studies are carried out to elucidate community structures in gas biofilters. To date, several molecular tools have been used to gain insight into the dynamic diversity of bacterial communities in biofilters: ARISA (Steele et al., 2005), SSCP (Khammar et al., 2005), RFLP (Khammar et al., 2005), amoA gene PCR, cloning and sequencing (Sakano and Kerkhof, 1998), DGGE on 16S rDNA (Sercu et al., 2005; Cai et al., 2006; Li and Moe, 2004; Shim et al., 2006; Chung, 2007), DGGE on 16S rRNA (Sercu et al., 2006), FISH (Friedrich et al., 1999; Friedrich et al., 2003). Adopting a microbial ecology approach is of prime interest to reach a better understanding of biological mechanisms occurring within a biofilter. This better understanding may help to control, stabilize and optimize the biological process. Biofilms involved in biofilters are constituted of a complex and uncharacterized microflora attached to the packing material. Thus an essential preliminary task for the investigation of microbial communities with molecular tools is to implement and optimize a methodology for the recovery of nucleic acids. To get samples as representative as possible of the initial diversity, this methodology has to be the least selective as possible. In other ecosystem studies, such as soil, a lot of work was carried out to compare and implement DNA recovery methods that exhibit an unbiased sampling of the investigated community (Robe et al., 2003). However, within biofiltration context, very little attention was paid to the methodological aspects of nucleic acids recovery, despite their crucial importance in final results significance (Khammar et al., 2004; Li and Moe, 2004). DNA recovery methods are very heterogeneous and have neither been standardized nor optimized to date. Indeed, they are hugely dependent on the packing material specificities (size, organic/inorganic nature, density, hardness, porosity). In this preliminary study we explored methodological aspects of the nucleic acids recovery from microbial communities involved in a laboratory scale biofilter
DEVELOPMENT OF A RELIABLE EXTRACTION METHOD FOR THE RECOVERY
519
filled with pine bark woodchips. Considering the packing material used in this work, direct DNA extraction could not be applied. Hence two successive steps had to be performed: cell desorption (crushing, shaking, sonication) and DNA extraction (three commercial kits –two of which being specific for soil– and a reference protocol). The objectives of the present work were (i) to optimize cell desorption from the packing material, and (ii) to select the optimal combination of desorption and extraction methods. To assess DNA recovery efficiency, importance was attached to three relevant criteria: extracted DNA amount and purity, as well as subsequent amplification yield (this latter data will only be presented orally).
2 MATERIALS AND METHODS 2.1 FIRST STEP: OPTIMIZATION OF CELL DESORPTION 2.1.1 BIOFILTER SETUP Experiments were conducted on the biomass which colonized a lab-scale biofilter (1 m height, 125 mm diameter) filled with pine bark woodchips (initial porosity of 37%) and treating an H2S stream (10 mgH2S/m3; 500 m/h). The packing material was kept at constant humidity by regularly spraying a salt mineral nutrient solution, whose composition was previously described (Lalanne et al., 2007), at a rate of 150 mL every six hours. The biofilter was run at ambient temperature. 2.1.2 DESORPTION METHODS AND OPERATING CONDITIONS FOR OPTIMIZATION Three commonly used detachment methods were investigated for microorganism removal from woodchip support: blending (performed by an Ultra-Turrax -T25 basic, Ika); shaking (performed by a Vibro-Shaker –Retsch MM200); sonication (performed by an ultrasonic bath -Branson sonifier bath, Energy). One or two parameters were retained as potentially influent to define optimal conditions for each method, as shown in Fig.1. Other parameters (revolution speed, rotating frequency) were maximal.
Figure 1. Desorption methods and optimization conditions.
520
L. CABROL, L. MALHAUTIER, J. ROCHER, F. POLY, X. LE ROUX, M. JOVIC, A. S. LEPEUPLE AND J. L. FANLO
Sonication duration range was chosen from previous results (Khammar et al., 2004), which demonstrated that ultrasonic treatment needed higher duration to suspend microorganisms. On the contrary, blending treatment did not need more than two minutes for a complete homogenization of the suspension. 2.1.3 SAMPLING SCHEME At the time of sampling, the biofilter had reached a steady state, with complete H2S removal. Each treatment was carried out in triplicate, on woodchips extracted from the biofilter at the same time and at the same location (about 0.5 m high). Sampling procedure is shown in Fig. 1. Each sample was constituted of 5 g of woodchips, suspended in 15 mL of sterile physiological serum (NaCl, 9 g/L). Aliquots of 1 mL of the liquid suspension were collected and enumerated. 2.1.4 MICROSCOPIC DIRECT COUNTS Total bacteria were enumerated by fluorescence microscopy using DAPI staining (Sigma, USA). Whole experiment is done in sterile conditions. After serial dilutions, samples were stained with DAPI at a final concentration of 20 μg/mL during 30 minutes in the dark in a shaker (200 rpm). Stained bacteria were collected on 0.2 μm polycarbonate membrane filters (Millipore GTBP, Ireland) by vacuum microfiltration. Filters were mounted on microscope slides in Mounting Medium (Sigma, USA) and observed with an epifluorescence microscope (Leica DMLB) equipped with a blue excitation filter (BP 340-380 nm) and a barrier filter LP 425. Thirty microscopic fields per slide were enumerated. 2.2 SECOND STEP: COMBINATION OF CELL DESORPTION AND DNA EXTRACTION 2.2.1 BIOFILTER SETUP Experiments were conducted on the same pilot-scale unit, but the biofilter was treating a VOC mixture made of acetaldehyde, acetone, butanal, MEK, DMDS, butanoic acid, isovaleric acid. At the time of biomass sampling, operating conditions were as follows: 10 mg/m3 for each compound; gas velocity at 100 m/h. 2.2.2 EXPERIMENTAL PROCEDURE The experiment involved 12 samples, each constituted of 5 g of woodchips suspended in 15 mL of sterile physiological serum (NaCl, 9 g/L). Each sample was repeated twice. As shown in Figure 2, each sample was submitted to one of the three desorption methods, under previously optimized conditions (described in section 3.1.). After centrifugation of liquid phase at 10 000 rpm for 10 minutes, the pellet was subjected to DNA extraction, using one of the four following methods: I. PowerSoil DNA Kit, MoBio (Ozyme, France); II. FastDNA® SPIN Kit for Soil, Qbiogene (MP Biomedicals, France); III. NucleoSpin® Tissue Kit (Macherey –Nagel, France); IV. An extraction protocol adapted from Godon et al. (1997).
DEVELOPMENT OF A RELIABLE EXTRACTION METHOD FOR THE RECOVERY
521
Figure 2. Schematic procedure to assess the optimal combination of desorption and extraction methods (in dotted lines: experiments in prospect).
2.2.3 EXTRACTION AND PURIFICATION OF TOTAL GENOMIC DNA Extraction by commercial kits was performed according to the manufacturers’ instructions, using a Vibro Shaker Retsch for cell disruption. In all cases, elution volumes were 50 μL. The fourth method was slightly modified from the one described by Godon et al. (1997), as follows, to ensure the largest sample size. The microbial cell fractioncontaining pellet obtained after desorption and centrifugation was resuspended in 385 μL of 4M guanidine thiocyanate-0.1 M Tris (pH 7.5) and 115 μL of 10%-Nlauroyl sarcosine. Samples were stocked at -20°C. After the addition of 500 μL of 5% -N-lauroyl sarcosine-0.1 M phosphate buffer (pH 8.0), the sample was incubated at 70°C for 1 h. One volume (500 μL) of 0.1 mm-diameter sterile zirconium beads (Sigma) was added and the sample was shaken at maximum speed (30 Hz) for 10 min in a Vibro Shaker (Retsch MM200). Polyvinylpolypyrrolindone (PVPP, 15 mg) was added. The sample was vortexed and centrifuged for 3 min at 12 000 rpm. The supernatant was recovered. The pellet was washed with 500 μL of TENP (50 mM Tris [pH 8.0], 20 mM EDTA [pH 8.0], 100 mM NaCl, 1 % PVPP) and centrifuged for 3 min at 12 000 rpm. The new supernatant was pooled with the first one. The washing step was repeated three times. The pooled supernatants were centrifuged for 3 min at 12 000 rpm to remove particles, and then split into two 2-mL tubes. Samples were incubated for 1h30 at 56 °C with 30 μL of proteinase K (20 mg/ mL), and then incubated 1 h more at 37°C with 20 μL of RNase A (10 mg/mL). Samples were split into 500 μL subsamples and crude DNA was purified by addition of 1 mL of phenol-chloroforme-isoamyl alcohol (25:24:1). After centrifugation for 5
522
L. CABROL, L. MALHAUTIER, J. ROCHER, F. POLY, X. LE ROUX, M. JOVIC, A. S. LEPEUPLE AND J. L. FANLO
min at 10 000 rpm, the upper phase was recovered. Nucleic acids were precipitated by the addition of 50 μL of sodium acetate 3M and 1 mL of cold absolute ethanol. Samples were incubated for 15 min at -80°C and 30 min at -20°C. After centrifugation for 30 min at 14 000 rpm, supernatant was discarded and DNA pellet was washed with about 1 mL of cold 70% ethanol, dried for 10 min at room temperature and resuspended in 50 μL of Tris-EDTA 0.1X. 2.2.3 EVALUATION OF DNA RECOVERY EFFICIENCY Sizing and quantification of extracted DNA were assessed by electrophoresis. 5 μL of extraction product were loaded in 1% agarose gel. Migration was performed at 85 V, for 45 min, in 1X TAE buffer and gel was stained with ethydium bromide. DNA amount was further determined by absorbance at 260 nm using an UV spectrophotometer (Biophotometer, Eppendorf). DNA concentration was calculated considering that 1 UDO corresponds to a double-strand DNA concentration of 50 μg/ mL, in 1 cm cuvettes. DNA purity was determined by the ratio of absorbance at 260 nm and absorbance at 280 nm (Biophotometer, Eppendorf), considering that the absorbance at 280 nm is mainly due to protein contamination.
3 RESULTS AND DISCUSSION 3.1 OPTIMIZATION OF CELL DESORPTION Bacterial counts after desorption are presented in Figure 3. When a single parameter was variable (treatment duration), results were statistically analyzed by ANOVA (analysis of variance). After verifying variance homoscedasticity with a Hartley test, the significance of differences between means was established by the Fisher-Snedecor test with a risk level of 0.05. It appeared that for blending desorption (Figure 3.A), treatment duration between 0.5 and 2 minutes did not influence the amount of recovered cells. Nevertheless, increasing blending duration led to more deviation. Indeed the longer was the blending, the more organic particles were suspended, which seriously hampered microscopic counting, thus leading to higher experimental errors. On the contrary, cell counts obtained after different sonication durations were not statistically equal (Fig. 3.B): cell removal was significantly improved by increasing sonication duration up to 60 minutes. Concerning shaking desorption, a Doehlert matrix was built. In the model provided by NemrodW analysis, the most significant coefficient is the constant one, i.e. the coefficient linked to none experimental factor (data not shown). Hence it can be concluded that neither shaking duration nor shearing force increase (by adding glass beads) improved shaking efficiency. These results are in accordance with those of Khammar et al. (2004), which detected no significant effect of glass
DEVELOPMENT OF A RELIABLE EXTRACTION METHOD FOR THE RECOVERY
523
Figure 3. Influence of treatment duration and glass bead amount on total microbial cells recovered after desorption and enumerated by DAPI (A: blending by Ultra-Turrax; B: Sonication; C: Vibro-Shaking). Graphs are in logarithmic scale and error bars represent standard deviation calculated on triplicates.
524
L. CABROL, L. MALHAUTIER, J. ROCHER, F. POLY, X. LE ROUX, M. JOVIC, A. S. LEPEUPLE AND J. L. FANLO
beads and treatment duration on cell detachment from peat and activated carbon by blending and shaking and highlighted treatment duration effect for sonication. As a result, optimal conditions for desorption treatments are chosen as follows: 1 min for blending; 60 min for sonication; 10 min without glass beads for shaking. Blending could have been thought to be the most efficient detachment method, as it allowed recovery and homogenization of the initial material in its entirety (no biomass was lost), while sonication and shaking only suspended microorganisms (a fraction of biomass left on the support may be lost). This was observed by Khammar et al. (2004). Nevertheless, in this previous study, biomass detachment was only evaluated by the number of viable and cultivable microorganisms. As enumeration results were in the same range whatever was the desorption method, whole three methods had to be further compared, on the basis of more accurate criteria. 3.2. OPTIMAL COMBINATION OF CELL DESORPTION AND DNA EXTRACTION METHODS Electrophoresis results are shown in Fig. 4. Only one sample is presented for each duplicate (except for MoBio extraction method, where results were not reproducible).
Figure 4. Electrophoresis of DNA extracted by four different methods, after three different desorption treatments (UT: UltraTurrax; US: UltraSonication; VS: VibroShaking).
Conclusions drawn from band intensity observations were confirmed and completed by absorbance measures, as shown in Fig. 5. It is important to note the bad reproducibility of the results between duplicates. This would be explained by the random aspect of bacterial colonization on the initial 5g-sampling. It is obvious that Godon-adapted extraction protocol led to significantly higher DNA amounts, when compared to commercial kit extraction (about 15 times higher on average). This is observed independently of the previous desorption methods used. But this protocol being highly time-consuming, it is not realistic to envisage its use for routine DNA extraction of numerous samples. It is thus considered as a reference protocol.
DEVELOPMENT OF A RELIABLE EXTRACTION METHOD FOR THE RECOVERY
525
A
B Figure 5. DNA concentration and purity after different methods of desorption and extraction, evaluated by A260 and ratio A260/A280, respectively (concentrations shown for both duplicates).
It should be noted that, whatever the extraction method was, desorption treatment by Ultra Turrax was clearly not suitable for high DNA extraction yield. DNA concentration after kit extraction was about 20 μg/mL and did not exceed 400 μg/mL with reference extraction protocol. Indeed, Ultra Turrax treatment led to a single-phase suspension where the whole initial material was homogenized. Therefore, in the pellet obtained after centrifugation, the relative proportion of cells was very low compared to the proportion of crushed wooden material. Moreover, blending detachment gave the worst results in terms of DNA purity (Fig. 5.B), probably because of the high amount of organic material in blended samples. In all cases, DNA recovery was higher when previous desorption was performed by Vibro Shaker. After shaking detachment, DNA concentration was about 130 μg/mL when kit-extracted and even reached 2000 μg/mL
526
L. CABROL, L. MALHAUTIER, J. ROCHER, F. POLY, X. LE ROUX, M. JOVIC, A. S. LEPEUPLE AND J. L. FANLO
with reference protocol. Cell removal by sonication gave intermediate results in terms of DNA recovery. DNA extraction by MoBio kit was the least efficient for DNA recovery (<30 μg/mL). In contrast, DNA extraction by both Qbiogene and MachereyNagel kits provided satisfactory DNA amounts, especially after biomass removal by shaking (about 200 μg/mL). Considering DNA purity, the worst results were obtained with the universal kit (Macherey-Nagel), followed by reference extraction protocol. Better DNA purity was gained with the two commercial kits specifically designed for DNA extraction and purification from soil samples. These kits (MoBio and Qbiogene) aimed to remove DNA contaminating organic substances, which seemed to result in improved DNA purity (Fig. 5.B). DNA extraction by Qbiogene kit was even more interesting as it was the only case where the ratio A260/A280 exceeded 1.75. To conclude, the experimental strategy implemented in this study allowed comparison and selection of a reliable combination of cell-desorption and DNAextraction methods, considering both DNA amount and DNA purity as decisive criteria (Fig. 6). As a result, the most appropriate methodology seems to be a desorption step with Vibro-Shaking, followed by an extraction step with Qbiogene kit.
Blending (Ultra-Turrax) Sonication Shaking (Vibro-Shaker)
MoBio kit -;
Qbiogene kit -;-
Macherey-Nagel kit -;-
Reference protocol ;-
-;-;-
-; ;
-;;-
; ; -
Figure 6. Summary of different combinations efficiency, basing on recovered DNA quantity (symbolized by ) and purity (designed by ).
DEVELOPMENT OF A RELIABLE EXTRACTION METHOD FOR THE RECOVERY
527
REFERENCES Ahn Y., Park, W., Tatavarty, R. and Kim, I.S. (2004) Comparative analysis of vertical heterogeneity of microbial community in sulfur-packed reactor used for autotrophic nitrate removal. J. Environ. Sci. Health. 39 - v3939(7): 1805-1818. Aizpuru, A., Malhautier, L., Roux, J.C. and Fanlo, J.L. (2001) Biofiltration of a mixture of volatile organic emissions. J. Air Waste Manage. Assoc. 51: 1662-1670. Cai, Z., Kim, D., Sorial, G.A., Saikali, P., Zein, M.M. and Oerther, D.B. (2006) Performance and microbial diversity of a trickle-bed air biofilter under interchanging contaminants. Eng. Life Sci. 6(1): 37-42. Chung, Y.C. (2007) Evaluation of gas removal and bacterial community diversity in a biofilter developed to treat composting exhaust gases. J. Hazard. Mat. 144: 377-385. Friedrich, U., Naismith, M.M., Altendorf, K. and Lipski, A. (1999) Community analysis of biofilters using fluorescence in situ hybridization including a new probe for the Xanthomonas branch of the class Proteobacteria. Appl. Environ. Microbiol. 65(8): 3547-3554. Friedrich, U., Van Langenhove, H., Altendorf, K. and Lipski, A. (2003) Microbial community and physicochemical analysis of an industrial waste gas biofilter and design of 16S rRNAtargeting oligonucleotide probes. Environ. Microbiol. 5(3): 183-201. Godon, J.J., Zumstein, E., Dabert, P., Habouzit, F. and Moletta, R. (1997) Molecular microbial diversity of an anaerobic digestor as determined by small-subunit rDNA sequence analysis. Appl. Environ. Microbiol. 63(7): 2802-2813. Khammar, N., Malhautier, L., Degrange, V., Lensi, R. and Fanlo, J.L. (2004) Evaluation of dispersion methods for enumeration of microorganisms from peat and activated carbon bioiflters treating volatile organic compounds. Chemosphere. 54: 243-254. Khammar, N., Malhautier, L., Degrange, V., Lensi, R., Godon, J.J. and Fanlo, J.L. (2005) Link between spatial structure of microbial communities and degradation of a complex mixture of volatile organic compounds in peat biofilters. J. Appl. Microbiol. 98: 476-490. Kim, N.J., Hirai, M. and Shoda, M. (2000) Comparison of organic and inorganic packing materials in the removal of ammonia gas in biofilters. J. Hazard. Mat. B72: 77-90. Lalanne, F., Malhautier, L., Roux, J.C. and Fanlo, J.L. (2007) Absorption of a mixture of volatile organic compounds (VOCs) in aqueous solutions of soluble cutting oil. Biores. Technol. (in press). Li, C. and Moe, W.M. (2004) Assessment of microbial populations in methyl ethyl ketone degrading biofilters by denaturing gradient gel electrophoresis. Appl. Microbiol. Biotechnol., 64(4): 568-575. Patra, A.K., Abbadie, L., Clays, A., Degrange, V., Grayston, S., Guillaumaud, N., Loiseau, P., Louault, F., Mahmood, S., Nazaret, S., Philippot, L., Poly, F., Prosser, J.I. and Le Roux, X. (2006) Effects of management regime and plant species on the enzyme activity and genetic structure of N-fixing, denitrifying and nitrifying bacterial communities in grassland soils. Env. Microbiol. 8: 1005-1016.
528
L. CABROL, L. MALHAUTIER, J. ROCHER, F. POLY, X. LE ROUX, M. JOVIC, A. S. LEPEUPLE AND J. L. FANLO
Prado, O.J., Veiga, M.C. and Kennes, C. (2006) Effect of key parameters on the removal of formaldehyde and methanol in gas-phase biotrickling filters. J. Hazard. Mat. B138: 543-548. Robe, P., Nalin, R., Capellano, C., Vogel, T.M. and Simonet, P. (2003) Extraction of DNA from soil. Eur. J. Soil Biol. 39: 183-190. Sakano, Y. and Kerkhof, L. (1998) Assessment of changes in microbial community structure during operation of an ammonia biofilter with molecular tools. Appl. Environ. Microbiol. 64(12): 4877-4882. Sercu, B., Nunez, D., Aroca, G., Boon, N., Verstraete, W. and Van Langenhove, H. (2005) Inoculation and start-up period of a biotrickling filter removing dimethyl sulphide. Chem. Eng. J. 113: 127-134. Sercu, B., Boon, N., Verstraete, W. and Van Langenhove, H. (2006) H2S degradation is reflected by both the activity and composition of the microbial community in a compost biofilter. Appl. Microbiol. Biotechnol. 72 (5): 1090-1098. Shim, E.H., Kim, J., Cho, K.S. and Ryu, H.W. (2006) Biofiltration and inhibitory interactions of gaseous benzene, toluene, xylene and methyl tert-butyl ether. Environ. Sci. Technol. 40(9): 3089-3094. Steele, J.A., Ozis, F., Fuhrman, J.A. and Devinny, J.S. (2005) Structure of microbial communities in ethanol biofilters. Chem. Engin. J. 113: 135-143. Venter, J.C., Remington, K., Heidelberg, J.F., Halpern, A.L., Rusch, D., Eisen, J.A., Wu, D., Paulsen, I., Nelson, K.E., Nelson, W., Fouts, D.E., Levy, S., Knap, A.H., Lomas, M.W., Nealson, K., White, O., Peterson, J., Hoffman, J., Parsons, R., Baden-Tillson, H., Pfannkoch, C., Rogers, Y.H. and Smith H.O. (2004) Environmental Genome Shotgun Sequencing of the Sargasso Sea. Science. 304: 66-74.
529
FT-IR characterization of biofilms formed on engineered biofiltration media treating volatile organic emissions for the forest products industry KIM JONES1, MILI KHILNANI1, ANAND KARRE1, SERGIO SANTOS1 AND JAN PACA2 1
Environmental Engineering, Texas A&M University-Kingsville, MSC 213, Kingsville, Texas 78363, USA 2 Institute of Chemical Technology, Department of Fermentation Chemistry and Bioengineering, Technicka 5, 160 28 Prague, Czech Republic
ABSTRACT The gaseous emissions from hardboard mill presses at lumber plants contain both volatile and condensable organic compounds, as well as fine wood and other very small particulate material. Biological emissions control for these compounds present several challenges. The biofiltration media provides support and contact between the gas phase contaminants and active microbial cultures attached as biofilms on the media’s surface. As the transformations in the biofilm and the media during optimal biofiltration operations are not well understood, the main aim of this project was to characterize the biofilm formed on the media during the biofiltration process using Fourier Transform Infrared Spectroscopy (FT-IR) and the FT-IR Microscope, and also examine the results along with the performance data of VOC biofiltration field and pilot scale tests. Some differences in the absorbance spectra were observed in the media and biofilm samples collected from the top and the bottom bed of the biofilters. This work suggests that while FT-IR spectral information can provide some useful insights to biofilm coverages and quality within media sections, more work and measurements will be needed to correlate the information to biofiltration performance and optimization.
529
530
KIM JONES, MILI KHILNANI, ANAND KARRE, SERGIO SANTOS
AND
JAN PACA
1 INTRODUCTION Industrial plants and processes use and emit various types of volatile organic compounds (VOCs), which are amenable to biological treatment. Biofiltration is an emerging and an attractive air pollution control technology for the removal of VOCs present at low concentration in the emissions from forest products plants. The filter media is one of the most critical elements of a biofiltration system as it provides contact between the gas phase contaminants and active microbial cultures either immobilized within and/ or attached as a biofilm on the media’s surface. Examination and characterization of the biofilms formed during bioloigcal treatment may provide some insight into process optimization and further enhancement of the technology. An infrared spectrum represents a fingerprint of a sample with absorption peaks which correspond to the frequencies of vibrations between the bonds of the atoms making up the material. Each different material is a unique combination of atoms, so no two compounds produce the exact same infrared spectrum. Fourier Transform Infrared Spectroscopy can thus result in positive identification (qualitative analysis) of different types of material. Additionally, the size of the peaks in the spectrum is a direct indication of the amount of material present. FT-IR can be very useful in determining the quality and consistency of multiple samples. Other researchers have begun to examine the utility of FT-IR characterization in biological systems. Haberhauer et al. (1999) used FT-IR to characterize the decomposition processes of spruce litter in organic soil layers from several forest regions. The investigators found a broad intense band at 3400 cm-1 wavenumber due to stretching vibrations of bonded and nonbonded hydroxyl groups and another broad band at 1630 cm-1 due to C=O vibrations of carboxylates and aromatic vibrations. A peak at 1510 cm-1 was attributed to amide vibrations and aromatic C=C vibrations and the transmission FT-IR band at 1630 cm-1 showed significant correlation to soil organic C content. When biological activated carbon was used as a support medium for biofilters treating gas contaminants, FT-IR analysis was performed on the carbon samples collected after the process (Duan et al., 2005). The fingerprint of the spectra showed peaks 1120 and 580 cm-1, which represented the vibration from SO42- indicating sulfuric acid as the dominant oxidation product of H2S. In this project, engineered biomedia samples from a field scale sequential biotrickling-biofiltration unit operated by BioReaction Industries LLC at the Stimson Lumber Plant in Gaston, Oregon, USA and samples from an experimental biofiltration experiment conducted in the laboratory at the Texas A&M University-Kingsville treating á-pinene emissions were evaluated with FT-IR techniques. An FT-IR fingerprint of the biofilms on the biofilter media located in different sections of the biofiltration process was obtained. Biofilm and media samples were collected from the top and the bottom bed of the biofilters. The FT-IR bench (Attenuated Total Reflectance) unit
FT-IR CHARACTERIZATION OF BIOFILMS FORMED ON ENGINEERED BIOFILTRATION
531
spectra and spectra from an FT-IR microscope (Transmission) were also compared and both found to be useful for certain evaluations. The main aim of this project was: 1) to characterize the engineered VOC biofiltration media biofilm samples from a field scale operation at the Stimson Lumber Company application using ATR FT-IR analysis and the FT-IR Microscope; 2) to compare the biofilm samples from different sections of the biofiltration units; and 3) to compare of the quality of biofilm samples from different locations on the media surfaces.
2 MATERIALS AND METHODS The Nicolet Nexus 470 FT-IR unit which has a relatively simple interferometer assembly was used in this study. The Fourier Transform aspect of the infrared response defines a relationship between a signal in the time domain and its representation in the frequency domain. The main component of the FT-IR is the interferometer. It splits and recombines a beam of light such that the recombined beam produces a wavelength-dependent interference pattern called interferogram. The interferometer consists of 2 mirrors and a KBr beamsplitter positioned at an angle of 45° to the mirrors. Incident light strikes the beamsplitter so that half of the light is transmitted through the beamsplitter and half to the mirrors. The two components are then reflected back and recombined into the beamsplitter with half of the light passing on toward the sampling areas and half traveling back to the source (Thermo Nicolet, 1999).
Figure 1. FT-IR laboratory at TAMUK.
Figure 2. FT-IR bench with Smart Miracle Accessory
532
KIM JONES, MILI KHILNANI, ANAND KARRE, SERGIO SANTOS
AND
JAN PACA
Several IR techniques are available for biological investigations. The analysis can be performed using Specular Reflectance, Diffused Reflectance and Transmission techniques. Different accessories are needed for each of these tests. The Thermo Nicolet Continuμm Microscope at TAMUK (Figure 1) provides both high performance infrared sampling and visible-light microscopy. It can be used in the Transmission or the Reflectance modes. It provides a continuous view of the sample while simultaneously collecting the data. The Reflex aperture provides redundant infrared masking, which reduces the effects of diffraction. With polychromatic light (radiation with more than a single wavelength), the output signal is the sum of all signals at the detector which is the Fourier Transform of the spectrum also called interferogram (Thermo Nicolet, 1999). The interferogram contains basic information on frequencies and intensities characteristic of a spectrum, which is then converted into more familiar forms using Fourier Transform methods. The interferogram is a function of time domain and the spectrum is frequency domain. Band intensities can be expressed either as transmittance (T) or absorbance (A). Transmittance is a ratio of radiant power (I) transmitted by a sample to the radiant power incident on the sample. Absorbance is the algorithm, to the base 10, of the reciprocal of the transmittance. The Thermo Nicolet FT-IR bench unit at TAMUK uses an ATR (Attenuated Total Reflectance) Smart Miracle accessory for the analysis of the samples (Figure 2). ATR is a non-destructive surface analysis of strong IR absorbing materials. The depth of penetration ranges from 0.6 μm to 2.0 μm. The FTIR ATR method offers the advantage of in-situ examination and also an ability to characterize chemical changes that occur due to the presence of moisture. The advantage with the ATR is that very little sample preparation is required. SAMPLE PREPARATION: Media samples were collected and preserved at 3oC until they were ready for measurement. Microtweezers and micro-sampling blades were used to lightly scrape the outer surface of the media and collect a small amount of wet sample for FT-IR analyses (usually less than 1 mg required). For the Continuum Microscope: The axes were perfectly aligned and the detector cooled using liquid nitrogen. The background was collected by placing the gold mirror in the objective and observing it under the microscope (Figure 3a). Without changing the settings, small particles of the media sample were collected and placed on the gold mirror and the spectra was measured. For the FT-IR Bench: A background spectrum was collected by collecting the raw media samples not used in the biofiltration process. The media samples from the field were observed visually and noted in terms of their color and dryness. The wet
FT-IR CHARACTERIZATION OF BIOFILMS FORMED ON ENGINEERED BIOFILTRATION
533
layer (biofilm) on these media samples was collected and also analyzed. The sample was kept in direct contact between the halide crystals of the bench-scale FT-IR unit (Figure 3b), and the spectra was obtained and replicated at least twice. This fingerprint spectra of the sample was then subtracted from that of the background dry material, and the resultant composition of the wet sample was obtained.
(a) (b) Figure 3. Sample compartment for the FT-IR showing (a) Microscope gold mirror slide, and (b) Bench (Smart Miracle sample compartment) for the ATR.
3 RESULTS AND DISCUSSION The forest products industry encompasses a vast number of wood processes, which can utilize large amounts of energy during product manufacturing and sometimes in emissions control applications. The forest products sector includes logging, sawmills and planning mills, softwood veneer and plywood mills, hardboard veneer and plywood mills, particleboard, oriented strand board (OSB), other reconstituted wood products mills and pulp and paper mills. The operations can produce air emissions, including fugitive emissions, which may amount to several thousand tons of VOCs and HAPs per year potentially released into the atmosphere. The Stimson Lumber Company is a private forest products and natural resource company with a plant located in the vicinity of the city of Gaston, Oregon, USA, which emits low concentrations of VOCs and HAPs into the atmosphere from the plant’s wet hardboard making process through the main ventilation system. In 2005 through mid 2006, a pilot scale biological treatment unit consisting of a biotrickling filtration section and a biofilter section in series was installed to treat a portion of the flow from the main press vent. The biofilter media beds were comprised of engineered media of organic substrate formed around inorganic support material (manufactured by BioReaction Industries as BioAIRSpheres®) and connected in series. The uniform-shape engineered media were 2.54 x 10-2 m hollow polyethylene spherical structures packed with organic, nutrient-providing, compost-based media. Both units, the BTF and BF, shared a water
534
KIM JONES, MILI KHILNANI, ANAND KARRE, SERGIO SANTOS
AND
JAN PACA
sump with volume of nearly 5.66 m3. The biological system operated at approximately 0.71 m3/s of inlet air, and provided an EBCT (empty bed contact time) of 45 seconds for the entire bio-oxidation system. In spite of the fluctuation of the press vent emissions from the wet-process hardboard plant of from 12-20 ppm VOCs in the inlet stream, the pilot scale biofiltration system successfully dampened the irregularities and kept the Total VOC emissions below the objective of 5 ppm throughout the project period (Santos et al., 2007). Most of the VOCs encountered were characterized by GC-MS as aldehydes, ketones and α-pinene emissions. Solid samples collected from several locations within the media beds were preserved at 3oC and shipped to Texas A&M Kingsville for FT-IR analyses. Several functional groups which might be encountered in the organic samples were targeted for FT-IR examination. Table 1 shows some potential wavenumbers and functional groups chosen for examination. Table 1. Functional groups of compounds potentially encountered in the biofiltration media treating α-pinene and formaldehyde. Functional Class Alcohols and Phenols
Aldehydes Carboxylic Acid
Alpha-pinene
Range (cm-1) 3580-3650 3200-3550 970-1250 2690-2840 1720-1740 2500-3300 1705-1720 1210-1320 700-900
Assigned Groups sharp O-H O-H (H-bonded) strong C-O C-H C=O (saturated aldehyde) broad O-H C=O (H bonded) strong O-C
Table 2 shows the unique identification and sample collection locations for the different samples used for analysis using the ATR FT-IR bench and the FT-IR Microscope.
FT-IR CHARACTERIZATION OF BIOFILMS FORMED ON ENGINEERED BIOFILTRATION
535
Table 2. Samples analyzed using the FT-IR ATR and the FT-IR Microscope.
Sample IDs I-030107 II-030107 III-030107 IV-031207 V-020806 VI-031207
Samples Analyzed Location and analyses Right surface of media collected on 3-1-07 Left surface of media collected on 3-1-07 Bottom surface of media collected on 3-1-07 Bottom of media collected on 3-12-07 Bottom of media collected on 2-8-06 Microscopic analysis of media collected on 3-12-07
Figure 4. Photograph of typical Bioairsphere® media prior to wet film sampling.
Figure 5. Bench data of samples I-030107 collected on the 3-1-07 for wavenumbers 2690-2830 cm-1, representative of aldehydes and carboxylic acid groups, showing a comparison of the spectra of the right surface of the top and the bottom media of the biofilter.
536
KIM JONES, MILI KHILNANI, ANAND KARRE, SERGIO SANTOS
AND
JAN PACA
Figure 4 shows a typical media sphere collected from the biofiltration units. Small scrapings of biofilms were collected from different locations on the media and examined. Figure 5 shows a comparison of the right surface of top and the bottom section media samples. Peaks of C-H bands were observed in both the bottom media and top media in the range of 2690-2830 cm-1, representative of aldehydes. This could suggest a slight build up of aldehydes in the top bed as the biofilm may be less active at this point. Samples V-020806 were collected on 2-8-06, from the top and the bottom of the field biofilter bed at Stimson Lumber. Physical examination of these samples showed slight variations in the biofilm coverage of the top and the bottom media of the biofilter. The media from the bottom biofilter was found to have a thicker coverage. It was more slimy and solid in color. The top media sample was drier due to moisture utilization in the biofilter. Figure 6 shows high absorption peaks representing C-H bands of aldehydes in the range of 2690-2790 cm-1 for both top and bottom media samples.
Figure 6. FT-IR bench data of samples V-020806 collected on 2-8-06, showing a comparison of spectra of the top and the bottom media of the Field VOC biofilter.
FT-IR CHARACTERIZATION OF BIOFILMS FORMED ON ENGINEERED BIOFILTRATION
537
Figure 7. Microscope derived FT-IR data of sample VI-031207 collected on 3-12-07, showing a comparison of the top and the bottom media of the Lab biofilter over the entire range of wavenumbers.
FT-IR measurements were also performed using the FT-IR Microscope on samples collected from the top and the bottom bed of a lab biofilter treating α-pinene. The spectra is shown in Figure 7. The broad bands in the bottom media in the range from 1800-3000 cm-1 are reflective of the higher water content and O-H stretch for the increased moisture in that range. A comparison of the two FT-IR techniques is possible which shows the increased sensitivity of the Continuum reflectance resolution, but more qualitative value of relative measurements and comparisons for the solid sample ATR.
4 CONCLUSIONS The FT-IR spectra obtained by examining the biofiltration media using the bench ATR unit and the Continuum FT-IR microscope showed biofilm differences at various wave numbers throughout the spectral ranges examined for these samples. Some significant differences in the absorbance spectra were observed in the media and biofilm samples collected from the top and the bottom bed of the biofilters. This work suggests that while FT-IR spectral information can provide some useful insights to biofilm coverages and quality within media sections, more work and measurements will be needed to correlate the information to biofiltration performance and optimization.
538
KIM JONES, MILI KHILNANI, ANAND KARRE, SERGIO SANTOS
AND
JAN PACA
5 ACKNOWLEDGEMENTS Portions of this study were supported by U.S. DOE contract DE-FC36-04GO14310, and by the National Science Foundation under Cooperative Agreement No. HRD0206259. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation. Also support by the Czech Science Foundation, Joint Project 104/05/0194 and the Ministry of Education of the Czech Republic, Research Project MSM 6046137305 is acknowledged.
REFERENCES Thermo Nicolet Cooperation (1999). Introduction to Fourier Transform Infrared Spectroscopy and the Continum Microscope publication on-line. Haberhauer, G. and Gerzabek, H.M. (1999) Drift and transmission FT-IR spectroscopy of forest soils: an approach to determine decomposition processes of forest litter. Vibrational Spectroscopy 19: 413-417. Duan, H. and Yan, R. (2005) Investigation on the mechanism of H2S removal by biological activated carbon in a horizontal biotrickling filter. Appl. Microbiol. Biotechnol. 69: 350-357. Santos, S., Jones, K., Abdul, R., Boswell, J., Paca, J. (2007) Treatment of wet process hardboard plant VOC emissions by a pilot scale biological system. Biochem. Engin. J. (in press).
539
Monitoring and characterization of bacterial populations of two biological air filters during the start up phase JOVIC, M., CABROL, L., DUCRAY, F., GAGNEUX, R. AND LEPEUPLE, A.S. Veolia Environnement – Centre de recherche sur l’eau (Anjou Recherche), Chemin de la Digue BP76, 78603 Maisons Laffitte, France
ABSTRACT This study aimed to monitor and characterize bacterial populations of two biological air filters during their start up phase (four months). The main objective of this work was to assess the potentiality of a microbiological approach to better understand the evolution of the bacterial populations within biofilters and therefore help to select biomass carrier media. The two biological filters were operated at full-scale (480 m3), filled with organic materials and dedicated to the removal of ammonia and Volatile Organic Compounds (VOCs). The first step of the work consisted in developing an extraction method for the biomass fixed on the solid supports. The second step investigated biofilters’ microbial ecology using molecular tools: DAPI (4,6-DiAmino-2PhenylIndole), TVC (Total Viable Counts), FISH (Fluorescent In Situ Hybridization) and SSCP (Single Strand Conformation Polymorphism). The findings of the experiments did not show a significant evolution of total bacterial concentrations in biofilms of both biological filters during their start up phase. However, SSCP data analysis underlined important variations in the composition of bacterial populations. Finally, examination of the results highlighted the interest to inoculate organic media in order to reduce the acclimation time of microbial populations.
1 INTRODUCTION Removal of Volatile Organic Compounds (VOCs) and ammonia has gained increasing importance within the composting sector. This issue is a result of the increases in regulation and in public expectations regarding the quality of their environment (Malhautier et al., 2005). The need for composting operators to design and apply
539
540
M. JOVIC, L. CABROL, F. DUCRAY, R. GAGNEUX AND A. S. LEPEUPLE
robust odour control strategies in order to minimise impact on residential areas has therefore become an important issue for the sector. A wide range of options is available to an operator to control the emissions and thereby reduce the impact on sensitive areas. In broad terms, odour control options can be characterized by two categories: process related control options and end-of-pipe abatement options. The principal behind end-of-pipe abatement techniques is to remove odorous compounds from the air collected from the process using a treatment system. This approach requires the odorous process to first be contained and collected with an extraction system. The emissions can then be treated using different techniques. One of the available technologies is the combination of an acid scrubber aiming at removing ammonia, and a biological filter for the treatment of VOCs. Indeed biofiltration has proved to be economical and environmentally viable (Aizpuru et al. 2001; Malhautier et al. 2005). Biofiltration processes are governed by two main phenomena. The first phenomenon is the mass transfer of contaminants from the gas phase to the liquid phase. Concentration in the biofilm is in equilibrium with the one in the gas phase according to Henry’s law (C*=P/H). The second phenomenon is the bioconversion of pollutants to biomass, metabolic-end-products, carbon dioxide and water. Chemical compounds are biodegraded in the biofilm, which contains organisms growing on the solid medium support. Considering these aspects, abatement efficiency in this process will depend on chemicals, physicals and microbiological parameters of the carrier media used (Malhautier et al., 2005). Nowadays physical and chemical parameters are commonly taken into account for the selection of biofilter carrier media. However, little attention has been paid so far to microbiological parameters. Therefore, the purpose of this study was to evaluate the potentiality of molecular tools to generate biological data useful for biofilter support media characterization. In order to compare the two biofilm carrier media, the monitoring and the characterization of bacterial populations of two biological air filters was realised during the start up phase (four months).
2 MATERIAL AND METHODS 2.1 COMPOSTING PLANT AND BIOFILTERS’ DESCRIPTION The composting facility studied was receiving compost from sewage sludge and green wastes. The ratio between sewage sludge and green waste varied from 1/4 to 1/3 depending on sewage sludge siccity. The composting building was fully enclosed and the odorous air extracted from this building was conveyed to a deodorization system designed to treat 184,000 m3/h of gaseous effluent. The extracted air was first treated by an acid scrubber and then injected into the biofilter.
MONITORING AND CHARACTERIZATION OF BACTERIAL POPULATIONS
541
The biofilter was composed of four units of a similar design and configuration. Each unit had a 320 m2 surface area, was 1.5 metre depth and was fed with 46,000 m3/ h of gaseous effluent containing ammonia and VOCs. Three units were filled with mixed peat/branches/wood fragments (biofilter B) and one was filled with mixed peat/ coco fibbers (biofilter A). The two biofilters (units A and B) studied were operated at full-scale (480 m3). 2.2 PHYSICAL AND CHEMICAL MONITORING OF BIOFILTERS 2.2.1 AMMONIA CONCENTRATION MEASUREMENT Ammonia concentration measurements were realised at the inlet and at the outlet of biofilters A and B using colorimetric Dräger tubes (Ammonia NH3-2 – reference D5085845 – MSA) and a Toximeter® II sampling pump (reference D6172755 – MSA). Abatement and elimination capacities were respectively calculated using equations A and B. % abatement = ((C(inlet) – C(outlet))/C(inlet)) . 100
(Equation A)
EC = ((C(inlet) – C(outlet))/C(inlet)).Q)/V
(Equation B)
EC: Elimination capacities (g/m3filter/h) Q: Flow at the outlet of the biofilter (m3air/h)
C: Ammonia concentration (g/m3air) V: Volume of the biofilter (m3filter)
2.2.2 VOCS CONCENTRATION MEASUREMENT VOCs samples were collected in TedlarTM SKC® bags. Quantification of total VOCs was performed within 24 hours using a Flame Ionisation Detector (APHA 360, HORIBA). The results were expressed in mg of ppmv equivalent CH4. 2.3 MICROBIOLOGY 2.3.1 CARRIER MEDIA SAMPLING AND BIOFILM DETACHMENT METHOD As stated by Khammar et al. (2005) and Malhautier et al. (2005), a stratification of bacterial population occurs along the biofilter depth. As a result, it was decided to collect samples at the same point during the time of the study. For each biofilter A and B, 1 L of carrier media samples was collected at 0.30 m depth under the surface. Samples were conserved at 4°C in a sterile bottle and no aqueous phase was added to avoid biofilm desorption during transport. In order to optimize the biofilm detachment technique from both biofilm carrier media, three methods were tested: (i) sonication, (ii) mechanical agitation and (iii) mechanical agitation with glass beads. Influences of parameters such as time (5, 10, 15 and 20 minutes), chemical dispersion agents (physiologic serum, Ringer Solution and PBS (Khammar et al., 2004) and mass ratio glass beads/biofilm carrier media (1/
542
M. JOVIC, L. CABROL, F. DUCRAY, R. GAGNEUX AND A. S. LEPEUPLE
2, 1 and 2) were investigated. For biomass carrier media A, mass media and aqueous phase volume were respectively equal to 100g and 500mL. For biomass carrier media B, these values were equal to 50g and 400mL. 2.3.2 DIRECT CELLS COUNTS Total and viable bacteria were enumerated by epifluorescence microscopy (Olympus) using staining with DAPI (reference D9542 - Sigma) and Total Viable Count (TVC) (AES-Chemunex). AOB and NOB bacteria quantification was carried out using FISH method. In order to evaluate data variability (RDS), three microfilters were examined for each TVC, DAPI and FISH samples. 2.3.2.1 DAPI Bacterial suspensions were incubated in the dark with in a 2.5 μg/mL final concentration DAPI solution during 30 minutes. Stained bacteria were recovered by microfiltration through a 0.2 μm porosity membrane filter (GTBP - Millipore). Microfilters were then mounted on microscope slides in mounting medium (Olympus) and examined using an epifluorescence microscope (Olympus) equipped with a blue excitation filter (330 – 385 nm) and a 420 nm barrier filter. 2.3.2.2 TVC In order to enumerate viable cells in samples, Esterase activity measurements, were undertaken using TVC kit (Doc 200–D0510–07 – AES-Chemunex). After cells staining, microfilters were mounted on microscope slides in a mounting medium (Olympus) and observed using an epifluorescence microscope (Olympus) equipped with a 470 – 495 nm excitation filter and a 519 nm barrier filter. 2.3.2.3 Cell viability Cellular viability rate (R) was evaluated by calculating DAPI/TVC ratio using equation C: R = (Ct / Cv).100 R : Cellular viability rate (%)
(Equation C) Ct, Cv: bacteria concentrations (cells/g)
2.3.2.4 AOB and NOB FISH FISH method is based on the capacity of labelled DNA probes to hybridize with a specific sequence of rRNA. Since its first application on biological reactor (Amman et al., 1990a; Amman et al., 1990b), this technique has gained an increasing interest and is now widely applied to study microbial ecology of biological filters (Friedrich et al., 1999; Moter et al., 2000).
MONITORING AND CHARACTERIZATION OF BACTERIAL POPULATIONS
543
In this study, FISH method was employed to investigate the colonization of biofilter carrier media by nitrifying bacteria (AOB and NOB). Three DNA probes were selected, two for NOB bacteria (NIT 3 and Ntspa 712) and one for AOB bacteria (Nso 1225) (Gieseke et al., 2001). As described by Gieseke et al. (2001), the control of DNA probes penetration and accessibility in cells was tested using Eubacteria probe EUB 338. This control also insured that rRNA quantity was sufficient for FISH analyses (Amman et al., 1990b). FISH analyses were performed as described by Gieseke et al. (2001) and Amman et al. (1990b). Quantification of nitrifying bacteria on carrier media were realised measuring total surface areas of AOB and NOB clusters (Cell, Olympus, Germany). 2.3.3 ANALYSIS OF TOTAL DNA BY PCR-SSCP SSCP is a separation technique based on single strand DNA secondary structure. As stated by Amman et al. (1995), rDNA is an ideal phylogenic marker. As a consequence, PCR-SSCP on the 16 S rDNA V3 region can be applied for the determination of molecular fingerprints. These fingerprints reflect ecosystem microbial richness and diversity (Khammar et al., 2004). 16 S rDNA PCR-SSCP analyses were realised by the Environmental Technology Institute (ITE, Narbonne, France). The protocol used for DNA extraction was the one designed by Godon et al. (1997). SSCP analyses were conducted as describe by Delbès et al. (2001). Similarity degrees between 16 S rDNA PCR-SSCP profiles (similarity matrix) were calculated using the Pearson coefficient (Dijkshoorn et al., 2001).
3 RESULTS AND DISCUSSION 3.1 OPTIMIZATION OF BIOFILM DESORPTION METHOD Statistical analyses based on DAPI numerations indicated that parameters such as time, chemical dispersion agents or the presence of glass beads did not have any impact on biofilm’s detachment (ANOVA, p > 0.05). As a result, a comparison of mechanical agitation and sonication was realised using an application time of 5 minutes and physiological serum as a chemical dispersion agent. The coupling of mechanical agitation and sonication was also evaluated. Statistical analyses of DAPI enumerations indicated that mechanical agitation desorbed significantly more biofilm than sonication (Student test, p < 0.05). Nevertheless, as shown in Figure 1, coupling sonication and mechanical agitation did not have a significant effect on biofilm’s desorption efficiency (Student test, p > 0.05). Finally, the desorption method selected for biofilter monitoring was a 5 minutes mechanical agitation in physiologic serum (8% (p/v) NaCl solution) using a Vortex (Scientific Industries) on position 10.
544
M. JOVIC, L. CABROL, F. DUCRAY, R. GAGNEUX AND A. S. LEPEUPLE
Figure 1. Comparison of three methods for biofilm extraction. The three methods compared were: (i) mechanical agitation (Vortex), (ii) sonication and (iii) coupling sonication/Vortex. Tests were realised on biofilm’s carrier media of biofilter B.
3.2 MONITORING OF POLLUTANT REMOVAL The monitoring of chemical and physical parameters confirmed that biofilters A and B were fed by VOCs and ammonia. For ammonia inlet concentrations varied from 20 to 40 mg/m3. During the time of the study, determination of the abatement of ammonia did not show significant differences for biofilters A and B. As shown in Figure 2, ammonia was fully eliminated by biofilters A and B after 90 days of functioning. Moreover EC for biofilters A and B was close to 2 g/m3/h.
Figure 2. Graphical representation of NH3 abatement in biofilters A and B during the time of the study. NB: for technical reasons the monitoring of VOCs was not performed. However, inlet concentrations were measured at the beginning of the study. They were close to 5 ppmv CH4 equivalent.
MONITORING AND CHARACTERIZATION OF BACTERIAL POPULATIONS
545
3.3 MONITORING OF BIOFILTERS MICROBIOLOGICAL PARAMETERS 3.3.1 TOTAL AND VIABLE BIOMASS MONITORING As described in Figure 3, densities of microorganisms observed on biofilters A and B were in agreement with those reported on other biofilters (Khammar et al., 2004). Statistical analyses of total biomass for biofilters A and B indicated a significant evolution of colonization of biofilm carrier media with time (ANOVA, p < 0.05). Nevertheless, as presented in Figure 3, modifications of the biomass concentrations were not relied with the start up of biofilters A and B. It was conclude that these changes were due to sampling fluctuations or biofilm washing.
Figure 3. Representation of viability rates® and colonization densities (DAPI, TVC) of carrier media of biofilters A and B.
546
M. JOVIC, L. CABROL, F. DUCRAY, R. GAGNEUX AND A. S. LEPEUPLE
As reported in Figure 3, concentration of viable biomass in both biofilter A and B presented significant modifications during the experiment (Kruskall-Wallis test, p < 0.05 (biofilter A) and ANOVA, p< 0.05 (biofiltrer B)). For biofilter B, modifications of the viable biomass concentration were attributed to sampling fluctuations or biofilm washing. However, for biofilter A, a 1-log increase in TVC enumeration appeared during the first week. This increase was relied to nutriments provision by gaseous effluent. The mean of TVC enumerations was calculated during the entire period of the study. It appeared that concentrations of viable biomass in both biofilters A and B were not significantly different (Student test, p > 0.05). Nevertheless, cellular viability rate was higher in biofilter B than in biofilter A. As indicated in Figure 3, cellular viability rate was ranged between 5 to 40 % for biofilter A and between 22 to 100 % for biofilter B. Also, at t = 0, it appeared that viable biomass concentration in carrier media B was significantly greater than in carrier medium A. It is important to point out that, as the carrier medium could be the only source of nutriments in abnormal working conditions, the biomass would better resist to stops of feeding in a biofilter containing medium B. 3.3.2 AOB AND NOB POPULATION MONITORING During the time of the study, the monitoring of nitrifying bacteria using FISH method did not show any presence of AOB and NOB in biofilters A and B. Positive results obtained with Eubacteria probe (EUB 338) underlined that there was not any problem of FISH probes accessibility in cells. Taking into account these results and ammonia abatement efficiency, the following hypothesis were assumed: (i) AOB and NOB were present in other strates of the biofilters (Khammar et al., 2004) (ii) ammonia was metabolized by others nytrifying bacteria genus, (iii) ammonia was not biodegradated but was absorbed by biofilm carrier media and (iv) ammonia was absorbed by biofilters scrubbing solution. 3.3.3 ANALYSIS OF TOTAL DNA BY 16S RDNA PCR-SSCP Analysis of 16S rDNA SSCP profiles based on similarity matrices allowed the comparison of microbial population of biofilters A and B. As shown in Figure 4, it appeared that microbial populations of biofilters A and B were modified in comparison to their initial state. The results also indicated that the most important variations occurred for biofilter A. As indicated in Figure 5, comparison of SSCP profiles with time for biofilter A and B established that microbial populations of both biofilters A and B converged and became nearly identical after 120 days of functioning. At this time, those microbial populations were closer than at their initial state.
MONITORING AND CHARACTERIZATION OF BACTERIAL POPULATIONS
547
Figure 4. graphical representation of evolutions of biofilters A and B microbial populations compared with their initial state.
Figure 5. graphical representation of evolution in time of similarities of biofilters A and B microbial populations.
Finally, as presented in Figure 6 comparison of 16S rDNA SSCP profiles between t and t-1, revealed that microbial populations of both biofilters A and B evolved and stabilized. Figure 6 demonstrated that biofilter B stabilized faster than biofilter A. After 60 days of functionning, biofilter B was stabilized whereas this state was reached after 120 days for biofilter A. Those results were attributed to the composition of
548
M. JOVIC, L. CABROL, F. DUCRAY, R. GAGNEUX AND A. S. LEPEUPLE
biofilm carrier media. Indeed, biofilter B was packed with a carrier media containing branches coming from the composting facility. Therefore, these results demonstrated that inoculation of biofilm carrier media with an acclimated microbial population significantly reduce biofilter start up phase.
Figure 6. Graphical representation of evolution of biofilters A and B microbial populations compared with their previous state (t-1 to t).
4 CONCLUSION AND PERSPECTIVES As a conclusion, it appeared that physical and chemical monitoring of biofilters A and B did not provide helpful information to select a biofilm carrier media. Indeed, during the time of the study, abatement efficiencies observed on the two biofilters were not significantly different. Microbiological parameters have provided valuable information. Indeed, it was established, by the quantification of viable biomass, that carrier media B allowed a greater supply in nutriments for the biomass than carrier media A. Dynamic analysis of microbial populations confirmed that important modifications of microbial populations occurred during the start up phases of the biofilters. It was also demonstrated that those modifications were dependant of initial states. Therefore, it appeared that biofilter’s stabilization phase could be reduced using an appropriate inoculum. Finally, the microbial approach proved its interest to better understand phenomena occurring during the start up phase of biofilters. It also demonstrated the interest to use molecular tools to caracterize a biofilm carrier media. Moreover, carrier
MONITORING AND CHARACTERIZATION OF BACTERIAL POPULATIONS
549
media B showed benefits compared to carrier media A, as it could significantly reduce the start up phase and provide a greater supply in nutriment. This last point could have a significative impact on biofilter’s stability in the case of feeding stops.
REFERENCES Aizpuru, A., Malhautier, L., Roux, J.C. and Fanlo J.L. (2001) Biofiltration of a mixture of volatile organic emissions. Air & Waste Manage. Assoc. 51: 1662-1670. Amann, R., Krumholz, L.and Stahl, D.A. (1990a) Fluorescent oligonucleotide probing of whole cells for determinative, phylogenetic and environmental studies in microbiology. J. Bacteriol. 172: 762-770. Amann, R., Binder, B.J., Olson, R.J., Chisholm, M.S.W., Devereux, R. and Stahl, D.A. (1990 b) Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analysing mixed microbial populations. Appl. Environ. Microbiol. 56: 1919-1925. Delbes, Leclerc, Zumstein, Godon and Moletta. (2001) A molecular method to study population and activity dynamics in an anaerobic digestor. Water Sci. Technol. 43(1): 51-57. Dijkshoorn, L., Towner, K.J. and Struelens, M. (2001) New approaches for the generation and analysis of microbial typing data. Elsevier. Friedrich, U., Naismith, M., Altendorf, K. and Lipski A. (1999) Community analysis of biofilters using fluorescence in situ hybridization including a new probe for the Xanthomonas branch of the class Proteobacteria. Appl. Environ. Microbiol. 65(8): 3547-3554. Gieseke, A., Purkhold, U., Wagner, M., Amman, R. and Schramm, A. (2001) Community structure and activity dynamics of nitrifying bacteria in a phosphate-removing biofilm. Appl. Environ. Microbiol. 67(3): 1351-1362. Godon, J., Zumstein, E., Dabert, P., Habouzit, F. and Moletta, R. (1997) Microbial 16S rDNA diversity in an anerobic digester. Appl. Environ. Microbiol. 63(7): 2802-2813. Khammar, N., Malhautier, L., Degrange, V., Lensi, R. and Fanlo, J.L. (2004) Evaluation of dispersion methods for enumeration of microorganisms from peat and activated carbon biofilters treating volatile organic compounds. Chemosphere 54: 243-254. Khammar, N., Malhautier, L., Degrange, V., Lensi, R., Godon, J.J. and Fanlo, J.L (2005) Link between spatial structure of microbial communities and degradation of a complex mixture of volatile organic compounds in peat biofilters. J. Appl. Microbiol. 98: 476-490. Khammar, N., Malhautier, L., Bayle, S. and Fanlo, J.L. (2005) Biofiltration of volatile organic compounds. Appl. Environ. Microbiol. 68: 16-22. Moter, A. and Gobel, U. (2000) Fluorescence in situ hybridization (FISH) for direct visualization of microorganisms. J. Microbiol. Met. 41: 85-112.
This page intentionally left blank
551
Biofilter response to biomass reactivation for VOC treatment ANA ELÍAS, ASTRID BARONA, GORKA GALLASTEGI, MIKEL LARRAÑAGA AND MARÍA FERNÁNDEZ Department of Chemical and Environmental Engineering, Faculty of Engineering, University of the Basque Country, Alda Urquijo s/n. 48013 Bilbao, Spain
ABSTRACT This research has undertaken a comparative study on using a fresh activated sludge or a refrigerated/ reactivated sludge as active biomass source for biofiltration purposes. A sludge sample was initially selected based on the ratio between volatile solid content and total solid content before and after refrigeration at 6 oC for 90 days. The degradation rate of the activated sample for three addition doses of toluene was established before and after refrigeration. The same procedure was also carried out for ethylbenzene and p-xylene after refrigeration/reactivation. Surprisingly, the degradation rate for toluene was higher after refrigeration and the results were very similar for an addition of 2 and 8 μL. Subsequently, one biofilter was inoculated with the activated sample and another with the reactivated sample, and both were fed with toluene ranging from 2.6 to 26.2 g toluene m-3 h-1. Concerning the elimination capacity of both biofilters, no relevant differences were found. It was concluded that the active biomass degrading toluene was not affected by refrigeration, in spite of the fact that the SV/ST ratio decreased after the storage period. The elimination capacity of the other two biofilters (ethylbenzene and p-xylene) was highly influenced by the gas flow rate.
1 INTRODUCTION Environmental regulations for pollution control are frequently enacted before «suitable» (affordable, effective and environmentally friendly) technologies have been fully developed. Amongst biotechnologies, biofiltration is a seemingly simple system whose effectiveness relies on the optimization of several operating parameters and the selection of a suitable packing material and degrading biomass.
551
552
A. ELÍAS, A. BARONA, G. GALLASTEGI, M. LARRAÑAGA AND M. FERNÁNDEZ
The biomass responsible for the degradation of the gaseous contaminant fed into the biofilter can be supplied by the support material itself (Barona et a.l, 2004), purchased from trade catalogues (Christen et al., 2002), isolated from other active bioreactors (Estévez et al., 2004) or even collected from locations contaminated with the target pollutant (González-Sánchez and Revah, 2007). Prior to the inoculation itself, the proper selection, storage, acclimation and activation of the microorganisms is crucial to ensure a long-lasting operation of the bioreactors. In fact, Prado et al. (2005) proved that the previous biomass concentration and biomass adaptation of the inoculum dramatically affected the start-up and performance of conventional biofilters treating methanol during the first stages of operation. Likewise, simple analytical techniques for biomass growth detection are needed. Several methods, such as extra-cellular enzymatic activity (Laurent and Servais, 1995) or electron transportation activity (Fontvieille and Moul, 1985), were developed by microbiologists to estimate biomass activity. Among other methods, respirometry, optical density at 600 nm, plate counting methods and 4’6-diamidino-2-phenylindole (DAPI) staining have also been used in literature (Hwang et al., 2003; Álvarez-Hornos et al., 2005; Kim and Jaffé, 2007). Other simpler analyses, such as carbon balance and volatile suspended solid content, are also practical tools for assessing biomass concentration and adaptation, although they have obvious disadvantages, such as no discrimination between living or dead biomass or no accounting for changes in microorganism physiology. Bearing in mind that biofilter controlling operators require simple and quick techniques for start-up and everyday operation, the objective of this study is to ascertain the influence of using a previously refrigerated and reactivated sample (a stored sample) as inoculum to set-up several biofilters. Likewise, the relevance of certain simple parameters for achieving biomass activity will be studied. The pollutants to be treated in the biofilters were toluene, ethylbenzene and p-xylene.
2. MATERIALS AND METHODS 2.1 SOURCE OF THE MICROORGANISMS AND MEDIA COMPOSITION Three sludge samples were collected in a wastewater treatment plant (W), in a small river close to a petrochemical company (P) and near a synthetic resin-producing industry (F) in Bizkaia (Spain). After sedimentation of the solid phase of each sample for 2 hours, the liquid phase was transferred into glass bottles for further experimentation. A nutrient medium (Barona et al., 2007) was used for enriching and maintaining the cultures.
BIOFILTER RESPONSE TO BIOMASS REACTIVATION FOR VOC TREATMENT
553
2.2 BIOMASS STORAGE AND ACTIVATION About 1 L of the supernatant phase of each sample was mixed with 1 L of the nutrient solution, and the total volume was transferred into 3 litre vessels that were continuously fed with toluene (inlet concentration ranging from 50 to 100 ppmv for 1000 hours at a gas flow rate of 2.5 L min-1). When this procedure concluded, the activated samples were stored at 6 oC for 90 days. The samples were subsequently warmed to room temperature (20 oC) and the reactivation procedure was then repeated for a further 1000 hours. 2.3 BIOMASS RECOVERY IN BATCH ASSAYS Two different parameters were used to compare biomass recovery before and after refrigeration and storage. The first parameter was the ratio between volatile solid content and total solid content (VS/TS), and it was basically used to select the best sample for subsequent experiments. After the preliminary selection of one of the samples based on VS/TS data, the second step was to determine degradation rates in order to ascertain the influence of refrigeration and storage lag on the selected sample activity. The originally activated sample was transferred into a hermetic vessel and 2 μL of toluene were added. The removal of the contaminant was monitored over time until the complete disappearance of the pollutant. Likewise, two consecutive doses of 4 and 8 μL of toluene were also added to determine the degradation rate of the biomass prior to inoculation. The same procedure was repeated with the reactivated sample (after refrigeration), and in this case, toluene, ethylbenzene and xylene were individually added in the respective hermetic vessels. After the addition of 2, 4 and 8 μL of each pollutant, the degradation rate was monitored over time until the complete disappearance of the compound. 2.4 BIOFILTER SETUP The packing material used for filling four biofilters was made up of composted pig manure and sawdust and has already been used in previous work (Elías et al., 2002). The outline of the pilot plant for each of the four biofilters has already been described in detail (Moura et al., 2006). The first biofilter (biofilter 0T) was inoculated with the original activated sample (without refrigeration and for toluene degradation) and the other three were inoculated with the refrigerated/reactivated sample, although the contaminant to be degraded was toluene, ethylbenzene and p-xylene (biofilter T, biofilter E and biofilter X, respectively). The inlet loading rate for the three contaminants ranged from 2.6 to 29.1 g m-3 -1 h . Two different gas flow rates were used during experimentation; the first one was 1 L min-1 and the second one was 2 L min-1, corresponding to a residence time of 180 and 90 s, respectively. Temperature was constant at 23 oC.
554
A. ELÍAS, A. BARONA, G. GALLASTEGI, M. LARRAÑAGA AND M. FERNÁNDEZ
2.5 ANALYTICAL METHODS Toluene, ethylbenzene and p-xylene were measured in a micro gas chromatograph (microGC CP 4900) equipped with auto-sampling injection mode and a TCD detector. Operating conditions were: injector temperature, 110 oC; column/ transfer temperature 80 oC.
3. RESULTS AND DISCUSSION The original three samples were activated for 1000 hours, and after a 90-day refrigeration and storage period, they were reactivated for a further 1000 hours. The evolution of the biomass during these activation assays was monitored by measuring the changes in the ratio between the volatile solid content and total solid content (VS/ TS). The results, shown in Table 1, revealed that none of the three samples was able to fully recover initial values of the controlling parameter after tempering. In all cases, the VS/ST ratio decreased by about 20% after storage. Nevertheless, the sample from the wastewater facility (W) showed the highest VS/TS ratio before (85) and after refrigeration (65). Consequently, W sample was preliminary selected as the best inoculum for subsequent operation, although further analyses were carried out to ascertain the degradation rate before and after storage. Table 1. The VS/TS ratio for the three activated samples before and after refrigeration.
Original After 1000 hours of activation After refrigeration and 1000 hours of reactivation
W sample 28 85 65
100 x VS/TS P sample F sample 21 20 60 72 48 58
The degradation rate before and after refrigeration for the toluene activated sample is shown in Table 2. The originally activated W sample achieved a relatively constant degradation rate for the addition of 4 μL of toluene (211 ppmv h-1) and for 8 μL (222 ppmv h-1). Surprisingly, the degradation rate was higher after refrigeration and very similar for an addition of 2 and 8 μL (291 and 280 ppmv h-1, respectively). Consequently, the decrease in the VS/TS ratio after refrigeration for W sample was not considered to be a decisive result (Table 1). Thus, the ratio reached after storage was high enough for the biomass to recover activity.
BIOFILTER RESPONSE TO BIOMASS REACTIVATION FOR VOC TREATMENT
555
Table 2. Degradation rate for the originally activated and refrigerated/reactivated W sample.
Sample type Originally activated Refrigerated/reactivated Refrigerated/reactivated Refrigerated/reactivated
Compound Toluene Toluene Ethylbenzene p-Xylene
Degradation rate (ppmv h-1) Dose 2 μL 4μL 8μL 66 211 222 291 213 280 263 258 526 194 233 204
The degradation rate for ethylbenzene and p-xylene is also shown in Table 2. The results for p-xylene revealed that the biomass in the refrigerated sample was able to degrade this pollutant at a rate ranging from 194 to 233 ppmv h-1, in spite of the contaminant dose. The refrigerated sample also showed a high capacity to degrade ethylbenzene, above all for the highest dose. The degradation rate was fairly constant for the addition of 2 and 4 μL, but it doubled when 8 μL of ethylbenzene were added. As an example, the remaining amount of ethylbenzene after the addition of 4 μL is plotted in Figure 1. Those preliminary results suggested that the refrigeration of W sample did not stop the biomass degrading the three simple alkylbenzenes. Furthermore, the highest degradation rate was reached for ethylbenzene with an addition dose of 8 μL.
Figure 1. Evolution of the remaining concentration of ethylbenzene along time after an addition of 4 μL (refrigerated/reactivated sample).
556
A. ELÍAS, A. BARONA, G. GALLASTEGI, M. LARRAÑAGA AND M. FERNÁNDEZ
In order to ascertain the behaviour of W sample (originally activated and refrigerated/reactivated) as inoculum for biofiltration purposes, four biofilters were started-up. Two of them were fed with toluene (biofilters 0T and T) and the other two (biofilters E and X) were fed with ethylbenzene and p-xylene, respectively. The inoculation of biofilter 0T was carried out with the originally activated W sample and the inoculation of the others involved the refrigerated and reactivated W sample. In all cases, two flow rates of 1 and 2 L min-1 were tested. The comparison of the response of biofilters 0T and T rendered similar results and, consequently, only the data for biofilter T have been plotted in Figure 2. It was concluded that the active biomass degrading toluene was not affected by the refrigeration, in spite of the fact that the VS/TS ratio reduced after the storage and reactivation period.
Figure 2. Response of Biofilter T treating toluene (inoculation with the refrigerated/reactivated sample).
As shown in Figure 2, when the inlet loading rate (IL) ranged from 2.6 to 26.2 g toluene m-3 h-1 for a residence time of 180 s (gas flow rate of 1 L min-1), the removal efficiency was close to 100%. When the residence time was reduced by half (80 s) as a consequence of increasing the gas flow rate to 2 L min-1, inlet load higher than 18 g m-3 h-1 rendered lower results for elimination capacity (data plotted below the dashed line). When the gas flow rate was 1 L min-1 and the inlet load of ethylbenzene fed into biofilter E ranged from 4.8 to 28.5 g m-3 h-1, the biomass was able to completely
BIOFILTER RESPONSE TO BIOMASS REACTIVATION FOR VOC TREATMENT
557
degrade the contaminant. In contrast, the change to the higher gas flow rate dramatically reduced elimination capacity and biofilter response was very variable. It is remarkable that the latter result is not consistent with the data shown in Table 2, where ethylbenzene recorded the highest degradation rate. Nevertheless, this behaviour may be explained by the random attachment of biomass to the support material, which is a phenomenon for further research.
Figure 3. Response of Biofilter E treating ethylbenzene (inoculation with the refrigerated/reactivated sample).
The results obtained for the p-xylene biofilter were similar to those obtained for biofilter E, and, for brevity, are not shown.
4 CONCLUSIONS The storage of active biomass under refrigeration (not in freezing conditions) is necessary for reproducing biofilters. In this study, the influence of a 90-day refrigeration lag was studied. Initially, the volatile solid content/total solid content (VS/TS) ratio of three activated sludge samples was measured before and after refrigeration. None of the three samples was able to fully recover initial values of the controlling parameter after tempering. Nevertheless, the sample with the largest VS/ST ratio was selected for determining the degradation rate of toluene before and after refrigeration.
558
A. ELÍAS, A. BARONA, G. GALLASTEGI, M. LARRAÑAGA AND M. FERNÁNDEZ
Surprisingly, the degradation rate for toluene was higher after refrigeration. Subsequently, one biofilter was inoculated with the activated sample and another one with the refrigerated/reactivated sample, with both of them being fed with a toluene loading rate ranging from 2.6 to 26.2 g m-3 h-1. It was concluded that the active biomass degrading toluene was not affected by the refrigeration, in spite of the fact the SV/ST ratio decreased after the storage and reactivation period. The elimination capacity of the biofilters for ethylbenzene and p-xylene was highly influenced by the gas flow rate, which is possibly related to the uneven attachment of biomass to the support material in these cases.
5 ACKNOWLEDGEMENTS The authors gratefully acknowledge the financial support of the University of the Basque Country (Research group GIU05/12), and the Spanish Ministry of Science and Technology (Project CTM2006-02460 with ERDF funding).
REFERENCES Alvarez-Hornos, F.J., Gabaldón, C., Martínez-Soria, V., Marzal, P. and Penya-Roja, J.M. (2005) Biodegradation of ethyl acetate and toluene mixtures by a peat biofilter: Respirometry monitoring and dynamics of living and dead bacterial cells. Proceedings of the International Congress on Biotechniques for Air Pollution Control, La Coruña, Spain, pp. 413-420. Barona, A., Elías, A., Arias, R., Cano, I. and González, R. (2004) Biofilter response to gradual and sudden variations in operating conditions. Biochem. Eng. J. 22: 25-31. Barona, A., Elías, A., Cano, I., Uriarte, A. and Artetxe, J. (2007) Additional determinations in a biofiltration system for toluene: Adsorption and partition in the nutrient solution. Chem. and Biochem. Eng. Q. CABEQ 763 (in press). Christen, P., Domenech, F., Michelena, G., Auria, R. and Revah, S. (2002) Biofiltration of volatile ethanol using sugar cane bagasse inoculated with Candida utilis. J. Hazard. Mat. B89: 253-265. Elías, A., Barona, A., Arreguy, A., Ríos, J., Aranguiz, I. and Peñas, J. (2002) Evaluation of a packing material for the biodegradation of H2S and product analysis. Proc. Biochem. 37(8): 812-820. Estévez, E., Veiga, M.C. and Kennes, C. (2004) Fungal biodegradation of toluene in gas-phase biofilters. In: Verstraete, W. (Ed.), Proceedings of the 5 th European Symposium on Environmental Biotechnology. Oostende, Belgium, 25-28 April, pp. 337-340.
BIOFILTER RESPONSE TO BIOMASS REACTIVATION FOR VOC TREATMENT
559
Fontvieille, D. and Moul, A. (1985) Dénombrement et mesure d‘activité, Bactériologie des milieux aquatiques , Point sur l‘épuration et le traitement des effluents (eau, air): Lavoisier Ed.: 79-131. González-Sánchez, A. and Revah, S. (2007) The effect of chemical oxidation on the biological sulphide oxidation by an alkaliphilic sulfoxidizing bacterial consortium. Enz. Microb. Technol. 40: 292-298. Hwang, S.C.J., Lee, C.M., Lee, H.C. and Fang Pua, H. (2003) Biofiltration of waste gases containing both ethyl acetate and toluene using different combinations of bacterial cultures. J. Biotechnol. 105: 83-94. Kim, H.S. and Jaffé, P.R. (2007) Spatial distribution and physiological state of bacteria in a sand column experiment during the biodegradation of toluene. Wat. Res. 41: 2089-2100. Laurent, P. and Servais, P. (1995) Fixed bacterial biomass estimated by potential exoproteolytic activity. Can. J. Microbiol. 41: 749-752. Moura, M. (2006) Sistemática de crecimiento y seguimiento de biomasa para sistemas de biofiltración. Project 2006. School of Engineering, University of the Basque Country. Prado, O.J., Veiga, M.C. and Kennes, C. (2005) Treatment of gas-phase methanol in conventional biofilters packed with lava rock. Wat. Res. 39: 2385-2393.
This page intentionally left blank
561
A comparative study of the characteristics and physical behaviour of different packing materials commonly used in biofiltration ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL1 AND JAVIER LAFUENTE1 Department of Mining Engineering and Natural Resources, Universitat Politècnica de Catalunya, Manresa, Spain 1 Department of Chemical Engineering, ETSE, Universitat Autònoma de Barcelona, Barcelona, Spain
ABSTRACT In this study, the characteristics and physical behaviour of 8 different packing materials were compared. The materials were selected according to previous works in the field of biofiltration including organic and inorganic or synthetic materials. Results pre-selected those more acceptable support materials for the main function they have to perform in the biological system: high surface contact, rugosity to immobilize the biomass, low pressure drop, nutrients supply, water retentivity or a commitment among them. Otherwise, pressure drop have been described by means of the respective mathematic expressions in order to include phenomena in the classical biofiltration models.
1 INTRODUCTION Biological treatment have become and effective and economical alternative to the traditional systems of gas treatment. However, several packing materials have been used in biofiltration without a global agreement about which one is the most adequate to immobilize biomass. Carrier materials may be organic, natural inorganic, or entirely synthetic. The nature of the packing material is a fundamental factor for successful application of biofilters because it affects the frequency at which the medium is replaced and other key factors such as bacterial activity and pressure drop across the bioreactor (Devinny et al., 1999).
561
562
ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE
Moreover, particles vary in size, which affects important medium characteristics such as resistance to air flow and effective biofilm surface area. If the size of the bed pellets is too small provides for large specific surface areas, available for essential mass exchange, but it also creates resistance to gas flow, while if it is too large, it favours gaseous flows but reduces the number of potential sites for the microbial activity (Delhoménie et al., 2002). Adu and Otten (1996) have reported that particle size is a parameter more influential to the performance than the gas flow rate. Among the naturals carriers reported, compost, peat, soil and the wood derivatives are the most extensively used while GAC, perlite, glass beads, ceramic rings, polyurethane foam, polystyrene and vermiculite are some of the several synthetic or inert carriers which have been studied (Kennes and Veiga, 2001). Specific surface area, porosity, density, water retention capacity and the nutrients availability are some of the most important characteristics of the filter media (Janni et al., 2001). In this work, a comprehensive study of physical parameters for different packing materials commonly used in biofiltration has been performed. Pressure drop was also determined for each packing material to determinate the inherent economical cost to flow the air through the bed. To this aim, pressure drop was evaluated in each case depending on the flow rate, the bed porosity and the water content circulating through the material media in countercurrent flow. Pressure drops have been described by means of mathematical expressions relating the effects of the studied factors in order to include this parameter in classical biofiltration models.
2 MATERIALS AND METHODS 2.1 EXPERIMENTAL SETUP Pressure drop assessment experiments were carried out using a lab-scale plant consisting of a PVC column with an inner diameter of 4.6 cm and a height of 70 cm (Figure 1). The compressed air was conducted by 2 different circuits. In the former, the air stream was passed through a water column in order to increase the relative humidity and in the latter, the air stream arrived completely dry to the fixed bed. The inlet air pressure and the flow rate were controlled and measured by means of a pressure regulator (Norgren Excelon) and a flowmeter (Tecfluid 2100) respectively. Throughout this study, the gaseous stream was supplied in up-flow mode. Tap water was sprinkled continuously at the top of the fixed bed be means of a peristaltic pump (Magdos LT10) and the water content was measured by an optical level sensor. Pressure drop was determined by means of two digital differential pressure meter used according to the limit detection and precision (Testo 512-20hPa and Testo 506-200 hPa).
A COMPARATIVE STUDY OF THE CHARACTERISTICS AND PHYSICAL BEHAVIOUR
563
Figure 1. Schematic of the lab-scale setup. 1: humidification column; 2: fixed bed for pressure drop study; 3: fixed bed for water retentivity study; 4: flow meters; 5: pressure regulator; 6: peristaltic pump; 7: digital differential pressure meter.
2.2 PACKING MATERIALS A total of 8 common packing materials used in biofiltration were studied and compared by determining their main physico-schemical properties. Organic packing materials analysed were coconut fibre, pine leaves, peat and compost from sludge of a waste-water treatment plant. The inorganic or synthetic packing materials studied were polyurethane foam, lignite from Mequinenza mines (Spain), lava rock and an advanced material based on a thin layer of compost over a clay pellet. 2.3 ANALYTICAL METHODS Characterization of packing materials were carried out according to standard methods (APHA, 1980; ASTM, 1990; TMECC, 2002). The following properties were compared in each case: specific surface area, elementary analysis, extractable phosphor content, organic matter, humidity, water holding capacity, retentivity, ph, conductivity and buffer capacity of the leachate. Specific surface area and material density were determined by the BET technique in a Micromeritics, model Tristar 3000, apparatus. Elementary analysis was realised by combustion in standard conditions using sulfanilamida as standard (EA-1108 ThermoFisher Scientific). Extractable phosphor was determined by the technique of ICP in a multichannel analyser in standard
564
ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE
conditions (Thermo Jarell-Ash model 61E Polyscan) using Baker Instra as digester of the sample. Surface rugosity was observed by means of a Scanning Electron Microscope (Jeol JSM-840). Humidity and organic matter were determined by drying and combustion standard procedures. Water holding capacity was measured keeping the material wet sparkling constantly tap water for 100 minutes and determining the weight changes. Water retentivity was measured by keeping wet material in constant contact with dry air flow circulating through the bed and measuring the loss of weight of the bed. Conductivity, pH and buffer capacity was determined for the materials leachate submerging them in water for 1 hour in controlled conditions of temperature and agitation.
3 RESULTS AND DISCUSSION 3.1 CHARACTERIZATION OF PACKING MATERIALS High nutrient, phosphorous, potassium and sulphate contents, as well as trace elements, are required for the establishment of a dense process culture. Regarding to the elementary composition of organic packing materials (Table 1), it is shown that the compost is the material with the highest content in nitrogen and phosphorus (2.7 % and 14.500 ppm, respectively. It must be pointed out that immature coal (lignite) studied showed a significant concentration of sulphur (8,8 %) which is related to the quality of the material. Also, presence of sulphur has been detected in compost as well. On the other hand, phosphorous concentration in lava rock (1800 ppm) is higher than expected probably due to the pre-treatment of this material to garden applications. Among the organic material analysed, coconut fibre and pine leaves present the highest organic matter content (higher than 85% by weight). The organic matter detected in coal (next to 80%) is a reflex of the immature nature of this material. In general, it is desirable to have media with a high water-holding capacity. Organic media are 40 to 80% water (by weight) when they are saturated (Devinny et al., 2002). Packing materials studied keep a water holding capacity inside the typical interval, being in the higher values for coconut fibre, pine leaves and peat (Table 2). The humidity of the materials is similar in all the studied cases but there are appreciable differences in water retentivity. Regarding to the specific surface, coal is the material with the highest value (6 m2 m-3), while compost is the highest among the organic materials (2,8 m2 m-3).
565
A COMPARATIVE STUDY OF THE CHARACTERISTICS AND PHYSICAL BEHAVIOUR
Table 1. Elementary composition of packing materials.
Coconut fiber Pine leaves Peat Compost Advanced magterial Lava rock Coal
Nitrogen (%) 1,17 0,56 1,26 2,68 0,34 0,00 0,85
Carbon (%) 45,05 45,18 21,99 33,86 2,45 0,40 44,37
Hydrogen (%) 6,18 6,10 2,56 4,63 0,18 0,00 4,06
Sulphur (%) 0,12 0,05 0,15 0,63 0,19 0,00 8,81
Phosphorus (ppm) 256 191 455 14487 1259 1821 98
Organic matter (%) 91,62 86,71 66,23 53,56 2,57 0,63 79,69
Table 2. Physical characteristics of packing materials.
Coconut fiber Pine leaves Peat Compost Advanced material Lava rock Coal Polyeurethane foam
Surface area (m2 m-3)
Humidity (%)
Water holding (g·g-1)
1,68 0,50 1,43 2,82 0,76 0,62 5,99 0,02
6,62 7,79 6,97 7,83 37,62 0,06 4,85 –
3,90 1,51 1,80 0,68 0,58 0,18 0,28 –
Water Conductivity pH retentivity (μS) (% dia-1) 192,24 422,78 66,38 57,89 41,90 23,33 41,62 416,45
315 216 338 470 226 33 205 –
5,93 6,90 5,13 7,24 5,72 7,21 6,51 –
Buffer capacity (ml So42-·l-1) 33 120 20 128 13 33 45 –
Packing materials studied showed a pH close to the neutrality or slightly acid (pH ~ 5 for peat) and a buffer capacity inferior to 150 ml SO42- l-1 in all the cases. Leachate conductivity of the materials was similar among them (excepting lava rock), being 470 μS the highest value determined in compost. The surface rugosity of the materials has been observed and compared by means of Scanning Electron Microscopy. As a sample of organic materials, coconut fibre shows an important surface rugosity which could aim to fix the microorganisms to the surface (Figure 2). Conversely, polyurethane foam shows the opposite situation where the surface observation at 1000 magnifications shows a completely flat surface.
566
a)
ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE
b)
Figure 2. Microscopic observation of the rugosity of a) coconut fibre at different magnifications by SEM (x30, x1000) and b) polyurethane foam (x30, x1000).
3.2 PARAMETERS INFLUENCE IN PRESSURE DROP Pressure drop tests were carried out at 7 different flow rates, 5 different water contents and 3 different bed porosities. Flow rates were selected in the range to obtain empty bed residence times commonly used in biofiltration (from 5 to 40 seconds). Water content circulating through the bed was regulated by means of the peristaltic pump avoiding flooding episodes. Porosity was selected through different particle size or different degrees of compactation depending on the materials as for instance, coconut fibre or pine leaves. Results were represented in surface plots to observe simultaneously the parameters influence pressure drop. Figure 3 shows the effect for coconut fibre and compost as examples of organic packing materials behaviour. Regarding water content, the influence is very similar for both materials in opposition to empty bed porosity influence where results do not show significant differences. Figure 4 shows an example of the behaviour of non-organic materials, concretely polyurethane foam and the advanced material. Polyurethane foam showed more important pressure drop in the range of study, presenting significantly differences for the different bed porosities tested. Water content is a parameter less influencing in comparison to organic samples. Drop pressure in advanced material, in the only possible porosity allowed by its shape and structure, shows a strong dependence on water content, being more important at high flow rates.
A COMPARATIVE STUDY OF THE CHARACTERISTICS AND PHYSICAL BEHAVIOUR
Figure 3. Influence of operational parameters in drop pressure for coconut fibre (a) and compost (b).
567
568
ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE
Figure 4. Influence of operational parameters in drop pressure for polyurethane foam (a) and advanced material (b).
A COMPARATIVE STUDY OF THE CHARACTERISTICS AND PHYSICAL BEHAVIOUR
569
3.3 MATHEMATICAL EXPRESSIONS IN DROP PRESSURE Drop pressure in a fixed bed has been described through several semiempirical mathematical expressions. In most of works, the pressure drop is described by the well-known Ergun equation (Ergun, 1952), which may be written as (Eq. 1):
(1)
Where: ΔP is the pressure drop in Pa; H is the height of the fixed bed in m; μ is the viscosity of the air in Pa s; ν0 is the superficial velocity in m s-1; ε is the porosity of the bed; dp is the equivalent spherical diameter of the particle in m; a is the first parameter of the Ergun equation, b is the second parameter of the Ergun equation. Parameters a and b are related to the friction factor. In addition the expression term related to parameter a is significant for flow under very viscous conditions while the parameter b term is only significant when viscous effects are not as important as inertia. Some authors have fitted satisfactorily experimental data to a modified Ergun equation adapting the coefficients of the expression be means of a correction factor (Delhoménie et al., 2003). Other authors have used a specific relation due to the heterogeneity of the material and the difficulty to model pressure drop with the classical Ergun equation (Comiti and Renaud, 1989). In this study, parameters a and b from Ergun equation have been fitted as a function of the material, the porosity and the water content. Pressure drop (ΔP/H) as a function of the empty bed velocity has been fitted by a lineal regression. In all cases the correlation coefficient R2 was superior to 0,990 indicating the correct linearity between operational parameters and pressure drop. This experimental study incorporated the effect of water content in pressure drop although this parameter is not present in Ergun equation. For this reason, parameter a and b were fitted as a function of water content in the bed in order to find a relationship which describes this effect in the pressure drop estimation. Table 3 shows the final results of this systematic study in order to compare the water content effect in each material. Water content for compost, lava rock and the advanced material biofilter showed the strongest effect in parameter a. In the case of parameter b the dependence on water content was markedly lower. Thus, it is possible to express a modified Ergun equation incorporating the water content effect in the pressure drop predictions for some packing materials. These results may be useful to incorporate pressure drop phenomena in classical biofilter models.
570
ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE
Table 3. Ergun equation parameters as a function of water content in biofilters.
Compost
Coconut fibre
Lava rock
Immature coal
Pine leaves
Advanced material
0,70 0,76 0,79 0,94 0,96 0,99 0,73 0,76 0,77 0,58 0,63 0,64 0,91 0,92 0,96 0,65
a ordinate 12,634 132,370 432,090 0,626 2,145 9,130 75,398 115,150 234,900 14,618 35,182 100,830 1,885 2,524 9,378 12,342
a slope -0,069 0,720 4,077 0,004 0,019 0,750 0,340 0,384 2,276 0,212 0,387 0,270 0,014 0,016 0,059 1,458
b ordinate 0,250 0,595 1,199 0,062 0,090 0,333 0,512 0,531 0,766 0,176 0,248 0,466 1,146 2,517 11,196 0,160
b slope 0,003 0,009 0,004 0,001 0,000 -0,011 0,005 0,001 0,002 0,001 0,002 0,007 0,058 0,144 0,000 0,009
4 CONCLUSIONS Commonly used packing materials in biofiltration have been characterized and compared for a better knowledge of their advantages and drawbacks. Coconut fibre, pine leaves, peat, compost, polyurethane foam, coal, lava rock and an advanced material have been studied. Organic materials, especially compost and coconut fibre, are suitable to release extra inorganic nutrients. Moreover, these materials are able to keep water content at optimal levels for microorganisms and show the highest specific surface. Surface observation by Scanning Electron Microscope shows a better condition to fix the biomass in organic materials. On the contrary, inorganic or synthetic materials offers higher contact surface and produce cleaner drainage water. Otherwise, pressure drop have been determined for each packing materials as a function of flow rate, water content and bed porosity in order to represent the several effects simultaneously and obtain a mathematical expression to include phenomena in classical biofilter models. A water content dependence has been found through a modified Ergun equation for several packing materials.
A COMPARATIVE STUDY OF THE CHARACTERISTICS AND PHYSICAL BEHAVIOUR
571
5 ACKNOWLEDGEMENTS This work was supported by the Spanish CICYT project CTQ2006 – 14997-C02-02. A.D. Dorado received a predoctoral fellowship from the M.E.C. (Ministerio de Educación y Ciencia).
REFERENCES Adu, B. and Otten, L. (1996) Modelling the biofiltration characteristics of volatile compounds. Proceedings of the 89th annual meeting and exhibition of the air & waste management association, June 23-26, 1996, Nashville, TN. Pittsburgh, PA: Air & Waste Management Association. APHA (American Public Health Association). (1980) Standard methods for the examination of water and wastewater. ASTM (Standard Test Method). (1990) Standard methods for volume weights, water holding, and air capacity of water saturated peat materials. In (annual book of ASTM Standards) Vol. 04.08. Bohn, H.L. (1996) Biofilter media. Proceedings of the 89th annual meeting and exhibition of the air & waste management association, June 23-26, 1996, Nashville, TN. Pittsburgh, PA: Air & Waste Management Association. Bohn, H.L. and Bohn, K.H. (1999) Moisture in biofilters. Environ. Prog. 18: 156-161. Comiti, J. and Renaud, M. (1989) A new model for determining mean structure parameters of fixed beds from pressure drop measurements: application to beds packed with parallelepipedal particles. Chem. Eng. Sci. 44: 1539-1545. Delhomenie, M.C., Bibeau, L., Gendron, J., Brzezinski, R. and Heitz, M. (2003) A study of clogging in a biofilter treating toluene vapors. Chem. Eng. J. 94: 211-222. Delhomenie, M.C., Bibeau, L. and Heitz, M. (2002) A study of the impact of particle size and adsorption phenomena in a compost-based biological filter. Chem. Eng. Sci. 57: 49995010. Deront, M., Samb, F.M., Adler, N. and Péringer, P. (1998) Biomass growth monitoring using pressure drop in a cocurrent biofilter. Biotechnol. Bioeng. 60: 97-104. Devinny, J. S., Deshusses, M.A. and Webster, T. S. (1999) Biofiltration for air pollution control. Lewis publishers, Boca Raton, Florida. Durham, D.R., Marshall, L.C., Miller, J.G. and Chmurny, A.B. (1994) Characteritzation of inorganic biocarriers that moderate system upsets during fixed-film biotreatment processes,.Appl. Environ. Microbiol. 60(9): 3329. Ergun, S. (1952) Fluid flor thorugh packed columns. Chem. Eng. Prog. 48: 9-94.
572
ANTONI D. DORADO, XAVIER GAMISANS, DAVID GABRIEL AND JAVIER LAFUENTE
Janni, K.A., Maier, W.J., Kuehn, T. H., Yang, C.-H., Bridges, B.B., Velsey, D. and Nellis, M. A. (2001) Evaluation of biofiltration of air-An innovative air pollution control technology. ASHRAE Transactions 107: 198-5214. Kennes, C. and Veiga, M.C. (2001) Bioreactors for waste gas treatment. Kluwer academic publishers, Dordrecht, The Netherlands. Morgan-Sagastume, F., Sleep, B.E. and Allen, D.G. (2001) Effects of biomass growth on gas pressure drop in biofilters. J. Environ. Eng. 127: 388-396. Morgan-Sagastume, J.M. and Noyola, A. (2006) Hydrogen sulfide removal by compost biofiltration: Effect of mixing the filter media on operational factors. Biores. Technol. 97: 1546-1553. TMECC (Test Methods for the Examination of Composting and Compost). (2002).
573
Suitability of dust and bioaerosols from a pig stable as inoculum for biological air filters ANJA KRISTIANSEN, PER HALKJÆR NIELSEN, AND JEPPE LUND NIELSEN Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 57, DK-9000 Aalborg
ABSTRACT Biofiltration for removing ammonia and odour compounds from ventilation air of pig stables is a promising approach. In order to reduce the time for starting up a well-functioning biofilter a good inoculum suited for the environment is needed. In this study the microbial identity and quantity of dust and bioaerosols coming from a pig stable were analyzed for its suitability as inoculum. Inoculation of biofilters with dust had similar good ammonia removal capabilities as biofilters inoculated with activated sludge, although analysis of the microbial dust community revealed clear differences. The organic fraction of the dust particles seems to be important for mediating biofilm development on the filter material.
1 INTRODUCTION Increasing productivity combined with a more centralized production of pigs has the last decades created odour nuisance problems and a high load on the environment through air emissions. Biofiltration is a promising and low cost technology for treatment of large volumes of air containing low concentrations of different compounds, such as ventilation air from a pig stable. The quantitatively most important components in ventilation air from pig stables are organic acids and ammonia, which can readily be oxidized by heterotrophic and nitrifying bacteria, respectively. A well-functioning biological air filter requires microorganisms capable of utilizing air components and adapted to live under extreme conditions. Inoculations by addition of activated sludge or material from existing filters often speed up the process of establishing a new biofilm. However, activated sludge might not be optimal
573
574
ANJA KRISTIANSEN, PER HALKJÆR NIELSEN
AND
JEPPE LUND NIELSEN
as the microorganisms herein are not adapted to the conditions on the new filter material nor the substrate fed into the biofilter. Microorganisms indigenous to the ventilation air, aerosols, and dust particles may already be pre-adapted to life in the environment and thus better suited as inoculum. However, only few investigations, most of which applying culture-dependent techniques, have dealt with the composition of airborne dust and aerosols from pig stables. By analyzing the microbial community herein its potential use as inoculum of nitrifying bacteria and heterotrophic microorganisms into biofilters can be evaluated. In the present study we have applied cultivation-independent molecular techniques to characterize the microbial community composition in dust and bioaerosols from a pig stable, and compared this to established biofilters treating air from the same stable.
2 MATERIALS AND METHODS 2.1 SAMPLING Dust and bioaerosols was obtained from a closed gestation unit with 450 sows near Aalborg, Denmark. Dust and bioaerosol samples were collected from the air by introducing a 2 cm thick (30 cm in diameter) water resistant cellulose filter with a pore size of <2 μm. Samples were collected in periods ranging from 2 to 4 days with an average air flow of 200 m3 h-1 in order to obtain homogenous samples. The dust load of the air was determined from the mass accumulated over time in the filters and the air flow (Windmaster, Kaindi). Dust and aerosol particles were extracted from the filters by washing them thoroughly in sterile filtered tap water on which all following analysis has been made unless otherwise stated. Biofilm samples from the biofilters were obtained by thoroughly shaking an aqueous suspension of the biofilter material until no more material fell off. 2.2 BIOFILTER SETUP Four identical biofilters were used to test the suitability of dust/bioaerosols as inuculum. The filters were cylindrical packed bed reactor (630 mm in diameter and 700 mm in height) packed with 300 mm of Lightweight Expanded Clay Aggregate beads, 10-20 mm in diameter (LECA®; Maxit A/S,). The filters had an initial porosity of 46% and a specific surface area of 300 m2 m-3 and a density of 215 kg m-3. The pig stable air was introduced through the bottom of the columns for counter-current flow operation and maintained with an airflow of 40-44 m3 h-1. The empty bed residence time was 8.1-8.7 s. The filters were kept moist by a discontinuous water addition of 3700 mL day-1. Homogenized activated sludge (400 mL), from nitrogen removing wastewater treatment plant (Aalborg East wastewater treatment plant) was added to two filters.
SUITABILITY OF DUST AND BIOAEROSOLS FROM A PIG STABLE AS INOCULUM
575
2.3 CULTIVATION AND IDENTIFICATION OF THE MICROBIAL COMMUNITY Quantification of viable microorganisms in the dust was done by serial dilutions of the dust sample using sterile filtered tap water. 100 μL of each dilution was plated on triplicate trypticase soy agar plates for bacteria and incubated for 48 h at room temperature (22ºC) and on triplicate malt extract agar plates (Smid et al., 1989) supplemented with 0.1 g L-1 chloramphenicol for quantification of fungi for 48 h at room temperature (22ºC). Different staining procedures were applied to the samples in order to gain information about the bacterial community. Total cell counts was obtained by DAPI (4',6-diamidino-2-phenyl indole) staining a sample concentrated on a polycarbonate filter (Osmonics Inc.) as described by Frølund et al. (1996). BacLight Live/Dead staining (Molecular Probes) and incubation with 1 mM Cyanoditolyl Tetrazolium Chloride (CTC) (Polysciences Inc.) was used to examine the viability of the cells in the sample. Identification and quantification of the major bacterial phyla in the dust and biofilters was performed with fluorescence in situ hybridization (FISH) and a range of oligonucleotide probes (Table 1). Further details on these probes are available in probeBase (www.probeBase.net). Cells were fixed by paraformaldehyde and ethanol as previously described (Kong et al., 2005) Fixed samples were stored at -20°C in 50% ethanol and 50% phosphate buffered saline until analysis. FISH was performed on gelatine-coated glass slides according to the method of Manz et al., (1992). Application of gelatine-coated slides and addition of 1% agarose on top of the sample reduced the loss of cells during FISH from 35-80% to only 7%. 52% of the total DAPI counts were detectable by hybridization of the general bacterial probemix (EUBmix). An Axioscop II (Zeiss) and a 630X magnification was used for all visualizations. For counting, a minimum of 400 cells were counted per sample distributed between at least 10 randomly chosen counting grids. DNA extraction was carried out using FastDNA SPIN kit for Soil (Vista). Undiluted extracted DNA was amplified by polymerase chain reaction (PCR) with the primers 8F and 1492R, purified, ligated, cloned and sequenced as previous described by Kong et al. (2005). 2.4 NITRIFICATION ACTIVITY Nitrification activity in the dust was evaluated by measuring the production of NO x under aerobic conditions (with and without the presence of 5mg L -1 Nallylthiourea) in dust suspended in 0.5 L 0.9% NaCl. No effect of NOx and excess of organic material supporting growth of heterotrophic microorganisms, was observed by implementing various pre-incubations prior to addition of 5 mM NH4+.
576
ANJA KRISTIANSEN, PER HALKJÆR NIELSEN
AND
JEPPE LUND NIELSEN
2.5 ANALYTICAL METHODS The water content and the content of organic matter in the dust was determined after drying at 105°C for 24 h followed by 550°C for 3 h. The fat content was quantified by mixing a diluted dust sample with Nile red to a concentration of 6.5 μg mL-1 and determining the fluorescence (excitation at 510 nm and emission at 580 nm) by fluorescence spectrophotometer, while protein was quantified with the CBQCA Protein Quantification Kit (Molecular Probes). The fluorescence was measured and quantified using oleic acid or gelatine as standard compounds. NH4+, NO2- and NO3- concentrations were determined after extraction with 1 M KCl for 2 h for biofilter samples and after sterile filtration for dust samples. They were analyzed by an autoanalyzer (Technicon TRAACS 800).
3 RESULTS AND DISCUSSION Inorganic filter material has the advantage over organic materials that it is not or only very slowly degraded. However, it often requires inoculation in order to minimize startup time. Addition of a proper inoculum can help reach the required performance of the filter faster. The challenge is to find the right inoculum. Complex microbial consortia, such as activated sludge are often chosen for a biofilter treating air with a broad specter of organic substrates, while commercially available monocultures are favored for treatment of a particular contaminant. Activated sludge is rich in diversity, but the cells herein might not be well-adapted for treating the components and the actual concentrations of the contaminated air. Thus better inoculum candidates should be found. An unknown fraction of the microorganisms indigenous to dust and aerosols in contaminated air systems can be speculated to be pre-adapted for living in the contaminated air. 3.1 DUST COMPOSITION AND MICROBIAL DETECTION In the present study we have investigated the microbial composition of dust and aerosols, compared it to the composition of biofilters treating the same air, and evaluated íts potential as inoculum for biofilters. Initial experiments aimed to analyze the quality of the air regarding load and composition of microorganisms. The dust load in the ventilation air was measured to 0.7 mg TS m-3 (total solids) of which 80% of the TS were carbohydrate, 15% were protein and 5% fatty acids. The culturable fraction of the dust particles averaged 1.31 × 105 colony forming units (CFU) m-3 for bacteria and 1.33 × 104 CFU m-3 for fungi. These numbers are largely in agreement with measurement from other pig stable environments (Chi and Li, 2005; Predicala et al., 2002) and show a high microbial input into the biofilters. The culturable part of
SUITABILITY OF DUST AND BIOAEROSOLS FROM A PIG STABLE AS INOCULUM
577
the total bacterial community, however, only constituted less than 1% as 3.45 × 107 cells m-3 was detected by DAPI. The equivalent number for activated sludge was 5 × 109 cells ml-1 with a culturable fraction of 10-15%. BacLight Live/Dead staining revealed 65% living cells in the dust, of which most were highly respiratory active, as determined by CTC staining. 3.2 NITRIFICATION ACTIVITY One of the main advantages of biofiltration is the high removal of ammonia through microbial nitrification. The biofilter system analyzed, when running steadily, had ammonia removal efficiencies varying between 70-90% (Figure 1) which are in agreement with other surveys (Melse and Mol, 2004; Sheridan et al., 2002).
Figure 1. Removal efficiency of ammonia in four biofilters receiving approximately 44 m3 h-1 air from a pig stable. Solid symbols indicate biofilters inoculated with activated sludge, while open symbols show biofilters which have only received air with 0.7 mg dust particles per m3.
More than 60% removal efficiency of ammonia was seen after approximately 44 days for the biofilters inoculated with activated sludge, while the two filters only supplied with dust, reached the same level after approximately 52 days (Figure 1). The corresponding control receiving filtered air without dust particles did not reach an ammonia removal efficiency above 60% within 120 days (results not shown). These results indicate the importance of inoculating the biofilters, and that a continuous supply of dust particles to the filters is almost as efficient as inoculation with activated sludge. The low dust load from the stable (most likely a consequence of the addition of fat to the pig feed to reduce dust), could further be speculated to be limiting for an even faster startup time of the filters. Other values of airborne dust concentration has
578
ANJA KRISTIANSEN, PER HALKJÆR NIELSEN
AND
JEPPE LUND NIELSEN
been reported to be between 1.66-21 mg m-3 (Crook et al., 1991). The relatively fast build up of ammonia removal obtained by inoculation with dust, compared to the control with dust free air, could be the result of providing a suitable microbial consortium to the filters with the dust, and/or a supply of organic coating of the filter for establishing a proper biomass. The nitrifying activity of the dust was investigated and it was found that ammonia was removed with a rate of approximately 0.05 μmol m-3 h-1 (74.8 μmol g TS-1 h-1). However, no measurable accumulation of nitrite or nitrate was observed, indicating that the removal of ammonia was most likely due to heterotrophic activity by being incorporated into the biomass. Further analyses of the microbial consortia were therefore carried out to evaluate this. 3.3 IDENTIFICATION OF THE MICROBIAL COMMUNITY The identity of the bacteria in the dust particles were analysed by FISH. Only 52% of all microorganisms were detected by this approach, with the majority (85%) being gram positive cells distributed with 10% Actinobacteria and 75% Firmicutes (Table 1). Higher taxonomic analysis revealed that 50% of the FISH-detected community derived from the genus Veillonella and 36% from the genus Streptococcus. Alpha-, Beta-, Gamma-, and Deltaproteobacteria made up only ca. 1% of the community each. In agreement with the rate measurements, use of specific probes for ammonia oxidizers showed, that they constituted less than 1% of all FISH detectable cells. No nitrite oxidizers (Nitrobacter or Nitrospira) were found.
Figure 2. Fluorescence in situ hybridization of dust particles with the general probe mix EUBmix (A) and a specific probe targeting Veillonella (B). The color has been artificially changed to black and white nuances to increase the contrast. The arrows indicate the presence of Veillonella positive cells which is not targeted by the EUBmix. Bar equals 20μm.
579
SUITABILITY OF DUST AND BIOAEROSOLS FROM A PIG STABLE AS INOCULUM
Table 1. List of microorganisms analysed with FISH showing phylogroup, probe used and its specificity and reference to each tested probe. The portion of the tested organism group compared to Eubacteria is presented in percent. Probe name
Specificity
Reference
ALF968 BET42a Nso190
Alphaproteobacteria Betaproteobacteria Ammonia oxidizing Betaproteobacteria Gammaproteobacteria Most Deltaproteo-bacteria and some other Bacteria Most Actinobacteria Many Actinobacteria Firmicutes Streptococcus Veillonella dispar, V. parvula, V. atypical Most enterococci
Neef, (1997) Manz et al., (1992) Mobarry et al., (1996)
<1 1 <1
Manz et al., (1992) Amann et al., (1990) Rabus et al., (1996) Roller et al., (1994) Erhart, (1997) Meier et al., (1999) Trebesius et al., (2000) Harmsen et al., (2002)
1 1 10 1 77 36 52
Meier et al., (1997)
1
GAM42a SRB385 + SRB385Db HGC69a HGC236 LGCmixLGC354b+A+C Str Veil223 Enc131
Percentage of Eubacteria [%]*
* Percentage of specific probe of EUBmix (Amann et al., 1990; and Daims et al., 1999). ** Hybridization of dust with the following probes did not result in positive FISH signals: Actino1011 (HGC1011) (Some Actinobacteria) (Liu et al., 2001), HGC1351 (Many Actinobacteria) (Erhart, 1997), Myb736a (Mybobacterium complex ) (De los Reyes et al., 1997), Sau (Kempf et al., 2000), Lis637 (Listeria) Schmid et al., (2003), CF319a+b (Bacteroidetes) Manz et al., (1996), Gnsb941+CFX1223 (Phylum Chloroflexi) (Gich et al., 2001), GCB532 (Green sulphur bacteria) (Tusehak et al., 1999) Ntspa662 (Nitrospira) (Daims et al., 2000), NIT3 (Nitrobacter) (Wagner et al., 1996) Arch915 (Archaea) (Amann et al., 1990) Pla46 (Planctomycetes) (Neef et al., 1998), NonEUB338 (Nonsense) (Wallner et al., 1993).
Through FISH analysis it was observed that 60% of the Veil223 positive cells were not targeted by the EUBmix probeset (Figure 2) although its phylogenetic affiliation in the phylum of Firmicutes is generally considered to be targeted by this probe mixture. The low fraction of FISH-detectable cells, was thus significantly higher than the 52%, and could be calculated to be around 83% of all DAPI counts. The fraction of FISH-detectable cells of all DAPI counts is often, although not entirely correct, used to interpret the actual activity level of the sample (Bouvier and del Giorgio, 2003). In the case of our dust and aerosol sample we have further seen that most (65% of all cells in the sample were viable and most of these were metabolically active as determined by Live/Dead and CTC staining. Therefore, very likely they could be
580
ANJA KRISTIANSEN, PER HALKJÆR NIELSEN
AND
JEPPE LUND NIELSEN
transferred to and trapped in the biofilters and there be actively respiring from the contaminants in the ventilation air. Both Veillonella and Streptococcus are known to carry out fermentation and therefore if established in the biofilm may play a role in the removal of organic odours compounds present in the air. Approximately 57% of all FISH-detectable cells in the biofilters inoculated with activated sludge were nitrifiers, while only 19% were nitrifiers in the biofilters receiving dust. The identity of these nitrifiers were tested with more specific oligonucleotide probes, but only minor fractions were identified. In the biofilters inoculated with activated sludge ca. 1% of the ammonia oxidizers were found to belong to the genus of Nitrosomonas and ca. 1% of cells belonging to the genus Nitrosococus mobilis. In the biofilters inoculated with dust, approximately 50% of all ammonia oxiders were affiliated within the Nitrosomonas oligotropha-lineage. No nitrite oxidizers (Nitrobacter or Nitrospira) were found in any of the biofilters. A generally higher number of FISH-detectable cells of all DAPI counts were seen in the biofilters inoculated with dust (ca. 86%) compared to biofilters inoculated with activated sludge (ca. 51%). No significant difference in the total number of DAPI counts (ca. 5-8 × 105 cells g dry LECA®) were seen between the two types of biofilters. To further investigate this difference in microbial consortia between the dust and biofilters we have screened the constructed clone library has revealed mainly two populations within the dust both belonging to the phylum of Firmicutes, namely sequences closely related to Clostridium and Bacillus species which originated from pig intestine samples. From cultured isolates Bacillus has been identified in dust and bioaerosols from swine barns (Predicala et al., 2002). Further investigations will focus on expanding the molecular investigations of the microbiology of the dust in order to compare larger fractions of the entire communities. In conclusion it appears that the microbial consortia in the dust is not directly reflected in the biofilters. Almost no nitrifiers were observed in the dust and aerosols indicating that the organic fraction of the dust particles is important for mediating biofilm development on the filter material. A clear difference in the diversity was observed between biofilters inoculated with activated sludge and with biofilters only receiving dust although no clear difference is seen between the performance of these biofilters. Thus it cannot be concluded from this study to which extent dust is preferable to activated sludge as an inoculum. However, dust particles and bioaerosols seems to hold sufficient capacity to be considered as a good candidate for inoculum for biofilters treating ventilation air from pig stables.
SUITABILITY OF DUST AND BIOAEROSOLS FROM A PIG STABLE AS INOCULUM
581
4 ACKNOWLEDGEMENTS The present research was funded by the Danish Ministry of Food, Agriculture and Fisheries (Project title: Function of Biological Airfilters).
REFERENCES Amman, R.I., Krumholz, L. and Stahl, D.A. (1990) Fluorescent-oligonucleotide probing of whole cells for determinative, phylogenetic, and environmental-studies in microbiology. J. Bacteriol. 172 (2): 762-770. Amann, R.I., Stromley, J., Devereux, R., Key, R. and Stahl, D.A. (1992) Molecular and microscopic identification of sulfate-reducing bacteria in multispecies biofilms. Appl. Environ. Microbiol. 58 (2): 614-623. Bouvier, T. and del Giorgio, P. A. (2003) Factors influencing the detection of bacterial cells using fluorescence in situ hybridization (FISH): A quantitative review of published reports. FEMS Microbiology Ecology. 44: 3-15. Chi, M.-C. and Li, C.-S. (2005) Fluorochrome and fluorescent in situ hybridization to monitor bioaerosols in swine buildings. Aerosol Sci. Tech. 39: 1101-1110. Crook, B., Robertseon, J.F., Travers Glass, S.A., Botheroyd, E.M., Lacey, J., and Topping, M.D. (1991) Airborne dust, ammonia, microorganisms and antigens in pig confinement houses and the respiratory health of exposed farm workers. Am. Ind. Hyg. Assoc. J. 52(7): 271-279. Daims, H., Nielsen, P.H., Nielsen, J.L., Juretschko, S. and Wagner, M. (2000) Novel Nitrospiralike bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Wat. Sci. Tech. 41 (4-5): 85-90. De los Reyes, F.L., Ritter, W. and Raskin, L. (1997) Group-specific small-subunit rRNA hybridization probes to characterize filamentous foaming in activated sludge systems. Appl. Environ. Microbiol. 63 (3): 1107–1117. Erhart, R. (1997) Thesis. Technische Universität München, Munich, Germany. Frølund, B., Palmgren, R., Keiding, K. and Nielsen, P.H. (1996) Extraction of extracellular polymers from activated sludge using a cation exchange resin. Wat. Res. (8): 1749-1758. Gich, F., Garcia-Gil, J. and Overmann, J. (2001) Previously unknown and phylogenetically diverse members of the green nonsulfur bacteria are indigenous to freshwater lakes. Arch. Microbiol. 177 (1): 1-10. Harmsen, H.J.M., Raangs, G.C., He, T., Degener, J.E. and Welling, G.W. (2002) Extensive set of 16S rRNA-based probes for detection of bacteria in human feces. Appl. Environ. Microbiol. 68 (6): 2982-2990. Kempf, V.A.J., Trebesius, K. and Autenrieth, I.B. (2000) Fluorescent in situ hybridization allows rapid identification of microorganisms in blood cultures. J. Clin. Microbiol. 38 (2): 830-838.
582
ANJA KRISTIANSEN, PER HALKJÆR NIELSEN
AND
JEPPE LUND NIELSEN
Kong, Y.H., Nielsen, J.L. and Nielsen, P.H. (2005) Identity and ecophysiology of uncultured actinobacterial polyphosphate-accumulating organisms in full-scale enhanced biological phosphorus removal plants. Appl. Environ. Microbiol. 71 (7): 4076-4085. Liu, W.T., Nielsen, A.T., Wu, J.H., Tsai, C.S., Matsuo, Y. and Molin, S. (2001) In situ identification of polyphosphate- and polyhydroxyalkanoate-accumulating traits for microbial populations in a biological phosphorus removal process. Environ. Microbiol. 3 (2): 110-122. Manz W, Amann R, Ludwig W, Vancanneyt M and Schleifer KH (1996) Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bacteria of the phylum cytophaga-flavobacter-bacteroides in the natural environment. Microbiology-UK. 142: 1097-1106. Manz, W., Amann, R., Ludwig, W., Wagner, M. and Schleifer, K.H. (1992) Phylogenetic oligodeoxynucleotide probes for the major subclasses of proteobacteria-problems and solutions. Syst. Appl. Microbiol. 15 (4): 593-600. Meier, H., Amann, R., Ludwig, W. and Schleifer, K.H. (1999) Specific oligonucleotide probes for in situ detection of a major group of Gram-positive bacteria with low DNA G+C content. Syst. Appl. Microbiol. 22 (2): 186-196. Meier, H., Koob, C., Ludwig, W., Amann, R., Frahm, E., Hoffmann, S., Obst, U. and Schleifer, K.H. (1997) Detection of enterococci with rRNA targeted DNA probes and their use for hygienic drinking water control. Wat. Sci. Tech. 35 (11-12): 437-444. Melse, R.W. and Mol, G. (2004) Odour and ammonia removal from pig house exhaust air using a biotrickling filter. Wat. Sci. Tech. 50(4): 275-282. Neef, A. (1997) Anwendung der in situ Einzelzell-Identifizierung von Bakterien zur Populationsanalyse in komplexen mikrobiellen Biozönosen. Doctoral thesis (Technische Universität München). Neef, A., Amann, R., Schlesner, H. and Schleifer, K.H. (1998) Monitoring a widespread bacterial group: in situ detection of planctomycetes with 16S rRNA-targeted probes. MicrobiologyUK. 144: 3257-3266. Predicala, B.Z., Urban, J.E., Maghirang, R.G., Jerez, S.B. and Goodband, R.D. (2002) Assessment of bioaerosols in swine barns by filtration and impaction. Current Microbiol. 44: 136-140. Rabus, R., Fukui, M., Wilkes, H. and Widdel F (1996) Degradative capacities and 16S rRNA-targeted whole-cell hybridization of sulfate-reducing bacteria in an anaerobic enrichment culture utilizing alkylbenzenes from crude oil. Appl. Environ. Microbiol. 62 (10): 3605-3613. Roller, C., Wagner, M., Amann, R., Ludwig, W. and Schleifer, K.H. (1994) In situ probing of Gram-positive bacteria with high DNA G+C content using 23S rRNA-targeted oligonucleotides. Microbiology. 140: 2849-2858. Schmid, M., Walcher, M., Bubert, A., Wagner, M., Wagner, M. and Schleifer, K.H. (2003) Nucleic acid-based, cultivation-independent detection of Listeria spp. and genotypes of Lmonocytogenes. FEMS Immunol. Med. Microbiol. 35 (3): 215-225. Sheridan, B., Curran, T., Dodd, J. and Colligan, J. (2002) Biofiltration of odour and ammouia from a pig unit-pilot-scale study. Biosyst. Engineering. 82 (4): 441-453.
SUITABILITY OF DUST AND BIOAEROSOLS FROM A PIG STABLE AS INOCULUM
583
Smid, T., Schokkin, E., Boleij, J.S.M. and Heederik, D. (1989) Enumeration of viable fungi in occupational environments: A comparison of samplers and media. Am. Ind. Hyg. Assoc. J. 50 (5): 235-239. Trebesius, K., Leitritz, L., Adler, K., Schubert, S., Autenrieth, I.B. and Heesemann, J. (2000) Culture independent and rapid identification of bacterial pathogens in necrotising fasciitis and streptococcal toxic shock syndrome by fluorescence in situ hybridization. Med. Microbiol. Immunol. 188 (4): 169-175. Tuschak, C., Glaeser, J. and Overmann, J. (1999) Specific detection of green sulfur bacteria by in situ hybridization with a fluorescently labeled oligonucleotide probe. Arch Microbiol. 171: 265-272. Wallner, G., Amann, R. and Beisker, W. (1993) Optimizing fluorescent in situ hybridization with ribosomal-RNA-targeted oligonucleotide probes for flow cytometric identification of microorganisms. Cytometry. 14 (2): 136-143.
This page intentionally left blank
585
PILOT-SCALE AND INDUSTRIAL APPLICATIONS
This page intentionally left blank
587
Biofiltration of bitumen vapours – Operational aspects MATTHIEU GIRARD1, JEAN-LOUIS FANLO2, NICOLAS TURGEON3, GERARDO BUELNA3 AND PAUL LESSARD1 1
Civil Engineering Department, Pavillon Pouliot, University Laval, Québec, Qc., Canada, G1K 7P4 2 École des Mines d’Alès, 6 Avenue de Clavières, 30319 ALES Cedex, France 3 Industrial Research Center of Québec, 333 Franquet street, Québec, Qc., Canada, G1P 4C7
ABSTRACT This study was carried out in response to odour problems around a bitumen mixing and storage plant. The general objective of this study was to determine the potential of biofiltration for the treatment of air containing bitumen vapours. Two pilot-scale biofilters, a single-stage system and a two-stage system, were operated using a synthetic gas for a period of 106 days. Results demonstrated that a period of about 50 days was necessary to reach a steady state. The two biofilters performed very well in regards to H2S, while VOC treatment was much less efficient. Maximum elimination capacities of 25 g·m-3·h-1 for H2S and 5.3 g·m-3·h-1 for VOCs were obtained with the two-stage system. It was possible to establish certain operating conditions necessary for proper operation and to determine that a two-stage system is more efficient for the simultaneous treatment of H2S and VOCs.
1 INTRODUCTION The deterioration of air quality, whether it is in terms of health risks or odour issues, has caused an increase in public awareness towards air pollution. Furthermore, since the beginning of the 20th century, emissions of atmospheric pollutants in North America have been in continual increase (Delhoménie and Heitz, 2005). This type of pollution is caused by a number of different components which are released by a variety of industries. Because of their effects on human health and their prevalence in industrial emissions, two types of pollutants are particularly interesting, VOCs and odours.
587
588
M. GIRARD, J. L. FANLO, N. TURGEON, G. BUELNA AND P. LESSARD
VOCs are organic compounds which have a vapour pressure above 10 Pa at 20°C according to European regulations (Fanlo, 2006). Due to the vague aspect of this definition, these components have a wide range of effects on human health and the environment. VOCs can cause nausea, headaches, irritation of the nervous and respiratory systems and even certain cancers (Delhoménie and Heitz, 2005). They also contribute to the production of tropospheric ozone which is an important component of smog (Thérrien, 2005). Odours can be defined as the result of interpretation by the brain of an interaction between a volatile compound and the olfactory senses (Fanlo, 2006). The measurement of odours is therefore difficult since there is no way to know beforehand whether a specific compound is odorous or not. Odours can be described by their quality, hedonic tone and intensity and are often composed of sulphur, nitrogen or organic compounds (Brandy, 1997). Reduced sulphur compounds, such as hydrogen sulphide (H2S), are reported to by highly odorous molecules. Symptoms can vary from simple discomfort, allergies or nausea to loss of concentration and depression (Météoglobe Canada Inc., 1993). Traditionally, the emission of air pollutants has been managed by physicochemical treatment methods. However, biological treatment technologies are rapidly gaining in popularity because of their low investment and operating costs as well as the simplicity of their operation (Michel, 1997). There are three types of bioreactors, the biofilter, the bioscrubber and the biotrickling filter, which are distinguished by the mobility of the biomass and the liquid phase (Devinny et al., 1999). Biofiltration is the oldest and most widespread biotechnology for the treatment of air pollutants (Jorio and Heitz, 1999). There are two distinct steps in the biofiltration process: transfer of the pollutants to the aqueous phase or the filter media and their degradation by the microorganisms (Malhautier et al., 2005). There are many parameters to consider when designing and operating a biofilter, such as filter media, air flow, pollutants, moisture content, pressure loss, microorganisms, temperature and pH. Biofiltration has been used to treat a great deal of different components, several authors have compiled some of the industrial and laboratory applications of this technology (Easter et al., 2005; Iranpour et al., 2005; Mulligan, 2002; Kennes and Thalasso, 1998). In recent years, the industrial sector of storage and use of bitumen has been in the spotlight concerning its gaseous emissions. Bitumen is a residue of the distillation of petroleum which is used in many industries, as in asphalt production for example. When the bitumen is stored, it must be kept hot, around 170°C, to maintain its fluidity. This causes an important production of bitumen vapours, which can then disturb a great deal of people. If solid bitumen contains mainly heavy hydrocarbons, its vapour is composed of much lighter compounds, such as VOCs and odourous substances of which H2S is the most important (Cook et al., 1999). Biofiltration of single substances or of compounds from the same chemical family (sulphur compounds, VOCs, etc.) is well documented but few studies have
BIOFILTRATION OF BITUMEN VAPOURS-OPERATIONAL ASPECTS
589
examined the simultaneous treatment of pollutants from different families as is the case of bitumen vapours. Koe and Liang (2005) studied two biofilter configurations for the co-treatment of toluene and H2S. They found that H2S significantly decreased the pH of the filter bed without pH control and that a two-stage biofilter was more efficient to treat H2S and toluene. Closer to bitumen vapours, Cook et al. (1999) published a paper on the biofiltration of emissions from a polymer-modified asphalt plant. The objective of their study was the removal of H2S since it was considered as the main contributor to the odour problem. This was achieved very well except during peaks of H2S which reached 30 000 ppmv. However, VOC removal was negligible. The general objective of this study was to determine the potential of biofiltration for the treatment of air containing bitumen vapours. Two biofilters configurations were studied: a single-stage system and a two-stage system. A secondary objective was to determine the operating parameters necessary for proper operation.
2 MATERIALS AND METHODS 2.1 BITUMEN VAPOUR ANALYSIS Considering the technical difficulties of reproducing bitumen vapours in a laboratory setting, it was decided to create a synthetic gas representative of these vapours to test the pilot-scale biofilters. A portion of solid bitumen was first heated to 170°C and then samples of vapour were drawn directly from the headspace above the bitumen. Analysis of the vapours was done first with a GC-MS (gas chromatograph with a mass spectrometer detector) for a general approach and then with a GC-PFPD (gas chromatograph with a pulsed flame photometric detector) for sulphur compounds. It was possible to identify over a hundred compounds, ranging from light to heavy hydrocarbons (C-2 to C-24) with several sulphur compounds. Subsequently, this analysis was used to select specific compounds to incorporate in the synthetic gas. The final composition of the gas is given in table 1. Table 1. Composition of the synthetic gas. Component Hexane Decane Toluene p-Xylene Thiophene H2 S
Concentration(mg·m3) 73.5 73.5 12.5 12.5 3 150
590
M. GIRARD, J. L. FANLO, N. TURGEON, G. BUELNA AND P. LESSARD
2.2 EXPERIMENTAL SETUP The experimental setup used for this study is represented by the schematic in Figure 1.
Figure 1. Schematic of the biofilters.
The biofilters are 25 cm in diameter and were packed with 1 m of an organic filter media made up of wood chips and peat moss. Biofilter 1 was composed only of column 1 and biofilter 2 was composed of columns 2A and 2B. The gas flow rate was constant at 4.9 m3·h-1, corresponding to a gas velocity of 100 m·h-1 and empty bed residence times of 36 and 72 seconds for biofilter 1 and 2 respectively. H2S was introduced with a bottle of liquefied gas while the blend of the five VOCs was injected using a micropump. A nutrient solution was periodically sprayed on the surface of the biofilters to maintain proper humidity and to ensure bacterial growth. Sampling ports were located at the inlet, the outlet and at 5 equidistant intervals over the height of each biofilter. The two biofilters were operated for a period of 106 days according to the schedule in table 2:
BIOFILTRATION OF BITUMEN VAPOURS-OPERATIONAL ASPECTS
591
Table 2. Operation schedule. Phase I II III IV
Description Start-up H2S starvation period pH control Removal of hexane
Duration (d) 50 23 29 4
2.3 ANALYTICAL METHODS H2S concentrations were measured with a gas chromatograph (GC) equipped with a photo-ionisation detector (PID) from HNU Systems with an oven temperature of 80°C, a temperature of 200°C for the injector and the detector and an analysis time of 2 minutes. Total VOC concentrations were determined with a Graphite 700 flame ionisation detector (FID) from Cozma which was calibrated with methane at 100 mg·m-3 before every series of measurements. Individual VOC concentrations were evaluated with an HP-5890 Series II GC-FID equipped with a 30m×0.53mm HP-1 column. Oven temperature was held at 80°C and analysis time was 6 minutes. Samples were injected automatically with the sampling pump on each of three systems.
3 RESULTS AND DISCUSSION 3.1 H2S ELIMINATION A graph of the H2S concentration versus time is presented in Figure 2.
Figure 2. H2S concentration versus time.
592
M. GIRARD, J. L. FANLO, N. TURGEON, G. BUELNA AND P. LESSARD
Immediately after start-up, the biofilters offered a certain treatment, system 1 reduced the H2S concentration to about 130 mg·m-3 while system 2 reached 45 mg·m-3. Two weeks later, H2S was removed below the detection limit by biofilter 2. The rapid drop for biofilter 1 around day 20 was due to an increase in liquid flow to rectify the inadequate moisture level in this biofilter. After 40 days of operation, both biofilters removed all of the H2S that was introduced. Even the starvation period (phase II) between days 50 and 64 had no effect on performance. During the first 70 days of operation, lixiviate pH slowly decreased and reached values around 1 due to sulphuric acid production from the biodegradation of H2S. To counteract this acidification and maintain a neutral pH, sodium bicarbonate was added in the nutrient solution during phase III. This adjustment of the pH had no obvious effect on H2S removal. Using the intermediate sampling ports, it was possible to determine that only 80 cm of biofilter 1 and 60 cm of biofilter were necessary to remove all the H2S. This corresponds to elimination capacities of 18.8 g·m-3·h-1 and 25 g·m-3·h-1. 3.2 VOC TREATMENT Results for total VOC concentration in this section are given in methane equivalents since the FID was calibrated using this gas. A graph of total VOC concentrations versus time is given in Figure 3.
Figure 3. Total VOC concentration versus time.
From Figure 3, it can be seen that there is no biodegradation for the first 25 days, then biofilter 2 starts removing the VOCs and their concentration falls to a stable value of 75 mg·m-3 on day 50. It is this stabilization that was used to define the start-up time of the biofilters. During this first phase, biofilter 1 only begins to treat
BIOFILTRATION OF BITUMEN VAPOURS-OPERATIONAL ASPECTS
593
VOCs, removing about 10 mg·m-3. There is an interesting connection to be made between the complete removal of H2S and the start of VOC treatment by biofilter 2. This could just be a coincidence, but it could also be caused by the acidifying effect of H2S (Koe and Liang, 2005). However, it could also be the standard time required by the microorganisms to adapt to the specific VOCs in the synthetic gas. During phase II, H2S starvation had a negligible effect on biofilter 2, but a significant improvement of VOC removal for biofilter 1 was observed. Adjustment of the pH in phase III didn’t seem to have any effect on VOC elimination. Maximum removal efficiencies of 20% and 60% were obtained with biofilter 1 and 2 respectively, corresponding to elimination capacities of 3.5 g·m-3·h-1 and 5.25 g·m-3·h-1. 3.3 HEXANE REMOVAL To explain the mediocre results concerning total VOC removal, individual VOC concentrations were examined. Figure 4 shows the average removal efficiency for days 53 to 67 for the five individual VOCs and the total VOC.
Figure 4. Individual VOC Removal Efficiencies for Days 53 to 67.
From the data shown in Figure 4, it is possible to establish an order of biodegradability: alkanes > aromatics > thiophene. However, hexane does not fit in this classification since it was barely treated by the biofilters. It was also this compound that greatly limited overall VOC elimination. According to Maier et al. (2000), medium size alkanes (n-alkanes ranging from C-10 to C-18) are more biodegradable than smaller and larger molecules of the same family. The low solubility of long alkanes reduces their bioavailability while it is a toxicity problem with short alkanes. This could explain why hexane was not removed by the biofilters whereas decane was almost entirely treated.
594
M. GIRARD, J. L. FANLO, N. TURGEON, G. BUELNA AND P. LESSARD
Knowing this, it is possible to recalculate removal efficiencies while considering only the four VOCs other than hexane. Maximum values of 41% and 98% were obtained for biofilters 1 and 2 using this method. It was also decided to remove hexane from the synthetic gas while maintaining the same total VOC load in phase IV. This change caused a rapid decrease in total VOC concentrations for both biofilters as can be seen in Figure 3. Unfortunately, due to technical difficulties, it was not possible to pursue the experiments further to observe complete VOC removal. Nevertheless, it was possible to obtain elimination capacities of 6.5 g·m-3·h-1 and 8.6 g·m-3·h-1 for biofilters 1 and 2.
4 CONCLUSIONS This study was carried out in response to odour problems around a bitumen mixing and storage plant. The potential of biofiltration for the treatment of a synthetic gas representative of bitumen vapours was shown. Of the two configurations tested, the two-stage system proved to be more efficient for the simultaneous removal of H2S and VOCs. Adequate filter bed moisture content and pH were essential operating parameters for proper performance. Maximum elimination capacities of 25 g·m-3·h-1 for H2S and 8.6 g·m-3·h-1 for VOCs were obtained with the two-stage biofilter.
5 ACKNOWLEDGEMENTS This project was funded by the Laboratoire du Génie de l’Environnement Industriel (LGEI) at the École des Mines d’Alès (EMA) in France and by the Industrial Research Center of Québec (CRIQ). Matthieu Girard was supported financially by a Master’s grant from the Natural Sciences and Engineering Research Council of Canada (NSERC).
REFERENCES Brandy, J. (1997) Hydrodynamique et transfert gaz-liquide dans un bioréacteur. Application au transfert d’émissions gazeuses par lavage biologique. Thesis from the Université de Pau et des pays de l’Adour, France. Cook, L.L., Gostomski, P.A. and Appel, W.A. (1999) Biofiltration of asphalt emissions: Full-scale operation treating off-gases from polymer-modified asphalt production. Environ. Prog. 18(3): 178-187.
BIOFILTRATION OF BITUMEN VAPOURS-OPERATIONAL ASPECTS
595
Delhoménie, M-C. and Heitz, M. (2005) Biofiltration of air: a review. Critical Reviews Biotechnol. 25(1-2): 53-72. Devinny, J.S., Deshusses, M.A. and Webster, T.S. (1999) Biofiltration for air pollution control. CRC Press LLC, United States. 109 p. Easter, C., Quigley, C., Burrowesa P., Witherspoon, J. and Apgar, D. (2005) Odor and air emissions control using biotechnology for both collection and wastewater treatment systems. Chem. Engin. J. 113: 93-104 Fanlo, J.-L. (2006) Odeurs et Composés Organiques Volatiles-Contexte général, Problématique et Procédés de traitement, Course Notes, École des Mines d’Alès, France Iranpour R., Cox H.H.J., Deshusses, M.A. and Schroeder, E.D. (2005) Literature review of air pollution control biofilters and biotrickling filters for odor and volatile organic compound removal. Environ. Prog. 24(3): 254-267. Jorio, H. and Heitz, M. (1999) Traitement de l’air par biofiltration, Canadian Journal of Civil Engin. 26: 402-424. Kennes, C. and Thalasso, F. (1998) Waste gas biotreatment technology. J. Chem. Technol. Biotechnol. 72(4): 303-319. Koe, L.C.C. and Liang, J. (2005) Comparison of two-stage and single-stage biofiltration for H2S and toluene co-treatment. Proceedings of the International Congress Biotechniques for Air Pollution Control, La Coruña, Spain, October 5-7 2005, pp. 95-102. Maier, R.M., Pepper I.L. and Gerba, C.P. (2000) Environmental Microbiology, Academic Press, United States, pages 380, 381 and 385. Malhautier, L., Khammer, N., Bayle, S. and Fanlo, J-L. (2005) Biofiltration of volatile organic compounds. Appl. Microbiol. Biotechnol. 68(1): 16-22. Météoglobe Canada Inc. (1993) Applicabilité de méthodes de mesure d’odeurs au voisinage des lieux de production, d’entreposage et l’épandage du purin de porc : rapport final, Québec: Ministère de l’Environnement, 193 pages, QEN/AE93-8/6 Michel, M-C. (1997) Traitement de l’hydrogène sulfuré par biolavage, Master’s Dissertation. Laval University, Québec. Mulligan, C.N. (2002) Environmental Biotechnology – Technologies for Air, Water, Soil, and Wastes, ABS Consulting, United States, pages 39-76. Thérrien, M. (2005) Les composés organiques volatils (COV) dans l’air ambiant au Québec, Bilan 1989-1999, Québec, Ministère du Développement durable, de l’Environnement et des Parcs, Direction du suivi de l’état de l’environnement, ISBN 2-550-46063-4, Envirodoq n° ENV/ 2005/0283, 24 pages.
This page intentionally left blank
597
Comparison of three pilot plants filled with organic materials for the treatment of air pollutants from a composting plant SÉBASTIEN BASSIVIÈRE, FLORENCE DUCRAY AND CHRISTOPHE RENNER Veolia Environnement, Research Centre for Water, Air & Odours Research Program, Chemin de la digue, 78603 Maisons Laffitte BP 76, France
ABSTRACT This study has compared the efficiency of 3 mixes of organic materials [peat + coco fibre; branches (1 vol) + wood (1vol) + peat (1 vol.); + branches (5 vol) + wood (2vol) + peat (1 vol.)] used in biofiltres pilot plants in order to remove Volatiles Organic Compounds (VOCs), Ammonia and Odours Units (O.U.) from process and ambient air of a composting plant. After a period of seeding, 3 different air velocities have been applied (125 m/h, 175 m/h and 200 m/h) during approximately 1 ½ month each. Results show a better NH3 removal in filters containing wood (92% to 98%) than in the one with coco fibre (between 17% and 63%) depending on the air velocity and on the pollutant concentration (from 35 to 60 ppmV). Concerning VOCs the efficiency was globally the same (approx. 75%) and that for each velocity and for a relatively constant inlet concentration of 20 ppmV Carbon equivalent. Differences are less obvious in terms of Odours Unit that is mainly due to the variability of the analysis itself. The coco fibres filter seems to be a little bit worst than the two others what would be in accordance with the pollutant removal efficiency observed.
1 INTRODUCTION Nowadays the increasing regulation on odours in order to protect the neighbourhood of composting plants does not more allows operators to underestimate the air treatment efficiency on their site. At the moment the best available technology for that kind of polluted gas seems to remain the association of an acid scrubber for ammonia removal and a heterotrophic biological filter for the treatment of Volatile Organic Compounds
597
598
SÉBASTIEN BASSIVIÈRE, FLORENCE DUCRAY and CHRISTOPHE RENNER
(VOCs). However the associated cost of such an air treatment often drives the composting plant’s owner to lower the investment particularly when the technical choice of the bacteria supporting material has to be done. Then the tendency could be to choose a cheap supporting media easily available in the surrounding of the composting plant Nevertheless the pollutant removal efficiency of biofiltres depends a lot on the quality and the characteristics of the material used. So the cost/efficiency balance of the supporting media has to be clearly determined in order to help the operator in his choice. The goal of this experimentation was to evaluate the removal efficiency of VOC, Odours Units and Ammonia by 3 mixes of organics materials that can be used in biological filtration and to estimate the main operational costs associated to each material (ΔP, sell cost, water consumption). To do so, three pilot plants have been operated since August 2006 on a composting plant in the south-west of France.
2 MATERIALS AND METHODS 2.1 ORGANIC MEDIA The 3 materials selected for biofiltration trials were: i)
a mix of peat (1 vol.) and of long coco fibres (1 vol.) (biofiltre F1). This media is widely used in usual biofiltres and is easily available on the market at a cost of 70 €/ton. ii) a mix of oversized branches /pallet fragments/blond peat in a volume ratio of 5/2/1 (biofiltre F2). This media is directly made at the composting plant at an approximate cost of 40 €/ton. iii) a mix of oversized branches /pallet fragments/blond peat in a volume ratio of 1/1/1 (biofiltre F3). Same availability and cost than media 5/2/1. Physical and chemical characteristics of those media have been previously determined at the lab scale (pH, density, void fraction, water holding capacity, pressure drop and compaction) to estimate their potential in being a useful bacterial support (table 1). Differences observed between the three mixes are mainly due to the peat content that lowers the pH value and gives bigger void fraction, water holding capacity but also compaction capacity. The higher compaction capacity of the peat/coco mix represents the only bad point in this evaluation compared to the other media. However those differences cannot lead to eliminate a potential media for biological filtration purposes.
599
COMPARISON OF THREE PILOT PLANTS FILLED WITH ORGANIC MATERIALS
Table 1. Physical and chemical characteristics of the media tested.
pH Density (kg/m3) Void fraction Water holding capacity (%) Pressure drop (Pa/m) (on wet material) Compaction (%) (on wet material)
Peat / Coco Branches/Pallet/peat (5/2/1) 5.2 6.6 353 677 0.83 0.56 81.8 60.2 20 20 17
Branches/Pallet/Peat (1/1/1) 6.1 552 0.58 76.1 30
4
7
2.2 PILOT PLANTS The characteristics of each pilot are: surface = 3.14m2, diameter = 2m, high of material = 1,15 m + 0,25 m of draining layer. Filters are closed with a plastic cover (figure 1). All of them are fed with the polluted air made of a mix of process air pumped out of the compost piles and of ambient air coming from the fermentation and maturation buildings. Each air pipe feeding each filter is fitted with an automated air valve controlled by online air flowmeter in order to ensure a constant flow in spite of pressure drop variations. Four periods have been defined to test increasing air velocities throughout the filters (125 m/h; 175 m/h, 210 m/h and 95 m/h). Filters are maintained wet with tap water sprayed during 15 sec. every hour what delivered a flow of 10Liters/day/m2. The water is not recirculated and is drained to the sewer network.
Figure 1. View of the three pilot plants and air pipes network.
600
SÉBASTIEN BASSIVIÈRE, FLORENCE DUCRAY and CHRISTOPHE RENNER
2.3 ANALYTICAL METHODS Ammonia and VOC were continuously monitored respectively with an Infra Red Gas Filter Correlation analyser (IR GFC) (Model MIR 9000, Environnement S.A.) and with a Flame Ionisation Detector (FID) analyser (Model 109A, JUM) that also measure methane (CH4) and Non Methane Hydro Carbons (NMHC). Sampling was alternatively made at the inlet and the outlet of each biological filter for periods of 20 min. Once a month, drain water of each filter was collected during 24 hours and analysed for NH4, NTK, NO3 and NO2 according to European or International Standardisation. At the same time, air samples had been collected in Tedlar® bags with a Teflon® coated pump. Then odours units (O.U.) determination had been done with a dynamic olfactometer (Model Odile, Odotech Inc.) according to the European Standard NF EN13725. Relative humidity (%) and Pressure Drop (mbar) in filters was measured by appropriate probe and sensor.
3 RESULTS AND DISCUSSION 3.1 PRESSURE DROP AND HUMIDITY The evolution of pressure drop (ΔP) follows clearly the increasing of air flow (and air velocity) (Graph. 1). This phenomenon is less important for biofilter 1 for which the void fraction is significantly higher than in other filters. Biofilter 2, with the lower void fraction, is the one that reaches the highest pressure drop. At the end of trials, the return to initial pressure drop is easily obtained when lowering the air velocity. Theses ΔP values remain very low and acceptable even after a long running time. Humidity remains very constant in every filter and is close to the recommended value of 30 %, and that in spite of a constant water flow for varying air velocities (Graph 2). This very important parameter assesses of the good running of filters and could have been more easily achieved with the use of covers above filters. 3.2 EFFICIENCY OF ODOURS UNIT ABATEMENTS Odours unit have been measure in three different laboratories in order to have a comparison and an idea of the large variations of results that can be associated with such an analysis. For example the sample taken from the air entering the biofiltres we had obtained values rising from 9000 O.U. to 15 000 O.U with the first lab, from 6000 to 28 000 with the second one and from 21 000 to 28 000 with the third lab. Fortunately the tendencies are the same when we consider the evolution of O.U. abatement versus the increases of air flow. Values were started at a very good abatement level of nearly 95% in each filter at an air flow of 390 m3.h-1. Then the more the air
COMPARISON OF THREE PILOT PLANTS FILLED WITH ORGANIC MATERIALS
Graph 1. Evolution of pressure drop in the three biofilters.
Graph 2. Evolution of Relative Humidity (%) in the three filters.
601
602
SÉBASTIEN BASSIVIÈRE, FLORENCE DUCRAY and CHRISTOPHE RENNER
flow increased, the more the abatement efficiency decreased and this to raise no more than 40% in the worst case (Graph 3). Every laboratory also showed that filter 1 was the less efficient when filter 2 and 3 could be considered as similar.
Graph 3. Abatement of Odours Unit (%).
3.3 EFFICIENCY OF AMMONIA AND NMHC REMOVAL The ammonia concentration in the polluted air was at a mean value of 25 mg.Nm3 at the beginning of trials and reached 75 mg.m3 at the end, what begins to be high for such biofiltres. The removal efficiency of NH3 is considerably lower in biofilter 1 (30% to 50% max) than in filters 2 and 3 (around 90% whatever the air flow was). For each filter we can not find any relationship between de NH3 removal and the air flow tested (Graph 4). This is of a great interest when we consider that the usual velocity applied on that type of filter is 100m.h-1 (air flow 315 m3.h-1) and that we could have applied more than twice time this velocity without decrease of the removal efficiency. Whatever the global ammonia abatement was, the analyses made on the drain water have shown that a large amount of NH4+ was found in this water. This led us to conclude that a great part of the removal was more due to a simple washing than to a biological activity. If we make the mass balance of nitrogen in the inlet air and outlet air of filters and in drain water we can see that 15% (filter 1) to 30% (filters 2 and 3) of entering ammonia is founded in drain water (Graph 5). That mean that for filter 1
COMPARISON OF THREE PILOT PLANTS FILLED WITH ORGANIC MATERIALS
Graph 4. Ammonia removal efficiency (%).
Graph 5. Part of Nitrogen (ammonia form) recovered in drain water (%).
603
604
SÉBASTIEN BASSIVIÈRE, FLORENCE DUCRAY and CHRISTOPHE RENNER
approximately 50% of ammonia is eliminate by washing. This value is of 30% for biofiltres 2 and 3. This difference can be explained by the bigger contact surface offered in Filters 2 and 3 compared to filter 1. Concerning the Non Methane Hydro Carbons, tendencies are more difficult to be distinguished. We just could observe that a global removal efficiency of 20% to 50% is obtained (Graph 6). The abatement of NMHC seems to best with an air velocity of 100m.h-1 and lower for higher flows.
Graph 6. Removal efficiency of NMHC (%).
4 CONCLUSIONS In conclusion we could have observed a better ammonia removal on filters filled with a cheaper home-made media than on a traditional media made of coco fiber and peat. This remains true even if we have seen water washing of NH3 on all the filters. The part of this physical treatment is approximately 50% on coco/peat against 25 to 30% on wood/pallet/peat. No great differences occur between filters 2 and 3. A better understanding of the performance differences is going to be conduced, including microbiological investigation. However the removal efficiency of NMHC is similar on all filters. If we consider the same tendencies on Odours Unit abatement, we could suppose that the main part of odours nuisance due to a composting plant seems to remains in the Volatile Organic Compounds content in the air treated
605
Biofiltration systems for the treatment of waste gas from industrial plants IAN PHILLIPS Bord na Móna Environmental Ltd., Main Street, Newbridge, Co. Kilare, Ireland
ABSTRACT Bord na Móna Environmental Ltd. has gained valuable experience over the years with the installation of over 500 biofilters, primarily in municipal and industrial wastewater treatment plants, food processing and more recently printing and coating applications. As a result, this paper will deal with the routes of successful biological air treatment and the pitfalls to avoid. Since the mid 1990s Bord na Móna have concentrated their efforts on process development for VOC and industrial applications. New applications include treatment of airstream containing high concentration of H2S, ammonia, VOC or indeed a combination of all three. The paper will cover new developments and new applications notably in solid waste composting plants, and in industrial applications. Success in biofiltration will be shown to be closely linked with in-depth analytical process engineering, accurate characterisation of waste gas streams, control of process conditions for optimisation of biological activity, and the physio-chemical properties of the filter media. Particular attention will be given to the successful application of biofilters for extremely high concentrations of sulphur compounds, up to 5000mg/m3, and the increasing use of biofilters for VOC elimination.
1 INTRODUCTION In response to a rapidly changing world, with ever increasing pressures on our environment globally, we continue to develop and implement more stringent environmental legislation. Arising from this, and an increased understanding of the impact of pollutants, there is an increasing requirement for economic and environmentally sustainable treatment options for various types of emissions to the environment.
605
606
IAN PHILLIPS
In the late 1980s Bord na Móna identified possible future use for peat as filtration media for treatment of waste air and water emissions. This led to the establishment of Bord na Móna’s Environmental business in 1991. The business has continued to grow at a pace, and is currently active in Ireland, the UK, USA, France, Spain and Italy. This paper traces Bord na Móna’s early experiences and developments with biofiltration systems for treating air emissions. These experiences, coupled with various innovations and ongoing developmental work over the years, have enabled Bord na Móna develop unique patented technologies. The Bord na Móna technologies can treat a range of applications from relatively straightforward odour emissions from municipal wastewater treatment plants through to difficult complex industrial emissions. Extensive application experience has been gained in treatment of off gases from municipal and industrial waste water treatment plants, various industrial emissions and from municipal solid waste handling and treatment facilities. By adopting a bi-directional approach to process development, Bord na Móna has developed a successful biological treatment technology for the treatment of high levels of VOC (Volatile Organic Components). This technology has been applied by coating industries, paint booths, printing processes and various applications in the pharmaceutical sector. Bord na Móna Environmental Ltd. have been awarded patents on the original MÓNASHELL technology. More recently world wide patents have been granted for the enhanced filtration technology for VOC treatment.
2 MATERIALS AND METHODS 2.1 BIOFILTRATION Biofiltration by definition is the aerobic degradation of pollutants in the presence of a carrier media. The early development work on biofiltration technology concentrated on organic media, such as, peat, compost, wood bark etc. In general terms, organic compounds are degraded to carbon dioxide and water, while inorganic compounds, such as, sulphur compounds are oxidised to form oxygenated derivatives. The formation of these acidic compounds can lead to a lowering of pH of the filtration media; which in turn impacts on the performance of the system. Removal of the oxidised compound from the media is an important consideration in the design of biofiltration systems. Biofiltration has long been considered to be something of a «Black Art» rather than a Science. Variations in system design, filtration media, process conditions, operational conditions and system controls all greatly impact on successful treatment. The track record for biofiltration system in the early 1990s was poor and these systems were not widely held to be reliable for air pollution application.
BIOFILTRATION SYSTEMS FOR THE TREATMENT OF WASTE GAS FROM INDUSTRIAL PLANTS
607
2.2 RESEARCH & DEVELOPMENT PROGRAMME Bord na Móna identified the UK municipal market as a target market for biofiltration systems. A research and development programme was set up to establish the optimum operational conditions for the treatment of sulphur compounds (principally H2S, Mercaptans and Alkyl Sulphides) using peat based biofiltration materials. The outcome of these results can best be summarised as follows: 2.2.1 Peat/organic media based filtration systems are suitable for municipal applications and can treat levels typically up to a maximum of 50mg/m3. 2.2.2 Two loading regimes were identified as optimum for airstreams with a maximum of 15mg/m3 H2S and 50mg/m3 H2S. 2.2.3 The limiting factor for treatment of sulphur compounds in the system was identified as pH. A control system based on the operation of an intermittent irrigation system was developed. 2.2.4 Various peat media were trialled and optimum media specifications identified. 2.2.5 Optimum operating parameters in terms of gas loading, temperature, pH and operation of irrigation system were identified. In addition to the above, the importance of the physical and chemical properties of the media in assisting in capture and in providing the conditions for successful treatment were identified. One of the most important aspects of biofiltration technologies is the physico-chemical characteristics of the media. The importance of having a homogenous matrix offering minimum resistance through which air will pass, making good contact with the surface of the media matrix cannot be over emphasised. The outcome of the R&D programme was that Bord na Móna could design and install systems offering process guarantees, conditional on the systems being operated within the correct parameters. 2.3 MÓNAFIL BIOFILTRATION TECHNOLOGY During this period Bord na Móna developed its patented MÓNAFIL granular peat media. This media is a fractionalised high density peat media. The media has a high «Air Filled Porosity» (85%) and exhibits excellent physical characteristics. It can be installed to a depth of 3 metres and has been shown to have a media life well in excess of five years. A further advantage is that the media can be regarded at the end of its life with up to 50% being available for re-use (2 years typical for compost –
608
IAN PHILLIPS
wood bark). This technology is now also used extensively for treatment of off-gasses from municipal solid waste treatment facilities, including composting plants.
3 RESULTS AND DISCUSSION Field experience on municipal applications confirmed the R&D findings that pH was indeed the limiting factor. If the system saw high levels of H2S, it was found that the oxidative by-products of H2S oxidation resulted in lowering in pH of the media. At low pH, while H2S removal still remains high, odour removal efficiencies tend to deteriorate particularly if Alkyl Sulphides and Mercaptans are present in the air. Many of the potential applications in the UK market were on coastal sites where saline infiltration was a feature resulting in difficult sludges with heavily concentrated air emission (levels up to 1000ppm H2S). It was found that the peat media biofilters became overloaded with a drop in media pH and thus reduction in efficiency. 3.1 MÓNASHELL BIOFILTRATION TECHNOLOGY Developments in biofiltration for treatment of high level sulphur contaminants enabled Bord na Móna compete in this sector. It was decided to develop a roughing filter for operation upstream of the Móna peat technology. Shells were identified as a media offering the following potential advantages: 3.1.1 In-built buffering capability due to the chemical make-up of shells (calcium carbonate). 3.1.2 High air-filled porosity. (AFP) 3.1.3 Ability to sustain high irrigation rates and capacity to retain large quantities of water. 3.1.4 Shape and size of packing is in the correct range for good mass transfer. 3.1.5 Calcium Carbonates are known to be a good media for supporting biological activities. Laboratory and field trials were carried out on shell-based systems. Early trial results indicated excellent results, so much so that it was decided to develop the process as a stand alone technology. Patents were applied for and the technology was launched in 1995. Since then, over 400 installations have been installed worldwide for airstreams with levels typically up to 500ppm H2S.
BIOFILTRATION SYSTEMS FOR THE TREATMENT OF WASTE GAS FROM INDUSTRIAL PLANTS
609
To date systems have been installed on airstreams treating levels up to 2,000 ppm H2S. Treatment for levels up to 10,000 ppm down to levels below 1ppm at the outlet have been achieved by operating multipass systems through a number of units in series. 3.2 ENHANCED FILTRATION TECHNOLOGY FOR VOC TREATMENT As a result of market demands, and with a view to pending and current European VOC directives and legislation, a significant requirement was identified for low cost treatment of low to medium, and medium to high levels of air-borne VOC from industrial processes (chemical / pharmaceutical), paint booths, printers and industrial coatings. Conventional wisdom held that biological treatment of medium to high level VOC air contamination was not possible due to poor solubility and the persistent nature of compounds. Conventional treatment was incineration, with the addition of supplementary fuel (high operation cost). The potential advantages that biological treatment offered were identified as low cost, low energy and environmentally sound and sustainable solutions. From application experience, it was known that the main limiting factors when treating VOC are as follows: • Limited solubility of many organic compounds leading to poor capture and treatment. • Excessive biomass production leading to plugging of filter media with excessive back pressure and reduced airflow. In 1997 a VOC research project was initiated. From the outset it was decided to adopt a bi-directional approach to the development of a biological technology for treatment of VOC levels as follows: • Explore the potential of Bord na Móna’s existing technologies for use on VOC application. • Enhancement of existing process by combining existing Bord na Móna processes with other technologies to enhance treatment and effect the following: – Increase mass transfer of contaminants to the aqueous phase – Treat high concentration regimes – Control excess biomass over growth. At the outset it was demonstrated that MÓNASHELL was capable of removing 15-20g of carbon per hour. This value is consistent with other biological systems. Two dynamics were explored to enhance solubility and capture as follows:
610
IAN PHILLIPS
3.2.1 RECIRCULATION OF AIR A dynamic had been developed on a high H2S application whereby treated air was recirculated to the inlet of the filter. This has a number of effects as follows: 3.2.1.1 Dilution of inlet concentration 3.2.1.2 Increased rate of mass transfer (gases to Aqueous) overcoming problem of insolubility. 3.2.1.3 Improvement of elimination capacity This dynamic effectively allows the increase of the elimination capacity of the filtration system from circa 20g/m3filter/hr to circa 60g/m3filter/hr for the same net contact time. Thus, high removal efficiencies are also achievable. 3.2.2 ELECTROMAGNETIC STIMULATION As part of a study to enhance solubility the use of electro-magnetic stimulation of the water was also examined. While carrying out this work it was noticed that use of electromagnetic stimulation enhanced system performance and prevented excess biomass formation. Supplementary nutrients are not added in the process. The system effects metabolism of solvents with only minimal synthesis of biomass. Conventional biological systems operate on the basis of converting pollutants to biomass. The dynamic in Bord na Móna’s enhanced technology is reliant on metabolic activity to produce CO2, H2O and exothermic energy. The first commercial installation on high level VOCs was commissioned at B.P.I. Ardeer, a printing application in Scotland in February 2002. This system is monitored remotely and has been operating successfully achieving specified outlet levels. The second phase of the project is currently under construction.
4 CONCLUSIONS The experience which Bord na Móna has gained over the years has proved that biological technologies can be successfully applied to difficult applications with remarkably predictable performance. Biological systems need to be engineered such that all critical parameters can be monitored and controlled. Biofiltration is successfully emerging from the shadows as a reliable, low cost option for a broad range of air treatment applications. It is now becoming apparent that biological treatment will play a far more significant role in achieving environmental control on air emissions.
BIOFILTRATION SYSTEMS FOR THE TREATMENT OF WASTE GAS FROM INDUSTRIAL PLANTS
611
Further related areas of study under review at present are: • • • • •
Comparison of different shell based media. Development of Multi-Pass system versus recirculation (reduced power requirements). Buffering of emission from non continuous processes. Optimum maintenance requirements for control of pressure drop through system. Co flow versus counter flow.
CASE STUDIES Table 1. MÓNASHELL case study Location Application Date of installation Biofilter size Total gas flow rate Inlet odour concentration Outlet odour concentration Odour removal efficiency
Sewage Treatment Plant, Ireland Treatment of emissions from picket fence thickener April 1998 3 m3 160 m3/hr 12,722 ou/m3 294 ou/m3 98%*
* Determined by Force Choice Dynamic Olfactometer Table 2. MÓNASHELL case study Location Application Date of installation Biofilter size Total gas flow rate Removal efficiency for individual compounds H2 S Mercaptans Dimethyl Sulphide * Determined by Force Choice Dynamic Olfactometer
Sewage Treatment Plant, Ireland Treatment of emissions from biotower September 1997 70 m3 1,000 m3/hr Inlet Concentration Removal 40-100 98-100% 88% 37 ou/m3 97%* 95%
612
IAN PHILLIPS
Table 3. MÓNASHELL case study Location Application Date of installation Biofilter size Total gas flow rate Inlet odour concentration Outlet odour concentration
Sewage Treatment Plant, Ireland Treatment of emissions from picket fence thickener April 1998 3 m3 160 m3/hr 12,722 ou/m3 294 ou/m3
* Determined by Force Choice Dynamic Olfactometer Table 4. MÓNASHELL case study Location Application Date of installation Biofilter size Total gas flow rate Removal efficiency for individual compounds H2 S Mercaptans Dimethyl Sulphide
Sewage Treatment Plant, Ireland Treatment of emissions from biotower September 1997 70 m3 1,000 m3/hr Inlet Concentration Removal 40-100 98-100% 37 ou/m3 88% 97%* 95%
* Determined by Force Choice Dynamic Olfactometer Table 5. MÓNAFIL case study Location Application Pre-treatment of non-condensable gases Biofilter volume Total gas flow rate Inlet odour concentration (after pre-treatment) Performance Date December 1992 May 1993 April 1993
Animal by-product Rendering Factory ventilation air, non-condensable process gases Cyclones for removal of particulates packed tower acid scrubber 250 m3 25,000 m3/hr 10,000 – 50,000 ou/m3 Mean odour removal efficiency across filter bed 98.7% 98.5% 99.5%
* Determined by Force Choice Dynamic Olfactometer
BIOFILTRATION SYSTEMS FOR THE TREATMENT OF WASTE GAS FROM INDUSTRIAL PLANTS
Table 6. MÓNASHELL case study Location Application Date of installation Biofilter size Total gas flow rate Inlet odour concentration Outlet odour concentration Odour removal efficiency Inlet H2S concentration
Industrial Plant, Ireland Treatment of emissions from anaerobic digestor January 1998 24 m3 200 m3/hr 434,531 ou/m3 508 ou/m3 99%* 1,600 ppm
* Determined by Force Choice Dynamic Olfactometer
Table 7. MÓNAFIL case study Location Application Date of installation Biofilter size Total gas flow rate
Pharmaceutical Plant, Ireland Treatment of emissions from industrial treatment plant September 1997 70 m3 3,500 m3/hr
Table 7.1 VOC analysis by GG-MS+ PARAMETER (mg/m3) Toluene Dichloromethane MIBK* TOH** Total Hydrocarbon
BIOFILTER INLET 8.50 106.60 3.00 1.80 119.90
* MIBK = Methyl is obatyl Ketone ** Total Other Hydrocarbons
BIOFILTEROUTLET 1.00 27.50 0.07 0.90 29.50
% REMOVAL 88.20 74.20 97.90 50.00 76.40
613
614
IAN PHILLIPS
Table 8. MÓNAFIL case study Location Application Date of installation Biofilter size Total gas flow rate
Composting plant, Italy Treatment of emissions from composting facility 1996 300 m3 20,000 m3/hr
Table 8.1 VOC analysis by GC-MS+ PARAMETER (mg/m3) Toluene Limonene Other C10H16 TOH* Xylenes/Benzenes Total Hydrocarbon
BIOFILTER INLET 3.6 28.9 1.8 9.9 5.9 50.1
BIOFILTEROUTLET 0.7 9.6 0.3 2.3 2.3 15.2
% REMOVAL 80.5 66.8 83.3 76.8 61.0 70.0
* Total Other Hydrocarbons
Table 8.2. Removal of odorous compounds from air streams utilising MÓNASHELL COMPOUND Mercaptan Mercaptan Ammonia Ammonia Ammonia Triethyl Amine Triethyl Amine
CONCENTRATION ppm 20 70 30 45 100 15 35
% REMOVAL 100 97 100 100 100 100 96
BIOFILTRATION SYSTEMS FOR THE TREATMENT OF WASTE GAS FROM INDUSTRIAL PLANTS
615
Table 9. MÓNASHELL enhanced biofiltration case study Location Application Date of installation Biofilter size Total gas flow rate Typical Emissions Components Removal Efficiencies
Printing Plant, Scotland Treatment of emissions from 8-colour printer 2001 2 x 46 m3 11,500 m3/hr 500-1200 mgC/m3 Isopropyl Alohol, MEK, Ethyl Acetate, Butyl Acetate, Ethanol to <150 mg C/m3